That is what we are going to do today.

Take Russia, the war in Ukraine, and the confrontation between Russia and the West, which has continued to spread into more areas of the economy.

This has already caused some major changes.

Europe has had to rearm. Russian oligarchs have had assets frozen. Governments and businesses have had to rethink supply chains they previously took for granted.

One consequence runs through all of this: the West wants to reduce its dependence on Russia.

Energy is an obvious part of that.

But there is one dependency that deserves more attention: uranium enrichment.

And that is the subject I want to explore with you today.

In 2025, global uranium requirements were estimated at 68,920 tonnes.

But uranium has to pass through several industrial stages before it can become usable fuel for most nuclear reactors: conversion, enrichment, and fuel fabrication.

And in 2025, Russia’s Rosatom accounted for an estimated 43% of the world’s uranium enrichment capacity.

That is considerable.

We are talking about global capacity.

The United States illustrates how significant this dependency remains.

In 2025, American nuclear utilities purchased 12.7 million separative work units, or SWU—the unit used to measure enrichment services—at an average price of $108.70 per SWU.

That represents approximately $1.38 billion in enrichment purchases.

Russia supplied 26% of the services purchased.

So this is a Western dependency with a real industrial chain behind it—and real money flowing through it.

The interesting part comes when we think about what reducing that dependency actually requires.

Governments can announce sanctions or set a deadline. But replacing a supplier requires another supplier with the technology, the licences, the facilities, and the capacity to deliver.

Those capabilities take years to build.

This is where the investment thesis begins.

When countries decide to change where they buy something strategically essential, they can create demand for new industrial capacity—even before total demand for the product increases.

In this case, that means new contracts, new investment, and potentially more value for the companies capable of supplying an alternative.

There is a useful paradox here: the world still has excess enrichment capacity overall, yet Western buyers are supporting the construction of more.

The location of that capacity—and who controls it—has become part of its economic value.

That is what I want to investigate with you today: how the effort to reduce dependence on Russian nuclear fuel is reshaping the industry, and where the resulting investment opportunities might be.

The rebuilding has already started

We can already see this shift turning into investment.

In January 2026, the U.S. Department of Energy announced $2.7 billion to strengthen domestic uranium enrichment over the next ten years.

The funding supports capacity for both the existing reactor fleet and future advanced reactors, with payments tied to specific milestones.

In France, Orano is investing more than €1.7 billion to expand its Georges Besse II enrichment plant.

The project will add approximately 2.5 million SWU of annual capacity. The new units are expected to start operating progressively in 2028, with full commissioning planned for 2030.

And Urenco’s order book reached €27.3 billion at the end of June 2026, up from €21.3 billion six months earlier.

Those contracts extend into the 2040s.

These figures measure different things: public funding, industrial investment, and future contracted revenue. But together, they show the same direction.

Buyers are making commitments that support the construction of an alternative fuel supply chain.

That matters because an enrichment plant requires substantial investment long before it starts delivering fuel.

A customer willing to commit for years can help make that investment possible.

This is one of the things I look for when researching an emerging investment thesis: the point where a strategic ambition starts becoming a commercial commitment.

Announcements tell us what governments and companies want to do.

Contracts help us understand what they are prepared to pay for.

Two forces are now moving through the same industry

The first is the effort to diversify away from Russian supply.

The second is the potential growth of nuclear power itself.

The World Nuclear Association’s reference scenario projects annual uranium requirements rising from approximately 68,920 tonnes in 2025 to more than 150,000 tonnes by 2040.

That would be an increase of roughly 118%.

This is a scenario, and its realization will depend on reactor construction, operating lifetimes, policy, and financing.

But it gives us a useful way to frame the opportunity.

The industry may have to support a larger nuclear fleet while also changing where part of its fuel comes from.

And these two forces operate on different timelines.

Diversification concerns reactors that are already running and need to secure future deliveries. Growth adds demand as new reactors enter service.

The investment opportunity is in how those needs become revenue.

An established supplier might benefit by signing new contracts, improving its realized prices, or expanding an existing facility.

A new entrant might benefit by building capacity that customers are already willing to support.

A technology developer might benefit if its process becomes commercially competitive.

Each is exposed to the same broad change, but each presents a very different investment case.

That distinction is where the research becomes important.

A company with an operating plant and long-term customers has a different risk profile from one still working toward commercial production.

A large contract can provide visibility, but its value depends on the delivery schedule, pricing terms, costs, and conditions attached to it.

And building strategically important infrastructure can still consume substantial capital before it generates an attractive return.

This is why I want to follow the money through the entire chain: from the customer’s commitment to the supplier’s investment, then to deliveries, earnings, and cash flow.

That is how we can assess whether the nuclear fuel sovereignty race is creating an attractive opportunity for shareholders.

In the rest of this edition, I will examine:

• The industry’s constraints: where conversion and enrichment capacity is concentrated, what buyers can access, and how the planned expansions could change the market.

• Three listed expressions of the thesis: Cameco, Centrus, and Silex—and the different businesses, timelines, and risks behind their exposure.

• The economics of the contracts: how prices, customer commitments, and new capacity could translate into earnings and cash flow.

• The additional opportunity in advanced fuels: what HALEU could contribute, and how much depends on reactors that still need to be financed and built.

• Valuation and scenarios: what would need to happen for each company to justify its price and deliver an attractive return.

• A framework for tracking the thesis: the milestones that would strengthen the investment case, and the developments that would make me reconsider it.

The question I want to answer is straightforward:

Which companies can turn this rebuilding of the fuel supply chain into lasting economic value—and how much are we being asked to pay for that opportunity?

Read more

While virtual AI tools like ChatGPT scale effortlessly in software, bringing AI into the physical world is a completely different story. Real-world robotics faces tough physical constraints—weight, friction, and precise spatial handling.

In today’s episode of Macro Notes Podcast (created with NotebookLM), we look beyond the hype of individual robot makers and focus on the real engine behind this $5 trillion transition: the foundational hardware. We break down why component suppliers—making sensors, smart motors, and actuators—represent the ultimate “picks and shovels” opportunity across the entire robotics ecosystem.

Note: This episode covers the insights from our free research edition. If you want direct access to our full stock selections, portfolio breakdown, and real-time trade updates, upgrade to a Premium or Founding membership at macronotes.com

Read the post

I recently secured MacroNotes.com for $10,000, and we’ve just launched our new website.

It brings together our research approach, membership plans, and wider newsletter collection, including Public Markets, which, like Macro Notes, is a Substack Bestseller.

You can explore it here (it’s a first version, with more to come): Visit MacroNotes.com →

The website is our new front door. The newsletter, subscriptions, and your existing membership remain on Substack.

Investing more in the research you receive

More of you are reading Macro Notes, sharing our work, and becoming Premium subscribers.

That support gives me both the means and the motivation to keep improving what we publish: deeper investigations, stronger analysis, and more original research.

So far in 2026, I’ve invested a little over $70,000 in developing Macro Notes, including detailed investigations into specialized sectors such as AI infrastructure.

In 2027, I plan to double that investment.

There is still so much we can build. I want to bring in more independent analysts with specific industry expertise, expand our coverage of global macro developments and investment strategies, and uncover insights that are difficult to find through conventional research.

The purpose is straightforward: give you better evidence to understand which investment theses deserve your attention—and why.

How we turn global shifts into investment theses

Macro Notes takes a global macro, top-down approach.

We start with the forces reshaping the economy: technological change, geopolitical shifts, industrial policy, infrastructure investment, and changing flows of capital.

Then we work down to the industries and businesses affected.

Where is demand growing? Where are the bottlenecks? Which companies can capture the value being created? And how much of that opportunity is already reflected in their valuations?

We follow more than 100 investment theses across areas such as AI infrastructure, defense, healthcare, and emerging industries. Dedicated market indexes help us track how these themes evolve over time.

But market performance is only one part of the picture. We also examine business fundamentals, industry developments, catalysts, and the evidence that could challenge our assumptions.

Each month, we select the thesis we believe offers the most compelling combination of potential, supporting evidence, and risk relative to the possible return.

The goal is to narrow a broad universe of ideas down to a few candidates worth researching in depth—and keep following them as the evidence changes.

A wider research collection with Founding

The Founding membership includes the full Macro Notes Premium research, access to our wider newsletter collection, and participating partner newsletters.

Public Markets and its Private Markets coverage are part of that broader perspective, alongside publications such as Solo Capitalist.

It’s a way to explore different markets and research approaches through one membership, while supporting the continued development of Macro Notes.

You can find the full list of included publications and membership details on our new website.

If you currently subscribe to Premium Monthly and would like to move to Premium Yearly or Founding, reply directly to this email.

I’ll work out an upgrade rate that takes into account how long you’ve been a subscriber and what you’ve already paid.

What’s next for Altis.finance

We’re also developing a more complete version of Altis.finance, the research platform behind our investment thesis indexes and portfolios.

The ambition is to go beyond viewing performance charts. We want to help you work through investment ideas: explore the businesses behind a thesis, compare opportunities, assess risks, and follow how the evidence develops.

We’re also working on a rebrand. “Altis” is already widely used across the financial industry, and I want the platform to have a more distinctive identity.

I’ll share more as the development progresses.

Our latest must-read editions…

Introducing a monthly guest edition

We recently published our first guest edition with MarketFighter, exploring his systematic approach to investing:

His approach differs from our usual thesis-driven research, which is precisely why I wanted to share it.

I plan to publish one guest edition each month, with a different contributor bringing a perspective, idea, or strategy worth exploring.

I’ll be selective. The quality of the work, the reasoning behind it, and its relevance to you will matter more than the author’s subscriber count.

If you publish investment research or write a newsletter and would like to contribute an edition to Macro Notes, reply to this email and tell me about your work.

Soon: follow our investment thesis indexes on MacroNotes.com

The website we’ve just launched is the first version of MacroNotes.com. It will evolve over time, with one goal: to improve your experience beyond the newsletter itself.

Our investment thesis indexes will soon be available directly on MacroNotes.com for Premium and Founding subscribers.

You’ll be able to track the market performance of the themes we research in one place, without having to sign up for Altis.finance or create a separate account on another platform.

This is the next step toward bringing our research and its ongoing performance tracking together, so you can more easily follow a thesis from the edition where we introduce it through its subsequent development.

Thank you for reading, subscribing, and helping this publication grow. Your support makes these next steps possible.

Pierre MJ
Founder, MacroNotes.com

Then the internet democratized information.

Then Substack, X, podcasts, and online communities democratized specialized analysis.

And now LLMs are democratizing a large part of the analytical work itself: summarizing earnings, comparing companies, calculating multiples, explaining industries, reading filings, identifying competitors, and more.

The result is almost paradoxical:

We probably have access to more information and analytical capacity than any previous generation of investors, yet that does not necessarily mean we make better investment decisions.

Because the nature of the problem has changed.

The challenge is no longer simply getting access to the right information, or even producing a reasonable first layer of analysis. Those remain important, but they are becoming easier to obtain, faster to produce, and increasingly available to everyone.

What is still scarce is the ability to decide which information actually matters, how much of it is already reflected in the price, what would change the thesis, how much risk to take, and how to behave once real money is involved.

And this is where psychology becomes much more important than it first appears.

I read Thami Kabbaj’s “Psychologie des grands traders” some time ago, and one of the ideas that stayed with me is that psychology is not something separate from the investment process. It is part of the process itself.

It is easy to think about investing as a purely analytical exercise. You gather information, build a thesis, estimate the potential upside and downside, and then make a decision.

In reality, the most difficult part often begins after the analysis is finished.

A stock you like suddenly falls 25%.

Another one you decided not to buy rises 60% in a few months.

A company you own reports weaker earnings.

A sector you have been studying starts moving quickly and everyone around you seems to already be positioned.

At that point, the question is no longer just whether your analysis was correct.

You are now making decisions while money, uncertainty, ego, regret, fear and opportunity are involved at the same time.

And that changes the way we think.

This matters even more today because the infrastructure around investing has changed too.

Modern brokerage platforms have done something extremely positive: they have made investing simpler, cheaper and more accessible than ever before. Opening an account, buying a stock and following a portfolio can now be done in minutes.

But the same simplification has another side.

The friction around execution has almost disappeared.

You see a price, a chart, a percentage move, and a buy or sell button. Everything is designed to make action easy.

What has not become easier is deciding whether that action actually makes sense.

In some ways, we have built extremely efficient interfaces for executing decisions without necessarily improving the process behind those decisions.

And that is where many of the most common investing mistakes begin.

See what Macro Notes Premium looks like

In The $600 Billion Sovereign AI Race, I break down one of the biggest emerging investment themes in AI infrastructure: why governments are building sovereign AI capacity, where the capital is going, and which parts of the ecosystem could benefit.

Read the Premium edition →

When price starts changing the thesis

FOMO is probably the most obvious example, particularly after several years of very strong performance across parts of the AI trade.

Imagine you start looking at a company when the stock is trading at $50.

You like the business, you understand the opportunity, but you decide that the valuation already looks demanding. So you wait.

A few months later, the stock is trading at $80.

At this point, something interesting often happens.

The business may not have changed enough to justify a completely different conclusion, but the environment around the stock has.

More analysts are talking about it. The company appears more frequently in financial media. Your feed is filled with investors explaining why the opportunity is still early. People who already own the stock are posting large gains.

And suddenly the same company that did not feel compelling enough at $50 starts feeling difficult to ignore at $80.

This is what makes FOMO more subtle than simply “buying because a stock is going up.”

The real problem is that price itself can gradually become part of the thesis.

A rising stock creates social proof. It makes the bullish case feel more credible. It reduces the discomfort of being wrong alone because the market appears to agree with you.

The move becomes evidence.

But a higher price does not necessarily make the investment safer. In many cases, it simply means that more of the optimistic scenario is already reflected in the valuation.

So the useful question is not whether the stock has gone up.

It is whether something has changed enough to justify paying more for the same future.

What changed between $50 and $80: the business, the expectations embedded in the price, or simply our perception of the opportunity?

That distinction is important because FOMO does not necessarily push investors into bad companies.

Quite often, it pushes them into good companies at increasingly difficult prices.

And being right about a company is not the same thing as making a good investment in that company.

You can be completely right about AI demand, datacenter growth, power consumption, fiber optics, memory or robotics and still generate a disappointing return if the price you paid already assumed an even better future.

Markets do not reward us simply for predicting what happens next.

They reward us for identifying the difference between what happens next and what the market already expects to happen.

That is a much harder problem.

A quick note for free readers

This edition is intentionally fully public.

Macro Notes Premium is focused less on general investing education and more on applying this type of framework to what is happening in markets right now.

I spend most of my research time looking for emerging investment themes, understanding what the market may still be underestimating, and following those ideas as the evidence develops.

Recent research has covered areas including AI infrastructure, European defense, fiber optics, power demand, robotics and other themes where I believe the market may still be early in understanding the full opportunity.

If that is the type of research you are looking for, you can join Macro Notes Premium here.

Subscribe now

Otherwise, keep reading — the rest of this edition is free.

When research becomes defense

Confirmation bias becomes even more interesting in the age of AI because research itself is becoming easier to bend toward what we already want to believe.

Imagine you buy a stock after spending several days building a thesis around the company. A few months later, the next earnings report disappoints. Revenue growth slows, guidance is weaker than expected, and the stock sells off.

At that point, the most useful question would probably be whether the new information materially changes the original thesis.

But that is not always the question we naturally ask.

Instead, once we already own the stock, we tend to look for explanations that allow us to preserve our existing view. We may ask an LLM why the earnings miss could be temporary, why the market might be overreacting, or which parts of the report were actually more encouraging than the headline suggests.

And the problem is that the model will usually be able to give us several perfectly plausible answers.

This does not make the analysis wrong. Some of those arguments may be completely valid.

The issue is that the direction of the research has changed.

Before investing, we were trying to understand whether the thesis was good.

After investing, we may quietly start trying to defend it.

That is where research can turn into confirmation bias.

AI makes it dramatically easier to explore an investment from multiple angles, which is extremely useful. But it also makes it easier to generate increasingly sophisticated arguments for a conclusion we had already reached.

So the real advantage is not simply being able to produce more analysis.

It is maintaining enough distance from our own position to ask the uncomfortable question first:

What evidence would make me change my mind?

That question should ideally be answered before the position exists, not after the stock has already fallen.

Because once money is involved, our interpretation of information changes.

Your entry price means more to you than it does to the market

Another good example is anchoring.

Suppose you buy a stock at $100.

A few months later, it trades at $70.

A very common reaction is to think:

I will sell when it gets back to $100.

But there is nothing inherently important about $100.

It matters to you because it is your entry price.

It determines whether the position appears green or red in your portfolio and whether selling feels like realizing a gain or admitting a mistake.

But the market does not know where you bought the stock.

The company’s customers do not care.

Its competitors do not care.

Its future cash flows do not care.

The more useful question is much simpler:

If I had no position today, would I buy this company at $70?

If the answer is yes, holding may make sense.

If the answer is no, then the fact that you originally paid $100 is not a particularly strong reason to continue owning it.

Anchoring also works in the opposite direction.

A stock falls from $200 to $100 and suddenly looks “cheap” because it is down 50%.

But cheaper than its previous price does not necessarily mean cheap relative to its future cash flows or the expectations embedded in the valuation.

The previous price may simply have been wrong.

This is one of the recurring problems in investing: numbers that are personally meaningful start influencing decisions even when they have little economic relevance.

Why we often sell the winners first

Loss aversion creates a related problem.

Imagine you own two stocks.

One is up 40%.

The other is down 40%.

Which one feels easier to sell?

For many investors, it is the winner.

Selling the winner creates a sense of completion. You made a good decision, you realized the gain, and the result is now secure.

Selling the loser feels very different.

It means accepting that the original decision did not work.

So we wait.

Maybe the next earnings report will be better.

Maybe the stock will recover.

Maybe we can at least get back to breakeven.

The problem is that none of those thoughts tells us very much about expected returns from today’s price.

The relevant question is not:

Am I currently up or down?

It is:

From this price forward, which investment offers the better expected return relative to the risk?

That sounds obvious in theory.

It is much harder in practice because we do not experience gains and losses symmetrically.

A realized loss feels like a mistake becoming permanent.

A realized gain feels like confirmation.

That can lead investors to sell positions that are still working simply because they are profitable, while keeping weaker positions because realizing the loss is psychologically uncomfortable.

Of course, sometimes selling the winner and keeping the loser is the correct decision.

The bias is not the outcome itself.

The bias is allowing the emotional status of the position to become part of the investment thesis.

The recent past is always easier to imagine continuing

Recency bias becomes especially important after long, powerful market trends.

When a sector has performed well for several years, the assumptions investors consider normal tend to shift.

Take AI infrastructure.

At the beginning of the cycle, the important question may have been whether spending on compute infrastructure would continue to grow rapidly.

After several years of accelerating investment, that assumption becomes much less controversial.

The discussion gradually moves from whether spending will grow to how quickly it will continue to accelerate.

That is a subtle but important change.

An optimistic scenario that once required strong evidence can slowly become the baseline assumption.

The same thing happens with valuation.

Multiples that would have looked extreme early in the cycle start to appear normal simply because investors have spent several years seeing them.

Then the recent environment becomes the reference point for what the future should look like.

This works in the opposite direction too.

After a sector falls 60% or 70%, investors often struggle to imagine what could make it attractive again.

After several weak quarters, temporary problems start to feel structural.

After a recession, cyclical weakness starts to look permanent.

We are very good at adapting to the current environment.

The danger is assuming that the current environment is permanent.

A useful question is therefore:

Which assumption in my thesis would look most obviously wrong if conditions simply returned to something closer to normal?

That does not mean every strong trend has to reverse.

Sometimes structural changes really do create a new regime.

But distinguishing a genuine structural shift from extrapolation is one of the hardest parts of investing.

Sometimes the best decision is no decision

There is another bias that I think receives less attention because it is partly created by the way modern investing works.

Action feels productive.

You research something, reach a conclusion and place a trade.

There is a clear outcome.

Doing nothing feels different.

You may spend several hours researching a company and conclude that the business is attractive but the price is not.

Nothing happens.

There is no position in the portfolio.

No confirmation screen.

No immediate reward for the work.

This is one reason action bias can become so powerful.

When markets are open every day and prices are constantly moving, there is always something to react to.

There is always another stock moving 10%, another earnings report, another headline, another opportunity that appears urgent.

But investing does not necessarily reward the frequency of decisions.

Sometimes the correct conclusion is to buy.

Sometimes it is to sell.

And sometimes the correct conclusion after doing all the work is simply to wait.

That is still a decision.

Cash is not necessarily the absence of conviction.

Waiting for a better risk/reward is not necessarily indecision.

The ability to do nothing when nothing needs to be done is probably one of the more underrated parts of investing.

And ironically, the easier execution becomes, the more valuable that ability may become.

The problem is not emotion itself

At this point, the obvious conclusion might be that better investors simply need to become less emotional.

I do not think that is realistic.

Knowing that FOMO exists does not prevent you from experiencing it.

Knowing about loss aversion does not make a 40% drawdown emotionally neutral.

Understanding confirmation bias does not make it pleasant to discover information that contradicts six months of your own research.

The objective is probably not to eliminate emotion from investing.

It is to build a process that gives emotion less control over the final decision.

And one of the simplest ways to do that is to make some of the important decisions before the emotional situation exists.

Build the decision before the emotion

Before buying a stock, I think it is useful to write down a few things explicitly.

Not because a checklist can predict the future.

It cannot.

But because it gives you a reference point created before price movements, P&L and emotion start changing the way you interpret information.

The first question is the thesis itself.

What do I believe that the market may be underestimating?

This is different from asking why the company is good.

A great business can still be a poor investment if everyone already understands exactly how good it is.

The second question is expectations.

What does the current price already appear to assume?

A company growing 30% is not necessarily attractive if the valuation requires something closer to 50%.

Conversely, relatively mediocre growth can generate excellent returns when the market is pricing an even worse outcome.

Then come the catalysts.

What could cause the market to change its view?

You can be right for a very long time without making money if there is no mechanism that causes expectations to move closer to your view.

Just as important is invalidation.

What evidence would prove the thesis wrong?

This is probably the question I would place above all the others.

It is much easier to define what would change your mind before you own the position than after the position is already down 30%.

Then there is valuation.

What future am I effectively paying for at the current price?

Position sizing.

How much capital am I willing to put behind this idea if the thesis is wrong?

And finally, behavior.

What would I do if the stock fell 20%?

What if it rose 50% very quickly?

What if the thesis appeared intact but the stock went nowhere for a year?

The objective is not to create rigid rules that can never change.

New information should change decisions.

The objective is simply to make the starting logic explicit.

In practice, a basic version could look like this:

THESIS
What do I believe that the market may be underestimating?

EXPECTATIONS
What appears to already be priced in?

CATALYST
What could make the market change its view?

INVALIDATION
What evidence would prove me wrong?

VALUATION
What assumptions are required to justify today's price?

POSITION SIZE
How much capital am I willing to risk?

BEHAVIOR
What will I do if the stock falls 20%, rises 50%,
or goes nowhere for a year?

I think of this less as a checklist and more as a small contract with your future self.

When the stock falls sharply, instead of immediately asking whether you should sell, you can return to the original thesis and ask whether anything has actually been invalidated.

When it doubles, instead of selling simply because the gain feels large, you can ask what the market is now pricing.

When a negative article appears, you can ask whether it changes one of the assumptions you wrote down before buying.

And when the price moves significantly without a corresponding change in the underlying business, you have something more useful to compare against than your emotional reaction to the chart.

None of this guarantees a good outcome.

Good investment processes still produce losing investments.

The point is to separate a bad outcome from a bad decision.

Those are not always the same thing.

The investing edge is moving

AI is going to make financial research significantly better.

It will make data easier to access.

It will make company analysis faster.

It will make it easier to compare businesses, build scenarios, understand industries and process information that previously required much more time.

I think that is an extraordinary development.

But it also changes what is scarce.

If everyone can summarize the same earnings call, the summary itself becomes less differentiated.

If everyone can produce a reasonable valuation model, simply having the model becomes less differentiated.

If everyone can generate ten arguments explaining why an industry will grow, producing those arguments becomes less differentiated.

The difficult questions move somewhere else.

Which information actually matters?

What does the market already expect?

Where might consensus be wrong?

What evidence would invalidate the thesis?

How much risk should I take?

And once money is involved, can I still follow the process I designed when I was thinking clearly?

The next generation of investors will probably have better tools than any generation before them.

Better data.

Better AI.

Better access to research.

But those tools will increasingly be available to everyone.

The harder advantage may remain much more human: knowing what matters, knowing when not to act, recognizing when your thesis is wrong, understanding what is already priced in, and having a process strong enough to survive your own emotions.

The goal is not to remove emotion from investing. It is to prevent emotion from becoming the decision-maker.

Find the next theme before it becomes obvious

Macro Notes Premium is where I publish my deeper research on emerging market trends, investment theses, catalysts, risks, and the companies positioned to benefit.

Premium also includes access to Altis.finance Premium to explore these ideas, companies, markets, and indexes in more detail.

Join Macro Notes Premium — $500/year →

Some of these theses have led to very attractive opportunities. They shared a common starting point: a change was becoming visible in the economy, and we wanted to understand which companies could benefit.

But past opportunities do not answer the question that matters to us today: where could the next ones come from?

So I wanted to share five investment theses we will be following closely at Macro Notes. They are at different stages. Some are beginning to translate into industrial investment. Others have recently passed an important clinical milestone. All five give us concrete reasons to keep researching.

When I describe these themes as early, I mean that their markets are still developing. That does not mean nobody knows about them, or that the stocks involved are necessarily cheap.

We are putting them on our active watchlist. We still need to see more before we consider a position justified in our indexes. For some, that means orders and better visibility into margins. For others, stronger clinical results. In every case, we will also need to find a company and a valuation that make the opportunity attractive for shareholders.

For each thesis, you will find a measure of what could change if it works, the developments beginning to support it, and the next milestones we want to see. We distinguish reported figures, third-party forecasts, and our own illustrative scenarios throughout.

This entire edition is free to read. I want you to understand our reasoning, the evidence behind it, and the risks we see.

Premium and Founding members will then receive our deeper research, updates when we take positions in our indexes, and explanations of any changes to our strategy. As we publish the relevant indexes, members will be able to follow those decisions on altis.finance.

An interesting idea is a starting point. The work that follows is deciding what to do with it, at what price, and when to change your mind.

1. Navigating When GPS Becomes Unreliable

Imagine a drone, a ship, or an aircraft suddenly losing access to a usable navigation signal.

For some operators, this is already a daily operational problem. On September 24, Reuters reported that Finnair had adapted its pilot training to GPS interference in the Baltic region. Aircraft have backup systems, but the frequency of the disruption adds another operational challenge.

That brings us to a simple question: which companies make it possible to keep navigating in these conditions?

The thesis focuses on inertial navigation systems, sensors, and equipment that make vehicles and other platforms less dependent on satellite signals. Established technologies and emerging quantum approaches will develop on different timelines.

Another signal came in June 2026, when EuroHPC launched a program focused on quantum sensors for navigation in environments where satellite signals are unavailable or disrupted.

The potential: protecting a function that supports an ecosystem worth hundreds of billions of euros.

In its May 2026 report, the EU Agency for the Space Programme, EUSPA, estimated global revenues from GNSS equipment and services at around €300 billion in 2024, with a forecast of €580 billion by 2034. It also expects the installed base of GNSS-enabled devices to reach almost 10 billion by then.

Those figures include a wide range of consumer and automotive applications. They are not the market size for navigation that can withstand jamming. They show how dependent we have become on this infrastructure, including in applications where reliability is critical.

If disruption persists, the ability to operate without a reliable satellite signal could become a much more important purchasing requirement for drones, ships, and certain aviation systems. That could create a cycle of new equipment purchases and upgrades to existing fleets, with suppliers integrated into their customers’ platforms for years.

What we need to see before taking a position: production orders following trials, repeat purchases, and a revenue contribution large enough to influence the supplier’s results. For quantum approaches, we also want performance demonstrated in real operating conditions and costs that make sense for the intended market. We will then assess how much of that growth is already reflected in the share price.

The risk is paying today for a market that takes much longer to develop. A promising technology still needs customers, and a small business line may have little impact on a large group’s earnings.

2. The Suppliers Behind Nuclear Medicine

Treating cancer with a radiopharmaceutical requires a drug that can deliver a radioactive substance to its target. It also requires the ability to produce that substance and deliver the treatment under the right conditions.

That industrial supply chain interests us as much as the drugs themselves.

In May 2026, Novartis presented new data on its radioligand therapies, including early results for an actinium-based treatment. These results do not guarantee success, but they strengthen the case for further research into this approach.

Manufacturing capacity is developing too. Eckert & Ziegler and DC Pharma recently announced the opening of an isotope production facility in China. A second phase is planned to include actinium-225 production.

Our thesis is that some suppliers could benefit from the development of several treatments by providing the isotopes, equipment, or manufacturing capacity they require.

The potential: supplying the expansion of a business that already generates billions of dollars.

In 2025, Pluvicto generated $1.994 billion in sales, while Lutathera generated $816 million, or $2.81 billion combined. Pluvicto’s sales grew 43% in US dollars that year.

In its November 2025 investor presentation, Novartis estimated Pluvicto’s peak annual sales potential at more than $5 billion, including marketed indications and expected future indications. That is more than 2.5 times its 2025 sales. This is the company’s estimate, dependent on clinical and commercial progress, with no guaranteed year for reaching that level.

If achieved, that would mean more than $3 billion in additional annual sales versus 2025 from a single treatment. The development of other radioligand therapies could expand industrial demand further.

Isotope suppliers would capture only part of that value. But more patients and treatments can mean more doses to manufacture, more qualified capacity, and more deliveries to organize. This is where we are looking for businesses with room to grow substantially.

What we need to see before taking a position: contracts translating into delivered volumes, approved and operational capacity, and evidence that higher production improves margins. We also want to understand each supplier’s dependence on individual drugs and make sure the share price leaves room for error.

There are two main risks: clinical trials may disappoint, and suppliers may build too much capacity. A supply constraint today does not guarantee lasting pricing power.

3. After Weight Loss, Preserving Muscle

GLP-1 treatments have opened up a huge field of research in obesity care. We are now interested in a related question: how can the quality of weight loss improve?

Lean mass can decline during weight loss. Lean mass is not exclusively muscle, and preserving it does not automatically demonstrate better strength or greater independence in daily life.

That is the next step we want to watch.

A Phase 2 study published in Nature Medicine in 2026 provided proof of concept for apitegromab in preserving lean mass during treatment with tirzepatide. The study remains preliminary and has methodological limitations, but it offers a direction worth investigating.

If an additional treatment demonstrates a meaningful enough benefit, it could find a place alongside obesity medications. It also connects with a broader question about aging: how to preserve physical function.

The potential: building a new treatment category around an already substantial market.

In an analysis published in 2026, Morgan Stanley estimated that the global market for GLP-1 treatments for obesity and type 2 diabetes could reach $190 billion by 2035, compared with $79 billion in 2025. That includes both conditions; it is not an estimate of the market for muscle preservation.

The relevance to our thesis is the growing number of patients receiving treatment. An effective add-on therapy would not need to be prescribed to everyone to become a significant business.

Consider a purely illustrative scenario, at some point after a possible approval: one million patients receiving an additional treatment that generates $3,000 in net annual revenue per patient would represent $3 billion in annual sales. At three million patients, the figure would be $9 billion. Both the price and patient numbers are assumptions for this calculation, not market data or an adoption forecast.

This gives a sense of the opportunity. Even targeted adoption could support a multibillion-dollar drug category if the clinical benefit and funding are there. The impact could be particularly meaningful if treatment also helps preserve mobility or independence, which remains to be demonstrated.

What we need to see before taking a position: results confirmed in larger trials, a clinically meaningful benefit beyond a lean-mass measurement, acceptable safety, and a credible path to approval and reimbursement. We will also assess developers’ funding needs and the risk of shareholder dilution.

The main risk is that an interesting biological result may not justify the cost and burden of another treatment. We want to understand the benefit to patients before building the economic case.

4. Mineral Independence Starts Beyond the Mine

Extracting a resource is only one part of securing a supply chain. It also needs to be separated, refined, and transformed into a usable material.

In rare earths, we are particularly interested in these intermediate stages, as well as recycling.

In July 2026, USA Rare Earth finalized an agreement to acquire a stake in France’s Carester. The transaction supports the development of rare earth separation and magnet recycling capacity in Lacq.

What interests us is the transition from a political objective to an industrial operation: suppliers, factories, and supply agreements.

The thesis is that supply security could support new capacity and customer commitments lasting several years.

The potential: building several times the capacity currently available outside China.

In its 2026 rare earths report, the IEA projects that demand for magnet rare earths outside China will increase by 50% by 2035 under current policy settings.

To meet all of that demand through diversified supply, its analysis indicates that refining capacity would need to grow roughly fourfold, and magnet manufacturing capacity roughly sixfold, relative to existing capacity including planned expansions. The report distinguishes this base from the broader pipeline of announced new projects, which could close part of the gap. These are requirements under a full-coverage scenario, not a forecast of what will actually be built.

The potential opportunity therefore extends beyond growth in consumption. It includes establishing industrial capabilities in new regions. A company that secures both its inputs and its customers could find a place in a supply chain that develops over many years.

The economic importance is striking relative to the volume of material involved. The IEA estimates that full implementation of the export restrictions it examines would put $6.5 trillion in annual downstream production outside China at risk. That measures industrial exposure under that scenario; it is not the revenue available to rare earth producers.

What we need to see before taking a position: enough financing to reach production, purchase agreements with credible customers, competitive costs, and initial deliveries that meet industrial specifications. The ability to protect margins if material prices fall will be important.

The risks remain significant: delays, additional financing needs, shareholder dilution, and competition from lower-cost producers. A country can need a factory without its shareholders earning an attractive return.

5. Replacing Organs: Another Approach to Longevity

Longevity is often discussed in terms of slowing the aging process. Another approach is to improve our ability to replace organs that no longer work.

Xenotransplantation, including the use of organs from genetically modified pigs, is one of the areas we will be following.

United Therapeutics announced the first transplant in its EXPAND kidney trial in November 2025. In May 2026, it also announced clearance to begin a clinical trial of its UHeart cardiac program.

These milestones allow the approach to be tested in clinical trials. They do not yet establish that it can become a widely available treatment.

This is the most experimental of the five theses in this edition. Its potential rests on the possibility of sustainably increasing the supply of transplantable organs.

The potential: adding tens of thousands of transplants to a system constrained by organ shortages.

As of July 30, 2026, US data recorded 95,492 people in the kidney waiting-list category, while 27,574 kidney transplants, excluding the separate kidney-pancreas category, had been performed in 2025. The waiting list is a snapshot, while transplants are an annual flow. Comparing them illustrates the shortage; it does not directly measure waiting times.

In an illustrative scenario, an additional 10,000 kidney transplants per year would represent an increase of roughly 36%over that annual volume. An additional 25,000 would bring the total close to double. This assumes genuinely additional transplants, eligible patients, and sufficient hospital capacity. It is not a deployment forecast.

If the approach becomes safe and reproducible, the impact could be very tangible: more patients receiving transplants and, for those with a lasting functioning graft, the possibility of coming off dialysis. On the industrial side, this could create a business around controlled biological production, with demanding qualification requirements and long-term monitoring.

There is also a new development in manufacturing: in August 2026, United Therapeutics opened a facility intended to produce xenografts for clinical trials and potential commercial use.

What we need to see before taking a position: durable graft survival across multiple patients, control of rejection and infections, reproducible outcomes, and a credible regulatory pathway. We will then need to understand production costs, reimbursement, and transplant centers’ ability to perform these procedures. We will separate the value of companies’ existing businesses from the much more uncertain value of their xenotransplantation programs.

Clinical risk is substantial, timelines may be long, and the ability to scale remains uncertain. Knowing when to wait will be part of the work.

How We Will Follow These Five Theses

These five themes form a research agenda. We will look for the companies best positioned to benefit, study their results, and examine what their valuations already assume.

Each opportunity is significant enough to deserve close attention. Each still has specific conditions to meet before we consider an investment justified.

The important transition will be from expected results to observed results: a trial confirming a clinical benefit, a factory delivering its product, a customer renewing a contract, or margins beginning to reflect growth. Our aim is to determine when the evidence and the price make the remaining uncertainty acceptable.

We can be right about a market’s development and still choose the wrong stock. We can also identify a good company at a price that leaves too little room for error.

That is why we will follow both the developments that strengthen each thesis and those that would force us to reconsider it.

When we decide to take a position in an index, Premium and Founding members will receive our analysis: why we chose that exposure, the strategy behind it, the risks, and what we will watch next. The indexes on altis.finance will let you review the documented positions and follow their evolution.

Our horizon is the medium to long term. We study changes that can unfold over several years, while remaining disciplined enough to adjust our approach along the way.

These five theses will be part of that work. I look forward to sharing it with you.

Follow What Comes Next with Macro Notes Premium

If you read Macro Notes for free, this edition gives you a preview of the themes we will explore over the coming months.

As a Premium member, you can follow the work behind these ideas: the research, the strategies, our decisions, and the reasons we adjust them.

We are still waiting for further confirmation before taking positions around these five theses. Premium coverage will help you understand the developments that bring an idea closer to an investment decision, those that make us wait, and those that lead us to drop it. When we decide to act in an index, you will receive the reasoning behind that decision.

Your membership includes Premium editions and the archive, in-depth analysis of the companies and theses we follow, and updates when we take positions or change strategies in our indexes.

It also includes Altis Premium access on altis.finance and access to more than 100 indexes from Macro Notes and Altis Research, where you can explore investment theses and follow their evolution as markets move. You can review charts, holdings, and documented trades. These indexes are simulated tracking tools; their performance does not represent returns from a live brokerage account.

Premium Monthly — experience our work for a month. Start here to read the research, explore the indexes, and see how our approach fits into your own investment process. Your subscription renews monthly until canceled.

Premium Annual — $500 a year, to follow the theses over time. This is the plan that best fits our approach. Investment decisions develop through the earnings reports, contracts, and events that shape a thesis over time.

Founding — $1,200 a year, to go further with your own research. You receive everything included in Premium, plus Premium access to more than 10 other Substack newsletters partnered with Macro Notes, including Public Markets, Tech Economics, and Future Insider, all included in your membership.

Founding also includes 10,000 agent credits per year to use the AI research tools on altis.finance, explore the themes that interest you, and develop your own investment ideas. It is also a way to provide additional support for the development of Macro Notes and Altis.

You can join today, explore the research already available, and receive our upcoming analysis of these five themes.

Explore Macro Notes Memberships

Instead of trying to forecast where markets are heading, his strategy systematically reacts to what markets are already doing.

I thought it was worth sharing because it offers a very different answer to a question every investor eventually faces: how much of investing should actually depend on our judgment?

Outperforming the market in the long run is a goal many of us share. We have different strategies, different mindsets, and different perspectives. But for most of us, the ultimate objective is the same.

I spent 15 years investigating and experimenting with all sorts of strategies. I read all the books and studied all the academic research.

Eventually, I settled on a simple, rules-based strategy that beat the S&P 500 in 22 of the last 26 years.

I’m an engineer from Denmark turned full-time investor and writer. I quit my job at 41 thanks to the returns from my strategy. Now I share my approach and my trades at The MarketFighter Strategy on Substack.

I was invited to write this guest post for you about how it works and how simple systematic investing can be the key to higher returns.

Why this approach is different

Before I explain how a rules-based approach can work, let me just briefly show the returns my strategy has delivered (measured in USD) since I started trading it actively in March 2021:

These results were achieved entirely through trading ETFs once a month (or less), spending approximately 15 minutes each time.

Investing in stocks usually involves some or all of the following:

• Political analysis

• Global macro analysis

• Individual business analysis

• Earnings analysis and company valuation

• Technical analysis and chart reading

• Forecasting

The systematic approach I finally settled on deals with none of the above. Instead, it focuses entirely on a set of predefined rules, rooted in academic research.

Traditional market analysis relies on investors being able to predict or anticipate future events that the market hasn’t priced in yet. Unfortunately, as you may know, all research shows that incredibly few are able to consistently succeed in this endeavor.

Even when you get the thesis right, the timing and execution has to be right as well. In the long run, you have to repeat this process, but with new trends, new markets, and new companies.

When I finally realized my chances of succeeding this way were minimal, I gave in to a humble investing mindset around six years ago.

This is a counter-intuitive philosophy that does not allow for any attempts to predict the future or be smarter than the market. I accept that I simply don’t know what will happen.

It’s the same way most index investors think. Index investing is the simplest form of systematic investing and the easiest way to apply the humble investing mindset.

But listen now: Most people stop here, which is a shame. Because there are so many options for optimizing the index approach with simple rules that most people never get exposed to.

Different results require a different approach

Although market-cap weighted indices like the S&P 500 compound at roughly 10% annually in the long run, the average investor only gains around 3% according to data.

None of us think we are average investors. Yet, all research shows it’s an incredibly small number of people on this planet who are able to somewhat consistently beat the market in the long run, without luck playing the major role.

Overconfidence, fear and greed, FOMO, recency bias, and all sorts of other human emotions play a much larger role than we like to think. All of these human instincts were designed for us to survive on the savannah. Not for doing well on financial markets.

Knowing how bad our chances of beating the market are and how most people never succeed, I came to an important conclusion: If you want a different result, you need a different process. I needed to find a different way to approach the market.

Now let me give you a brief idea of my journey towards this.

How my journey shaped my strategy

I started investing with a small amount of money when I was 18. I got completely hooked, but like most beginners I was overconfident and had no idea what I was doing.

I took some hits, and got humbled by the market. I was not the natural talent I thought I would be, but I kept learning from my mistakes, and my eagerness to succeed in the market only grew with the effort I put into learning and growing.

I’m a software engineer. I spent 15 years developing complex software applications for a living. There’s no doubt this background has shaped the way I approached the market and eventually ended up trading it.

You can find opinions and advice pointing in all directions, but if you apply a more rational and scientific approach to optimizing returns, there’s really only one thing that matters, if you ask me:

What are the most reliable and consistent factors for outperformance in the markets, historically?

We all know that past performance is never a guarantee of future returns. But if something has worked well across multiple decades and business cycles, history shows it has a high probability of continuing to do so.

Utilizing persistent market anomalies

The factors I’m talking about are based on persistent market anomalies resulting from human behavior. I’m sure you’ve heard of the most well-known examples:

• Value

• Momentum

• Quality

• Size

These are all quantitative and rules-based factors, and if you had bought stocks filtered by these over the past century, you would have significantly outperformed the market.

The problem with these factors is that each of them also exposes investors to long multi-year periods of underperformance from time to time.

That’s where strategic asset allocation and rotational strategies become interesting. My research led me to several other sources of more persistent market patterns that turned out to work with relatively high consistency.

Eventually, I spent months trying to put together all these pieces into a system I could easily trade for myself and my family.

Enter rules-based investing

When you invest systematically (rules-based), the core idea is that every trade you execute is entirely based on predefined rules. No personal opinions, no second-guessing, no emotions or discretionary analysis.

It’s the same approach many successful hedge funds apply.

But in my case, I wanted a long-term system that didn’t require my attention on a daily basis. I wanted a strategy with a high probability of long-term outperformance without being riskier than a market-wide index fund.

The system I settled on

Unlike most systematic strategies that focus on individual stock picking (and mostly apply to active traders), I chose to go for an optimized version of index investing, following a more passive approach.

I built a system based on the best ingredients I found in research: Quantitative factors, relative momentum, sector rotation, and defensive crisis rules.

Instead of trying to predict future events in the world, it reacts to market price trends regardless of what caused them.

Focusing on historical consistency and robustness in returns, I defined a universe of 15 ETFs divided into two baskets: One for factor-focused ETFs and one for sectors.

Then I set up rules to check the relative momentum in each basket once a month. I let the system pick the strongest one from each basket, based on these rules. Here’s an illustration of the concept:

I run this process once a month, when the closing prices of the last trading day are settled. Then I allocate my money 50/50 into the two selected ETFs. After that, I do nothing until the next month has passed.

On top of this setup, I apply a drawdown protection rule consisting of three simple checks, all based on momentum as well. If all of them apply, I exit the market and wait for the next monthly signal. That’s it.

Testing and results

The beauty of entirely rules-based strategies is that they can be back-tested by applying the same rules on historical data. This way you can simulate historical performance to see how well it has worked.

Not all backtesting is beautiful, though. It comes with a lot of pitfalls, and simplicity often beats a system with many parameters and inputs. My system reads one thing only: Monthly closing prices. Nothing else.

I did the backtest on data from 2000 to 2020. Since then, I have traded it actively from 2021 until now. Here are the results by calendar year, measured in EUR and compared to the MSCI World Index (also in EUR):

The numbers are return percentages, and the 2026 number is a year-to-date figure as of closing prices on September 23.

I chose the MSCI World Index as my main benchmark, because my strategy involves both US and European stocks. But as I mentioned in the beginning, a comparison to the S&P 500 (and measuring everything in USD) shows outperformance in 22 of the last 26 years.

Now you’re probably thinking: So, what’s the catch? Here it is: Executing a strategy like this requires a specific mindset. You can’t succeed with a systematic strategy if you’re not ready to follow the rules, no matter what happens. This demands strict discipline.

It may sound easy, but if you have strong opinions about the market and the system signals a trade that goes against these opinions, you may find it very hard to stick to the strategy. I’ve had to unlearn old habits and force my brain not to interfere.

My personal outcome

For me personally, this strategy has had a huge impact on my life. What works for one investor will not work for everyone, but I finally found my own way of conquering the market.

After five years of quietly trading this for myself and my family, I decided it was time to share it with others who might be interested.

Due to the returns I had achieved, I was able to quit my job earlier this year to live off the strategy and spend more time on my passion for investing as well as other things in life.

That’s when I started sharing my strategy on Substack. Once a month, before the markets open on the first trading day, I send out the trading signal my system generates. This reveals the two ETFs I will be holding for the coming month.

I strongly believe in transparency and open communication, which is why I share complete performance updates in public every month.

Thousands of investors are already following along, and I’m truly proud of what I have achieved just by approaching investing from a different perspective and taking the path less traveled.

If you found this post interesting and want to learn more about how it works, feel free to join our growing community at The MarketFighter Strategy.

Thanks for reading!

The MarketFighter Strategy

Join the simple ETF strategy that beat the S&P 500 in 22 of 26 years. Built on rules and evidence, not opinions.

Disclaimer: The MarketFighter Strategy is for educational and informational purposes only. It is not financial advice, and the author is not a licensed investment advisor. Investing in ETFs involves significant risk, and past performance is never a guarantee of future results. You are solely responsible for your own trades and financial outcomes. Read the full Disclaimer here.

And not because I suddenly became interested in aviation, but because a company whose mission is to bring commercial supersonic flights back has actually developed a very innovative solution to one of the biggest problems facing artificial intelligence infrastructure.

And that’s what I want to talk about today.

Something Elon Musk often says, and which is probably behind his ability to build such successful companies, is that the only laws that are truly real and impossible to change are the laws of physics.

Regulations and laws are very human things. They can be changed, adjusted, or removed.

Physics is the real rule of the game.

And SpaceX is probably the best example of this philosophy: trying to accomplish everything that physics allows. If something is not physically impossible, SpaceX will try to do it.

Boom Supersonic is pursuing a somewhat similar idea. The problem with supersonic travel is regulation and cost. Physics itself is not the problem. Boom’s bet was essentially to say:

If we can build a supersonic aircraft without the traditional sonic boom — which has been one of the main arguments behind restrictions on supersonic flight — then supersonic flights could eventually be allowed again.

And if we can do it at a much lower cost, then it can also become economically viable.

At that point, you can build a profitable business around it.

Since the 1960s, there has been almost no improvement in the speed of international air travel, even though international travelers place a very high value on time.

By offering a premium product to customers who would normally fly business class, you potentially have a market.

The idea is to target exactly this customer.

A normal economy flight remains long.

A premium flight could eventually become supersonic.

If someone is already willing to pay several thousand dollars to fly from London to New York, Boom wants to offer a product where the value proposition is no longer simply a seat that turns into a bed, but potentially three hours saved.

New York to London could, for example, fall from around seven hours to roughly three and a half.

And that allows Boom to target a very different market from a Boeing 787 or Airbus A350.

A 787 has to spread the economics of the flight across more than 200 passengers.

Boom essentially wants to operate a:

64–80 seat “shared business jet.”

The small number of seats is both an advantage and a problem.

The advantage is that filling 70 premium seats is easier than filling 250 premium seats.

The problem is that all of the aircraft’s costs are spread across only around 70 passengers.

And physically, flying almost twice as fast is expensive.

Of course, for this model to work, the cost per seat becomes extremely important.

You need to offer a price customers are willing to pay.

Otherwise, they will simply continue flying regular business class, spend a few more hours on the plane, and save a few thousand dollars.

At supersonic speeds, you notably face wave drag, the drag created by shock waves.

You need more thrust.

The engines cannot use the enormous bypass ratios found in modern turbofans optimized for Mach 0.85.

Temperatures are higher, structures face more stress, and the aerodynamics become much more demanding.

That is precisely one of the reasons Concorde consumed so much fuel and was never really economically viable.

But we are now more than 50 years later, and a lot of technological progress has been made since then.

And one of the challenges for a new aircraft manufacturer trying to compete with Boeing or Airbus and control its costs is to build its own engine.

So to understand why Boom’s Symphony engine was interesting, I first had to understand what a medium-bypass turbofan actually is.

The bypass ratio measures how much air passes around the engine core compared with how much air passes through the core.

A Boeing 787 or Airbus A350 engine has a high bypass ratio.

That is excellent for fuel efficiency at around Mach 0.85, because it is more efficient to accelerate a large amount of air moderately than a small amount of air very strongly.

But at Mach 1.7, having an enormous fan becomes problematic: diameter, drag, airflow speed, aerodynamic constraints.

Symphony therefore needs a lower or medium bypass ratio compared with a modern long-haul aircraft engine.

It sits somewhere between: a very “hot” turbojet optimized for speed and a large ultra-efficient turbofan optimized for Mach 0.85.

It is a compromise between: supersonic speed and acceptable fuel consumption.

But probably the most important difference between Symphony and Concorde’s engines is the afterburner.

An afterburner, or reheat, consists of injecting fuel after the turbine, directly into the exhaust gases while they are still extremely hot.

Very simply: compressor → normal combustion → turbine → additional fuel injection → enormous hot exhaust jet

This produces a very large increase in thrust.

It is spectacular and is widely used on fighter jets.

But it is also extremely inefficient in terms of fuel consumption.

You get much more thrust, but you burn a huge amount of extra fuel to get it.

Concorde used afterburning — known as reheat — particularly during takeoff and while passing through the transonic region around Mach 1.

That was one of the reasons Concorde consumed so much fuel.

Boom wants Symphony to be able to do: takeoff → acceleration → pass Mach 1 → Mach 1.7 without afterburning.

So instead of getting additional power by essentially “burning a huge amount of fuel behind the engine,” Boom wants the engine itself to be efficient enough to provide the required thrust.

Economically, this is extremely important.

To simplify:

Concorde: engine + afterburner → enormous thrust → enormous fuel consumption.

Overture / Symphony: optimized turbofan → enough thrust without afterburning → better efficiency.

That does not mean Overture — Boom Supersonic’s aircraft — will be as efficient as a 787.

It probably won’t be on a per-seat basis.

But Boom is trying to reduce the difference enough to make supersonic travel commercially viable, which was much more difficult with Concorde’s architecture.

And of course, there are many other technological elements that make the model potentially viable.

The goal is to bring the price of a supersonic flight closer to that of a long-haul business-class flight, while also bringing the cost per seat closer to something economically sustainable.

Every small technological improvement Boom makes can have a huge impact on cost and on the potential of the model.

And that is also why airlines have already placed a significant number of pre-orders.

For them, if the model becomes viable, the question then becomes whether they want to be among the first airlines to launch a supersonic route under their brand rather than wait for a competitor to do it first.

But this is not actually the main subject of this edition.

As I said at the very beginning, what I really want to talk about is AI infrastructure.

Boom created its own engine.

And building an engine adapted to supersonic flight is one of its main competitive advantages in aviation.

Because, as we understood earlier, building a supersonic aircraft is not simply about making an aircraft more aerodynamic.

You also need to build an engine specifically adapted to those speeds.

A turbine engine can be simplified like this:

AIR → COMPRESSION → COMBUSTION → TURBINE

The compressor pulls air into the engine and greatly increases its pressure.

Fuel is then injected into this compressed air.

It burns inside the combustion chamber.

This creates extremely hot, high-pressure gas.

That gas then expands through a turbine.

The turbine spins.

And that rotation creates mechanical energy.

In Symphony, the high-pressure core is mainly made up of: High-Pressure Compressor → Combustor → High-Pressure Turbine

Boom itself describes the engine core around these three fundamental components.

The interesting part is that combustion is not directly what pushes the aircraft forward.

It first creates energy in the gas.

The engine then decides how that energy will be used.

Now imagine taking essentially that same engine core, but instead of using it to move an aircraft forward, you want to use it to produce electricity.

You keep: Air → Compressor → Combustion → Turbine

But instead of using the remaining energy to move the aircraft, you add another turbine designed to recover this mechanical energy.

That turbine is connected to a shaft.

The shaft turns a generator.

And the generator produces electricity.

Boom has created a similar solution called Superpower.

Of course, Superpower is not simply an aircraft engine.

I went through this explanation so that you can understand just how closely related the underlying technologies are.

And this is why it actually makes sense for a company like Boom Supersonic to develop something like this.

They have not suddenly changed industries.

The core of their business is still building the next generation of commercial supersonic aircraft.

But Symphony — the aircraft engine — and Superpower — the system designed to generate electricity — share around 80% of their hardware.

They also share the same high-pressure compressor and high-pressure turbine.

The low spool is slightly modified.

Where Symphony uses its large titanium fan to produce thrust, Superpower adds two additional compression stages and then a three-stage free power turbine, connected to an independent electrical generator.

The fuel injectors are also different:

Jet A / SAF for the aircraft,

natural gas for Superpower.

And of course, the reason Boom is now offering a solution to generate electricity is AI data centers.

As you already know if you have been reading Macro Notes, demand for data centers has exploded with the emergence of AI.

But at this point, you might ask why anyone would bring aircraft technology into this market when turbines capable of doing something similar have already existed for years.

GE Vernova, Siemens Energy, Mitsubishi, Baker Hughes and others obviously did not wait for Boom.

But aircraft engines have several interesting characteristics.

They have to produce enormous amounts of power in very little space and with very little weight.

That creates very high power density.

They also need to withstand extremely high temperatures.

They need to be able to change their power output quickly.

And aerospace technology has been pushed extremely far in areas such as materials, compressor aerodynamics, blade cooling and control systems.

That is actually why an entire category of ground-based turbines already exists:

aeroderivative gas turbines.

GE Vernova is probably the best example.

Its LM2500 comes from aircraft-engine technology and produces around 35 MW in its LM2500XPRESS version.

Crusoe has already ordered 29 units, representing close to 1 GW of power capacity for its data centers.

Each package arrives around 95% factory assembled and can start in roughly five minutes.

So: Boom is not inventing the concept of using aircraft-engine technology to generate electricity.

This is already an existing industry.

What Boom is trying to do is build a new generation of aeroderivative turbine around an engine architecture it was already developing for Overture.

And to simplify, Superpower is not really trying to create the cheapest electricity possible.

There are other solutions that are much more interesting if your only objective is to minimize the cost of electricity.

Instead, Boom is trying to solve another problem that has become critical for AI:

getting a very large amount of MW, very quickly, directly where the data center needs them.

I think this is the most important part of this edition.

A traditional data center may consume a few dozen MW.

The new generation of AI campuses can require hundreds of MW, and in some cases more than 1 GW.

And most importantly, this demand is concentrated in one location.

That is very different from 1 GW of electricity demand spread across millions of households.

Here, you can have a single campus saying:

“I need another 500 MW here, in this exact location, ideally within the next two years.”

The IEA estimates that global data-center electricity consumption could rise from around 485 TWh in 2025 to roughly 950 TWh by 2030, while electricity consumption from data centers specifically related to AI could roughly triple.

In the United States, data centers could represent close to half of total electricity-demand growth through 2030.

So the problem is not simply:

“The world needs to produce more electricity.”

It is more specifically:

“We need to deliver enormous amounts of electricity to a few very specific locations, very quickly.”

And that is much harder.

Of course, this electricity can be more expensive.

But that is simply part of the supply-and-demand equation.

If you have already spent billions of dollars building an AI campus, installing GPUs, networking equipment and cooling systems, getting access to power quickly can be far more important than getting the absolute cheapest possible electricity.

There are also several other interesting points.

The natural gas used by these systems is already widely available in many of the regions where data centers are being built.

And gas can provide continuous power, which is obviously important for data centers designed to operate 24 hours a day.

Superpower is also relatively compact.

And aircraft-engine technology is naturally designed to operate in very high-temperature environments, which can also be useful for power generation.

Boom claims that Superpower can maintain its full 42 MW output at temperatures above 110°F, or roughly 43°C, without using water.

In short, what I found extremely interesting while researching Boom is that it gives us a way to understand two different subjects at the same time.

On one side, the return of supersonic travel and the challenges involved in making high-speed commercial aviation economically viable.

But on the other side, it shows how a startup building aircraft engines can end up developing a product that directly addresses one of the biggest emerging problems in AI infrastructure.

And for Boom itself, this potentially makes the business model much more viable.

Overture has not launched yet.

Building a commercial aircraft takes time.

Regulation has also started evolving as Boom works on solutions around the sonic-boom problem.

But aviation obviously remains a highly regulated industry — and fortunately so.

For a company like Boom, having a potentially very profitable secondary market for a large part of the technology it is already developing is therefore very good news.

The Premium Edition: The Next AI Infrastructure Trade

Today’s edition is free and open to everyone.

But while researching Boom, we also started building a new index on Altis.finance around the larger investment thesis behind this story.

Because the interesting part is not simply Boom.

Boom is private.

The bigger opportunity is the infrastructure being built around the same problem:

AI companies are able to deploy compute faster than the electric grid can deliver the power required to run it.

And that opens a completely new layer of the AI infrastructure trade.

In the Premium edition coming in the next few days, I’ll go much deeper into this thesis and share the full index we built around it.

We’ll look at the public companies positioned across the entire private-power stack — from gas turbines and distributed generation to electrical equipment, power distribution, cooling and the infrastructure required to bring those megawatts all the way to the GPU.

I’ll also share:

• the full Altis.finance index built around the thesis;

• the trade history and performance behind it;

• the companies already in the portfolio and why each position is there;

• the new trades and positions we are currently considering;

• which parts of the value chain I think have the strongest exposure to time-to-power;

• and the major risks that could invalidate the thesis.

The broader idea is simple:

For the first phase of the AI boom, everyone focused on the chips.

Then came networking, memory, optics and cooling.

Now the bottleneck is moving again.

And this time, it may be moving all the way into power generation itself.

Macro Notes Premium and Founding members will receive the full research and portfolio update in the next few days.

Go Deeper With Macro Notes

Macro Notes is built around a simple idea:

identify structural changes early, understand where the bottlenecks are moving, and find the companies positioned to benefit before those themes become obvious.

The free newsletter will always remain an important part of what we do.

But most of the work happens after the initial idea.

That is what Macro Notes Premium and Founding are built for.

Macro Notes Premium — $500/year

Premium is where we turn the themes discussed in the newsletter into actual investment research.

Premium members get access to:

Full Premium research

Deeper investment theses, company analysis, industry mapping, and follow-up research on the themes we introduce in the free newsletter.

Our Altis.finance indexes

Instead of publishing a thesis once and forgetting about it, we track many of our ideas through live indexes on Altis.finance.

You can see the companies behind each thesis, the trades we record, how the thesis evolves over time, and the performance of the index.

Portfolio updates and new positions

When we add a new company, reduce a position, identify a new part of the value chain, or materially change our view on a thesis, Premium members get the update.

Access to Altis.finance Premium

Macro Notes Premium includes access to the research tools and indexes we publish on Altis.finance.

The goal is to move beyond simply reading research.

You can actually follow how an investment thesis develops over time.

Macro Notes Founding — $1,200/year

Founding includes everything in Macro Notes Premium.

But it is designed for readers who want to go much further and use the same research infrastructure we are building internally.

The biggest difference is Altis.finance Agents and Radar.

Founding members receive:

10,000 Altis Agent & Radar credits every year

These credits can be used to research companies, industries, emerging investment themes and, increasingly, to turn an idea directly into a structured investment thesis.

And this is becoming one of the most important parts of Altis.

Introducing the New Altis.finance Radar

We originally built Radar to help answer a difficult question:

What investment themes are beginning to emerge around a specific trend, company, technology or event?

But we are taking the product much further.

The new version of Radar doesn’t simply return a list of companies.

It can research the underlying theme and begin building complete draft indexes around the opportunities it identifies.

Imagine asking:

What are the second-order beneficiaries of the electricity bottleneck created by AI data centers?

Radar first researches the problem.

It looks for the different investment theses hidden underneath it.

For example, it might identify: on-site gas generation, electrical equipment, grid bottlenecks, data-center cooling, natural-gas infrastructure or other adjacent opportunities.

But instead of stopping there, Radar can turn those themes into draft indexes.

For each thesis, it can generate:

A name and clear investment thesis

An explanation of exactly what structural change the index is trying to capture.

Suggested companies

The public companies with the strongest exposure to that specific thesis.

Suggested trades

Potential entries and changes based on the development of the thesis.

Historical event-based trades

Radar can research when important events around the thesis actually occurred and reconstruct how an index could have evolved through those events, while checking historical dates and prices.

A generated performance chart

So you can immediately visualize how the thesis would have behaved historically.

And Radar can present different ways of approaching the same opportunity.

For example:

Early

Companies positioned before the thesis is broadly reflected in fundamentals.

Balanced

A combination of established beneficiaries and earlier-stage exposure.

Late

Companies where the theme is already more visible in revenue, orders or earnings.

The result is not simply an AI answer.

It becomes a draft investment framework that you can research, edit and eventually turn into your own Altis index.

From a Question to an Investment Thesis

This is the direction we want Altis to move toward.

Today, doing serious thematic research normally means jumping between:

Google, earnings calls, company filings, news, spreadsheets, charts, and portfolio software.

Radar is being built to compress that process.

You start with a question:

“Who benefits if AI data centers increasingly generate their own electricity?”

Radar researches the theme. ↓

Identifies the bottlenecks. ↓

Maps the value chain. ↓

Finds the companies exposed to each part. ↓

Researches the important historical events. ↓

Builds potential indexes. ↓

Suggests trades. ↓

And gives you a structured starting point for deeper research.

The objective is not for AI to make the investment decision for you.

It is to dramatically reduce the amount of time required to go from:

“This trend looks interesting.”

to:

“I understand the thesis, the value chain, the public companies exposed to it, and how I could track it.”

That is where I think Radar becomes genuinely valuable.

Premium or Founding?

If you mainly want to read our research, follow our investment theses and access the indexes we build, Macro Notes Premium is the right plan.

Macro Notes Premium — $500/year

If you want everything in Premium but also want to use Altis as your own research platform, explore investment ideas with Agents, and use the new Radar to discover and build theses yourself, Founding is designed for that.

Macro Notes Founding — $1,200/year

Founding includes 10,000 Agent & Radar credits every year, alongside full access to Macro Notes Premium and Altis.finance.

We are increasingly building Macro Notes and Altis as two parts of the same product:

Macro Notes helps you discover the themes we are researching.

Altis helps you follow them, test them, and increasingly discover new ones yourself.

Join Macro Notes Premium or Founding

Macro Notes is becoming one of the largest macro and finance newsletters on Substack.

We’re now close to reaching 1,000 Premium subscribers and 100 Founding members — a milestone I expect us to hit by the end of the year, in line with the goals we set for 2026.

On the readership side, we recently removed around 5,000 inactive accounts, which temporarily brought our subscriber count down to roughly 55,000. In just three months, we gained those 5,000 subscribers back through entirely organic growth, bringing us back to around 60,000 readers.

This has also helped us reach an average 30% open rate, which is a particularly strong level for a newsletter of this size and makes the audience very attractive to advertisers.

One of our recent partnerships was with Coinrabbit. Here is the dedicated edition we produced in collaboration with them:

If you’d also like to advertise on Macro Notes and get exposure to a highly qualified investor audience, contact me at pierre@altis.finance to discuss a tailored partnership.

A dedicated edition presenting your company, written in the style of a regular Macro Notes issue, can be sent to our 60,000 readers.

This is far more effective than a simple ad placement, because this type of edition creates much deeper engagement with your brand and helps build trust directly with the audience.

Of course, we carefully select every partnership. Our goal is not to run a large number of sponsored campaigns, but to work with a small number of companies that are genuinely relevant and valuable to our readers.

The price of Macro Notes Premium has now permanently increased to $500 per year to reflect the additional value we’ve added, including access to Altis.finance and our full archive of past research.

The price will not increase again and will remain at $500 per year for the Premium plan.

Of course, we’ll continue improving Altis.finance over time — expanding the indexes included with your subscription, including those published by Macro Notes and Altis Research, and building new tools to make the platform more complete and useful.

If you’re already a subscriber, the price you pay will never change. You’ll remain on the rate that was in place when you originally subscribed.

The launch of Altis.finance has received very positive feedback, even though the platform is still an MVP.

We also recently renamed Portfolios to Indexes.

Altis.finance is a platform that lets you follow our investment theses through indexes — portfolios of stocks built around a specific theme or thesis. You can track their performance over time and receive ongoing insights and updates on the indexes you follow.

The idea behind Altis.finance is simple: when we introduce an investment thesis in a Macro Notes edition, you now have a dedicated place to follow how that thesis evolves and how it performs against the market.

You can see how we are positioned, the trades we make over time, and the key developments affecting each thesis.

Macro Notes covers a very wide range of macro and investment themes, so it was never realistic to publish full updates on every single thesis directly in the newsletter. Some readers are mainly interested in AI infrastructure, others in financial themes, others in biotech, and so on.

With Altis.finance, you can now access more than 100 indexes that we publish and continuously update with the help of AI agents.

And most importantly, you can also create your own indexes and share them with your own subscribers — on Substack or elsewhere — by becoming a contributor for free.

Last week, more contributors have started sharing their portfolios on Altis.finance, including:

Aerospace Stock

Aerospace Stocks

I love aerospace stocks.

By Cade | Aerospace Stocks

He created and published Aerospace Compounders on Altis.finance—an index tracking the aerospace companies with the greatest potential to outperform over the long term.

You can explore and follow it here, and I encourage you to subscribe to his newsletter. Personally, I often find writers with a clear area of expertise, like Cade, far more interesting than those who cover investing in overly broad terms.

View Portfolios

Solo Capitalist

Solo Capitalist

High-conviction, concentrated investing. One deep-dive per week into the companies and sectors where a single position can define a portfolio.

He has published several portfolios on Altis.finance to track his investment ideas, focusing on global entrepreneurship and the rise of expats building online businesses in thriving cities around the world.

View Portfolios

Macro Engine

Macro Dashboard

A simple way to view macro: charts, indicators, flows

By Macro Engine

Macro Engine has shared his portfolio on Altis.finance to track how his macro convictions play out over time. It gives readers a practical way to follow the investment ideas behind his research.

Explore his portfolio on Altis.finance, and subscribe to his newsletter for the analysis behind it.

View Portfolios

That’s where Macro Notes stands today.

What started as a newsletter is gradually becoming something broader: a research platform, a growing network of contributors, and a place where investment theses can be followed over time rather than disappearing after a single edition.

There is still a lot we want to build — especially around Altis.finance, the indexes, and the tools available to Premium members.

As always, thank you for reading, subscribing, sharing the newsletter, and supporting the project.

We’re still early.

Team Macro Notes & Altis.finance

It is one of the easier themes for me to investigate because I spend a lot of time in Eastern Europe, not very far from the Ukrainian border. The consequences of what is happening in Ukraine are much more visible here than they are from London.

But after the drone and missile attacks that reached Dubai earlier this year, there was another defence theme I started paying much more attention to: counter-drone defense.

In March, drones reached the area around Dubai International Airport, while attacks across the Gulf disrupted airports, energy infrastructure and some of the busiest air routes in the world.

And the reason I find this theme particularly interesting is that it is fundamentally different from the European rearmament story we have discussed before.

This Is Not Just Another Rearmament Cycle

Europe’s defense problem is, in many ways, an enormous supply-and-demand imbalance. Russia represents a growing threat, European governments therefore need more ammunition, more missiles, more air-defense systems, more vehicles and more production capacity.

Governments increase their defense budgets, manufacturers receive longer contracts, factories expand, and companies across the defense supply chain benefit. We have discussed this dynamic extensively on Macro Notes.

But drones introduce something different. They change the economics of warfare itself.

For decades, when people talked about military drones, they were largely talking about unmanned aircraft. Think of something like a Predator: a large aircraft, long range, expensive sensors, sophisticated communications equipment, capable of remaining airborne for hours and striking high-value targets.

What is changing warfare today looks very different.

The modern battlefield increasingly involves relatively inexpensive drones manufactured at enormous scale. Many use relatively simple airframes, commercial electronics, navigation systems and inexpensive propulsion. They can fly low, they can be launched in large numbers, they can be mixed with decoys, and they don’t necessarily need to survive.

A drone worth tens of thousands of dollars can simply be sent in the direction of an infrastructure asset worth hundreds of millions.

That creates a very unusual economic equation.

Imagine spending $30,000 or $50,000 to launch a drone. The defender detects it, but the system available to destroy it was originally designed to intercept aircraft or sophisticated missiles. The interceptor can cost hundreds of thousands — sometimes millions — of dollars.

Even if the defender successfully destroys every drone, it can still lose economically.

This is the important part: the attacker doesn’t necessarily need to defeat the air-defense system. It can simply force that system to consume expensive interceptors faster than they can be replaced.

Ukraine has made this problem impossible to ignore. Russia is now fielding increasingly sophisticated drones at significant scale, including newer jet-powered systems that are harder to intercept, while Ukraine is simultaneously racing to develop cheaper interceptor drones and other countermeasures.

This is why I think the most important metric in air defense is gradually changing. It is no longer simply can you shoot it down? It is becoming how much does it cost you to shoot it down?

And that creates an entirely new defense market.

The Defense Stack Is Changing

Traditional air defense is not disappearing. You still need Patriot, advanced missiles, fighter aircraft and extremely sophisticated systems capable of intercepting ballistic and cruise missiles.

But using those systems against every $30,000 drone doesn’t make economic sense.

So another defense layer is being built underneath them: detection, small radar, radio-frequency sensors, electronic warfare, jamming, AI-powered identification, command-and-control software, low-cost kinetic interceptors, interceptor drones, programmable ammunition and, increasingly, directed-energy systems such as lasers.

The Congressional Budget Office recently described exactly this layered approach: radar and radio-frequency systems to detect small drones, followed by different methods to defeat them ranging from communications disruption to projectiles and directed energy.

This isn’t simply another product category inside the defense industry. It represents a new cost layer of modern air defense, and governments are beginning to spend accordingly.

$40 Billion

The event that made me revisit this thesis recently came from NATO.

On July 7, NATO allies announced that they intend to invest more than $40 billion in counter-drone capabilities over the next five years. NATO is also creating a dedicated counter-drone marketplace designed to make NATO-tested and NATO-compatible systems easier for member countries to purchase at scale.

I think the marketplace is almost as important as the headline number because one of the biggest problems in defense technology has historically been the gap between developing an interesting technology and actually getting governments to buy it at scale.

A marketplace designed around tested, interoperable systems potentially shortens that path: prototype, testing, approval, procurement, production, deployment.

That is exactly the transition I want to follow.

We created a portfolio around this thesis. Before the new trades we are making today, our Counter-Drone Defense Indexhad returned +66.1% on contributed capital.

After today’s additions, that figure moves closer to +45%. This does not reflect a deterioration in performance, but simply the fact that we are injecting new capital into the index, which mechanically reduces the return on contributed capital even though the existing gains remain intact.

The original thesis was relatively simple. We wanted exposure to the companies building the intelligence, software and systems required to detect and respond to increasingly cheap autonomous threats. Our initial positions included companies such as Palantir, CACI and Leidos.

As early as November 2023, the CSIS was already describing small drones as a significant military threat because of their low cost, rapid proliferation, and ability to overwhelm defensive systems.

By 2024, the war in Ukraine had clearly demonstrated the growing role of inexpensive commercial and FPV drones, while NATO held a major exercise in September 2024 specifically focused on counter-drone systems.

And recently, the thesis has started moving into another phase.

CACI is a good example. When we initially added the company in November 2024, our investment note was simply: “Core — Counter-UAS & signals intelligence.”

In July 2026, CACI received a $500 million counter-UAS IDIQ contract for its SkyValor system. More importantly, after operational evaluations, the system is now moving into full-rate production.

Palantir is seeing a similar transition elsewhere in the defense stack. At the end of August, the U.S. Army moved its TITAN battlefield intelligence system from prototype into production, awarding Palantir a $127 million production order as part of the initial deployment.

So while I would not attribute every recent move in the index to one contract or announcement, there is a broader pattern emerging across the companies we track: defense technology is moving from experimentation toward procurement and production.

And that matters for this thesis.

When we created the index, we were primarily positioned around software, intelligence, signals and systems integration.

In 2026, I think the opportunity is expanding toward another part of the stack: hardware deployment, mass production, interceptors, electronic warfare and directed energy.

In other words, the thesis hasn’t changed. The investable surface area has.

Recently, we imported this index into Altis.finance, the platform we built to help you follow our investment theses and track how they evolve over time.

Following Our Investment Theses

This is exactly why we built Altis.finance.

An investment thesis shouldn’t simply disappear after we publish an article about it. I want to be able to track it for years, add companies when the thesis changes, remove exposure when something is invalidated, record every trade and see whether the thesis actually produces returns over time.

Macro Notes Premium members have access to the indexes published by Macro Notes and Altis Research on Altis.finance, with more than 100 investment theses currently tracked across AI infrastructure, defense, energy, robotics, commodities, healthcare and other structural themes.

One important note: this Friday at midnight, Macro Notes Premium will move to $500/year, aligning its price with Altis.finance Premium, as both plans now provide access to the same platform benefits and indexes.

If you subscribe through Substack today, you can still join for $300/year, and as long as you remain subscribed, that founding subscription price will not change.

Unlock:

• Premium access to Altis.finance

• Access to all our previous and future editions

• Access to the community space we’re building on Altis.finance, launching soon, where you’ll be able to comment on portfolios, interact with investment ideas, and connect with other members.

Upgrade to Macro Notes Premium

Below, I want to go one level deeper.

The first thing I want to understand is where this $40 billion counter-drone spending cycle is actually going. Counter-drone defense is not one market, it is an entire stack, from detection and electronic warfare to command systems, low-cost interceptors and directed energy. We’ll look at which parts of that stack are becoming most important, where procurement is accelerating and, more importantly, where I think the economics are becoming attractive for investors.

Then I’ll come back to our Counter-Drone Defense Index.

We started tracking this thesis in 2024, before the current procurement cycle had really begun. I’ll go through what worked, what didn’t, how our original positions in Palantir, CACI and Leidos fit into the thesis today, and which parts of our initial view have now been validated by actual contracts and production orders.

But the most important part of this edition is what comes next.

The portfolio we built in 2024 was primarily positioned around the software, intelligence and systems-integration layer. In 2026, the opportunity is expanding toward hardware deployment, mass production and cheaper ways of actually defeating drones at scale.

So today, we are making three new trades in the index.

In the premium section, I’ll cover:

• Where the counter-drone market is heading and which parts of the defense stack I expect to capture the largest share of this new spending cycle.

• What our existing index got right, and how the thesis has evolved since we first started tracking it in September 2024.

• The three new companies we are adding today, why each one fills a missing part of the index and why I think the timing is becoming more interesting now.

• How I am positioning the index for this next phase, including the role each new position plays inside the broader counter-drone stack.

• What could invalidate the thesis, because a $40 billion spending announcement alone is not enough. What matters is whether procurement actually translates into production, deployment and sustained revenue growth.

The opportunity today is different from the one we were investing in two years ago. The thesis is the same, but the companies positioned to capture the next phase are starting to change.

Read more

Today, let’s look at AI in the real world, because this is a macro trend that is booming right now, and the deployment of AI in robotics and autonomous systems is accelerating.

Amazon celebrated its millionth robot in 2025. The IFR also recorded 542,000 industrial robot installations in 2024, more than double the number ten years earlier.

Industrial robotics has been growing for a long time, but it hasn’t been particularly visible. Now we have applications that are much easier to see: autonomous cars, delivery robots.

The difference with virtual AI applications is that we don’t get the same rapid deployment. ChatGPT reached an estimated 100 million users in a few months. In a virtual environment, deploying an application has fewer physical constraints.

Those constraints mainly come down to compute and data centers, which is why demand for all kinds of infrastructure components has exploded alongside AI adoption.

But for AI applications in the real world, demand goes beyond data centers. It also means complex factories to produce these robots, including the large factories Tesla plans to use to build Optimus robots and autonomous cars at scale.

So when we invest in robotics, it is usually a long-term investment by nature. We expect deployment to be gradual, and demand to grow progressively.

But AI is starting to change this. The more autonomous industrial production becomes, the closer we could get to bringing the kind of mass adoption we already know in the virtual world into the real world.

A startup could invent a physical product and potentially reach hundreds of millions of users by outsourcing its production to a fully autonomous, highly efficient production line. Digital instructions could translate into real-world actions with far fewer steps in between.

This won’t remove every physical constraint, but it could make the process much faster.

And that is why I think we are heading toward a paradigm shift, and why investing in robotics now is probably a good idea over the long term.

On altis.finance, we share ten portfolios tracking different theses around autonomy and robotics. Here are the ten, along with their current returns:

The only thesis that hasn’t really worked out for us so far is surgical robotics. I will write a dedicated edition explaining why the constraints on robotics in surgery are different from those in other sectors, and what we will change in the portfolio to address this.

The best-performing one, on the other hand, is Actuators, Sensors & Smart Motors, up +93.7%. It focuses on the components behind robotics: allowing machines to sense, move, and interact with the physical world.

All the portfolios deserve a dedicated edition too. There is a lot to say about each one.

We now share more than 120 portfolios on altis.finance, published by Macro Notes and more than 15 other contributors.

If you’d like to become a contributor and share your own investment theses, visit altis.finance/contributors or contact me at pierre@altis.finance to learn more.

The idea is to make altis.finance an open platform where everyone can share portfolios and investment theses, and where you can follow their evolution in real time. A platform that goes beyond just the ideas we publish through Macro Notes.

You can see the thesis behind each portfolio and the reasoning behind each trade. Every purchase and sale is documented.

Access to all portfolios published by Macro Notes and Altis Research is included in Macro Notes Premium.

If you’re a Macro Notes Premium subscriber, you don’t need to pay for a separate altis.finance subscription, which normally costs $499 per year. That subscription is for people who discover the platform without being Macro Notes Premium members.

Subscribe today to unlock all our research and newsletter editions, plus access to altis.finance.

Subscribe to Macro Notes Premium for $300/year

Subscribe to Macro Notes Premium

1. The next challenge: turning intelligence into movement

To understand this trend, let’s take something simple: picking up an object and placing it on a shelf.

For us, this is almost automatic. But a robot has to identify the object, understand where it is, determine how to grasp it, apply enough force to hold it without damaging it, then adjust its movement if the object slips or encounters an obstacle.

And the constraints change depending on whether it is handling a cardboard box, a piece of clothing, or a piece of fruit.

This is a fundamental difference from digital AI applications. In the physical world, understanding an instruction is not enough. You have to execute it, with all the constraints of weight, friction, precision, and safety that come with it.

Amazon provides a concrete example with Vulcan, introduced in May 2025. This robot uses force sensors to detect contact and adjust its movements when storing or retrieving products. According to Amazon, it can handle around 75% of the types of items in its fulfillment centers. Situations it cannot handle are passed on to employees.

What interests me here is that we don’t need to wait for a robot that can do everything before we start seeing economic benefits.

A machine that reliably performs some tasks can already be useful, as long as integrating and supervising it does not cost more than the value it creates.

AI could gradually expand that range of tasks: recognizing more objects, adapting to more situations, and learning new tasks with less specialized programming.

But this requires several layers of technology.

Nvidia, for example, describes its approach to robotics around three functions: training the models, simulating the situations in which robots will operate, and then running those models on the machines themselves. Simulation allows developers to test behaviors before putting them into practice, although it does not replace testing in the real world.

This extends our thesis around AI infrastructure. Developing robotics still requires compute, but it also creates demand for everything needed to turn a digital decision into a physical action.

You need sensors to measure. Motors and actuators to move. Control systems to coordinate those movements.

The more complex the tasks we want to automate, the more important the quality of this entire chain becomes.

And this is the idea behind our components thesis, which I introduced earlier as the best-performing robotics and autonomy thesis among the portfolios you can follow on altis.finance.

At the time of writing, the portfolio is up around 93%.

2. Components: exposure to several areas of robotics

When we think about the winners in robotics, we naturally think about the companies making the machines we see: Tesla, Figure, industrial robot manufacturers, or companies developing autonomous vehicles.

It is similar with LLMs. You think about Claude and OpenAI. But when we look deeper, we see that some of the biggest beneficiaries so far have been the infrastructure suppliers that make this AI possible, receiving a substantial share of the investment behind its deployment.

Our theses around fiber optics and memory have covered this extensively.

And in robotics, there is a similar idea. That is what sits behind our Actuators, Sensors & Smart Motors portfolio.

We are talking about the components inside robots, and here is why they matter.

An actuator turns a command into movement. Depending on its design, it can integrate a motor, a transmission mechanism, and measurement or control components. A gearbox adjusts the motor’s speed and torque to suit the task. Sensors tell the system about its position, the force being applied, or what is happening around it.

These elements may look less impressive than a humanoid robot. Yet they determine a large part of what the machine can actually do.

Harmonic Drive provides an example of this industrial specialization. The company offers precision gears and actuators combining components such as gears, motors, and encoders. Its technologies are used in industrial and collaborative robots.

The investment thesis is quite simple: some suppliers could benefit from the growth of several manufacturers and applications, without depending on the success of a single robot model.

But we need to be precise. A drone, a surgical robot, and a humanoid do not all use the same components. Their requirements for weight, power, precision, and cost are different. We therefore need to identify the markets each supplier is actually exposed to.

And above all, an increase in the number of robots does not guarantee higher profits for every supplier.

Let’s take a purely illustrative example. If a manufacturer sells twice as many components, but its average selling price falls by 30%, its revenue increases by 40%, not 100%. And if production costs do not fall sufficiently, profitability can disappoint despite a growing market.

This is an important tension within the thesis: for robots to be adopted more widely, their cost often needs to come down. Suppliers may be asked to absorb part of that reduction.

So the question becomes: which components are difficult to replace? Which manufacturers offer precision, reliability, or durability that customers are willing to pay for?

In the premium section of today’s edition, we’ll explore:

• The components thesis in more depth: where demand could grow, which suppliers could retain pricing power, and what could weaken the investment case.

• The portfolio and its positions: the companies we have selected, what they supply, and how each one fits into the thesis.

• The trade history: when we added or reduced positions, the reasoning behind those decisions, and how the portfolio has evolved.

• What we are watching next: the developments that could strengthen our conviction or lead us to adjust our positions.

Your Macro Notes Premium subscription also includes access to this portfolio on altis.finance, where you can follow its evolution in real time and see the reasoning behind each recorded trade.

That access extends to all the other portfolios published by Macro Notes and Altis Research, including the robotics and autonomy portfolios introduced earlier.

Read more

For me, the most effective way to invest over the medium to long term has always been a global macro approach, combined with top-down stock picking.

Put simply: identify a macro thesis — a structural change in a market, for example — then work your way down to the companies best positioned to benefit.

Europe’s rearmament is a good example. European countries are responding to the Russian threat by increasing defense spending. Which companies will benefit from those budgets, and where will that money actually go?

Often, following one thesis leads us to another.

In this case, a less obvious investment thesis is emerging from Europe’s rearmament: Europe wants to manufacture more weapons, but it still depends on China for some of the ingredients it needs.

Europe can increase its budgets, order missiles, and finance new factories. But those factories need specific materials and components to operate.

If an essential input is missing, having the building, the workers, and the money is not enough. A small component can hold up the delivery of equipment worth far more.

That is what this edition is about: military security also depends on industrial supply chains that are far less visible than aircraft or tank manufacturers.

Why does China matter so much?

Because it dominates several stages of raw-material processing.

Take the rare earths used to manufacture certain high-performance magnets. According to the International Energy Agency, in 2024, China accounted for approximately:

• 60% of global mining output for magnet rare earths;

• 91% of their refining;

• 94% of the corresponding sintered permanent magnet production.

These figures describe the global supply chain for rare-earth magnets. They do not mean that 94% of European weapons depend on China.

The crucial detail is the jump from 60% to 91%. Finding a mine in another country does not fully solve the problem if its output still has to pass through China for processing.

You need the entire chain: mining, separation, refining, manufacturing components, and qualifying those components for their intended use.

Rare earths and tungsten are two examples of the materials Europe needs to secure. The dependence also extends to specialized components and the chemical inputs used in ammunition production.

This is one of the emerging investment theses flagged by Radar, the new product we’re developing inside Altis.finance.

Radar is still in beta. Its purpose is to help identify promising theses while they are still taking shape, before they become obvious.

On Altis.finance, you can already follow more than 100 portfolios tracking investment theses, many of which are further along in their development.

We continue tracking them because an established thesis can reveal where the next opportunity is forming. As companies expand production, governments change policy, and spending moves through the economy, new constraints and new beneficiaries emerge.

Following a thesis over time gives us the context to recognize its next chapter. That is part of what we are building Radar to do.

Europe’s dependence on China shows how this works. Following the rearmament cycle leads us to a deeper question about industrial sovereignty — and a different set of potential investment opportunities.

Imagine a brand-new ammunition factory. The orders are signed, the machinery is installed, and the workers have been hired. But one supplier cannot deliver an essential material.

The entire production schedule can slip.

How much would you then value a company capable of securing that supply?

Its strategic importance can be far greater than the value of what it sells you.

Europe aims to mobilize up to €800 billion in additional defense spending under its rearmament plan. That is potential funding capacity, dependent on national spending decisions. Part of that ambition rests on less visible suppliers: companies that refine a metal, produce a propellant, or manufacture a component that is difficult to replace.

For suppliers with qualified production capacity already in place, this can change the relationship with their customers.

A manufacturer seeking to secure several years of production may commit to long-term purchases. A government may help finance a new facility. Reliable supply may justify a higher price.

For a smaller company, a few major contracts can transform its revenue outlook — and, if execution follows, its profits.

That is the transformation we want to identify: the moment an industrial supplier becomes essential to a national priority.

But the opportunity depends on exactly where the bottleneck sits. Opening a mine is not enough if refining still depends on China. Building a factory is not enough if customers have not yet qualified its output.

Our research therefore focuses on finding the companies that can actually relieve these constraints, then assessing how much of that opportunity their share prices already reflect.

Who already has the capacity Europe will need tomorrow — and where could the next orders materially change the scale of the business?

That is where our selection begins.

Today, we’re launching a new live portfolio on Altis.finance dedicated to this thesis.

Subscribe to Macro Notes Premium to access our research and follow the portfolio, including its positions and trade history.

Or choose the Founding plan if you also want access to Radar, which will be available exclusively to Founding Members.

Subscribe to Macro Notes Premium

It trades around the clock, seven days a week. As I write this, its market capitalization sits at roughly $1.57 trillion, with tens of billions of dollars changing hands on active trading days.

Yet for most of Bitcoin’s history, a holder who needed liquidity faced a strangely simple choice:

Sell Bitcoin.

Need capital to fund a business?

Sell Bitcoin.

Need cash for a large purchase?

Sell Bitcoin.

Want to redeploy capital into another opportunity?

Sell Bitcoin.

And every sale solves the immediate liquidity problem by creating another one: you give up part of your exposure to the asset.

For a trader, that may not matter.

But for someone who owns Bitcoin because they expect to hold it for five, ten or twenty years, it is a structural mismatch. You are financing a short-term cash requirement by liquidating a long-duration position.

This is especially relevant to Bitcoin because the long-term investment thesis has always been built around scarcity and continued financial adoption. Bitcoin’s protocol caps supply at 21 million coins, while its integration into traditional financial infrastructure has continued to expand. In January 2024, for example, the SEC approved the listing and trading of multiple spot Bitcoin exchange-traded products in the United States.

Whether that thesis ultimately proves right or wrong isn’t really the point here.

The point is that if you believe Bitcoin will be worth more ten years from now, selling it because you need liquidity today is an inefficient way to manage the position.

And this isn’t how large pools of wealth generally operate elsewhere.

A founder doesn’t necessarily liquidate his company shares every time he needs capital.

A real-estate owner can borrow against property rather than selling the building.

Investors can borrow against securities portfolios. Schwab, for example, explicitly offers a pledged-asset credit line designed to give investors liquidity without liquidating their investments.

Even assets we normally think of as highly illiquid can become collateral. Bank of America and J.P. Morgan offer financing against fine-art collections so wealthy clients can access liquidity without selling the underlying works.

The asset remains on the balance sheet.

The liquidity comes from somewhere else.

And Bitcoin is beginning to make the same transition.

That is what I want to explore today.

A Quick Note Before We Go Further

For this edition, we partnered with CoinRabbit, a crypto-backed lending and digital-asset management platform.

This is a sponsored edition, but CoinRabbit’s involvement went considerably further than a normal sponsorship placement. Their team gave us access to several pieces of research and proprietary data that helped shape the thesis you are reading.

Specifically, they provided four resources:

1. CoinRabbit × CryptoQuant — Buying the Bear, an August 2026 study of whale and institutional-scale positioning across Bitcoin, Ethereum and XRP.

2. CoinRabbit × GoMining — The Bitcoin Mining Playbook 2026, examining how post-halving mining economics are pushing operators toward more sophisticated treasury and balance-sheet management.

3. CoinRabbit × ChangeNOW — Financial Privacy in the Digital Age, focused on the implications of putting increasingly large amounts of private wealth onto publicly observable blockchains.

4. CoinRabbit’s proprietary borrower data, covering the last four months of borrowing activity and breaking down repeat borrowing and collateral preferences between retail and high-net-worth users.

CoinRabbit explicitly gave us freedom to use the metrics, research and insights wherever they fit our analysis.

The sponsorship made access to those materials possible. The interpretation and conclusions below are mine.

And together, these datasets point toward a much broader change than simply the growth of another crypto-lending product.

Learn more about CoinRabbit.io

Bitcoin Is Changing Function

Bitcoin’s financial history can be simplified into something like this:

Buy → Hold → Sell

You acquired Bitcoin.

You held Bitcoin.

Eventually, if you wanted to extract economic value from your position, you sold some of it.

That model is now expanding into something more interesting:

Reserve → Collateral → Liquidity

Bitcoin isn’t just something that can sit on a balance sheet.

Increasingly, it can be used to finance the balance sheet itself.

That distinction may sound technical, but I think it matters.

There is a major difference between owning $1 million of an asset and owning $1 million of an asset that can support credit.

In the first case, accessing $200,000 means selling part of the asset.

In the second, the asset can potentially remain economically exposed while supporting a separate source of liquidity.

Traditional finance already works this way at enormous scale.

Real estate supports mortgages.

Treasuries support repo markets.

Equity portfolios support securities-backed credit lines.

Fine art can support specialty lending.

And we are now beginning to see Bitcoin inserted into similar structures.

In March, Better Home & Finance and Coinbase launched a structure allowing qualified U.S. homebuyers to pledge Bitcoin or USDC against a separate loan used to fund the down payment on a conventional mortgage. The primary mortgage itself remains a standard conforming loan eligible for Fannie Mae backing.

That may look like a niche product today.

But conceptually, it represents something much larger.

Bitcoin is starting to move from:

an asset you can own

to:

an asset you can finance against.

And if that continues, the next stage of Bitcoin’s institutionalization may not only be about who buys it.

It may increasingly be about what holders can do with it once they own it.

The First Signal: Large Holders Are Accumulating

Before thinking about credit, we first need to understand what the holders with the most capital are actually doing.

This is where the research CoinRabbit produced with CryptoQuant becomes useful.

Their August report tracks large-holder behavior across Bitcoin, Ethereum and XRP using wallet balances, order flow and realized-price data.

The Bitcoin data is particularly interesting.

According to CryptoQuant, whale balances — excluding exchanges and mining pools — bottomed at roughly 2.87 million BTC in December 2025.

By the time of the report, those holdings had climbed to around 3.06 million BTC.

More importantly, accumulation accelerated when Bitcoin fell below $60,000 during the June selloff.

Now, there is an important timing distinction here.

Bitcoin is no longer below $60,000. As of August 30, it has recovered toward roughly $78,000. So this isn’t a claim that whales are currently buying BTC below $60K; it is evidence of how they behaved when the market gave them that opportunity earlier this summer.

And that behavior is exactly what we would expect from investors with a longer time horizon.

They weren’t simply surviving the drawdown.

They were using it to increase exposure.

Ethereum shows an even more striking divergence.

CryptoQuant found that wallets holding between 1,000 and 10,000 ETH had reduced their aggregate holdings from roughly 15.6 million ETH in January to about 12.9 million.

But the larger 10,000–100,000 ETH cohort moved in the opposite direction, increasing from around 14 million ETH in mid-2025 to approximately 19.6 million ETH by the time of the report.

The 100,000+ ETH cohort added roughly 1.8 million ETH from mid-2025, according to the same research.

This doesn’t guarantee that the bottom is in. CryptoQuant itself explicitly cautions that further downside remains possible.

But I don’t think “whales are bullish” is actually the most interesting conclusion from this data.

The more interesting question is what happens after they accumulate.

Because as an investor’s exposure increases, another problem becomes increasingly important:

How do you access capital without selling the assets you have deliberately spent a bear market accumulating?

And that is where the collateral layer begins to matter.

Separating Liquidity From Liquidation

Historically, the two were almost the same thing in crypto.

Need liquidity?

Sell crypto.

But collateralized lending separates those decisions.

The old equation was:

Need liquidity → Sell Bitcoin

The emerging equation is:

Need liquidity → Sell Bitcoin OR Borrow Against Bitcoin

The mechanics are relatively straightforward.

A holder places Bitcoin as collateral.

A lender advances capital against a portion of its value.

The borrower can use that liquidity for another purpose.

Interest accrues.

When the loan is repaid, the collateral is released according to the lender’s terms.

The CoinRabbit × GoMining research describes essentially this flow: deposit crypto as collateral, receive liquidity, use the borrowed funds, pay interest, repay principal and receive the collateral back.

This obviously isn’t free money.

Selling eliminates part of your exposure.

Borrowing preserves the economic thesis but introduces debt.

That means interest expense, counterparty risk, collateral risk and potentially liquidation.

The important change isn’t that one option is universally better.

It’s that holders now have two different tools for two different problems.

And when I looked through the proprietary CoinRabbit data, one number made me think this behavior is already becoming much more routine than I expected.

For readers who want to see the mechanics of a Bitcoin-backed loan directly, CoinRabbit explains its current BTC lending structure here. Explore Bitcoin-backed loans on CoinRabbit

65% Come Back

CoinRabbit shared a dataset covering its last four months of borrower activity.

It divides users into two groups: retail clients holding less than $500,000 on the platform and high-net-worth clients above $500,000.

The number that jumped out at me was this:

65.1% of borrowers in the dataset were repeat borrowers.

I think that’s more revealing than simply knowing how many loans were originated.

If collateralized borrowing were primarily an emergency mechanism — something people discovered once when they desperately needed liquidity — you might expect usage to be mostly transactional.

Borrow once.

Solve the problem.

Leave.

Instead, a large proportion return.

That suggests that for at least part of CoinRabbit’s customer base, borrowing against crypto is starting to behave more like a liquidity-management tool than a one-off financial event.

The collateral mix is also interesting.

Among the HNW segment in the internal dataset, CoinRabbit reports BTC at 30.53% of the collateral mix, followed by XRP at 26.85%, ZEC at 24.20%, and XMR at 6.17%.

Retail customers look different: XRP represents 35.19%, Bitcoin 30.22%, while ETH is much further behind at 3.19%.

Retail borrowers show a noticeably different mix.

I would not extrapolate these percentages to the entire crypto market.

This is a CoinRabbit-specific dataset covering four months, not a representative survey of every crypto holder.

But that’s precisely why I find the repeat-borrowing figure useful.

We’re not looking at what investors say they might do with crypto-backed credit.

We’re looking at a group of actual borrowers and seeing that once some of them use their assets this way, they tend to use the mechanism again.

And there is another part of the Bitcoin economy where this problem is even more obvious.

Miners.

Bitcoin Miners Have the Perfect Liquidity Problem

I think miners are one of the best laboratories for understanding what could eventually happen to Bitcoin holders more broadly.

Their business creates an unusually clean mismatch.

They produce Bitcoin.

But almost everything required to produce it is paid for in fiat currency.

Electricity. Employees. Hosting. ASICs. Cooling. Land. Debt service. Infrastructure.

For years, the obvious solution was:

Mine BTC → Sell BTC → Pay Opex

That worked reasonably well when mining margins were wide.

Then came the April 2024 halving.

The network’s block subsidy fell from 6.25 BTC to 3.125 BTC, cutting the primary block reward in half.

At the same time, network competition continued increasing.

CoinShares estimated that the weighted average cash cost to produce one Bitcoin among publicly listed miners had reached approximately $79,995 in Q4 2025, while hashprice fell as low as roughly $29 per PH/s/day during Q1 2026.

The result was predictable.

Public Bitcoin miners sold more than 32,000 BTC during Q1 2026, according to data analyzed by TheEnergyMag — more than their net sales during all of 2025 and a new quarterly record.

Think about what is happening here.

These companies exist to produce Bitcoin.

Many presumably have a long-term positive view on the asset.

Yet their operating model can force them to sell precisely when mining economics deteriorate and Bitcoin is weak.

They are long Bitcoin strategically but short liquidity operationally.

That is exactly the problem collateralized credit is designed to solve.

From Mining Bitcoin to Managing Bitcoin

This is the central argument in the CoinRabbit × GoMining report.

The mining industry is gradually moving from a pure production mindset toward a balance-sheet mindset.

The mining industry is gradually moving from a pure production mindset toward a balance-sheet mindset.

The report describes the historical model as:

Mine → Sell BTC → Pay Opex

and the more capital-efficient alternative as:

Mine → Keep BTC → Access Liquidity → Pay Opex

That seems like a small adjustment.

It isn’t.

It changes what the mined Bitcoin is.

Under the first model, Bitcoin is inventory.

Under the second, Bitcoin starts to become a reserve asset and financing asset.

The report’s broader finding is that mining competitiveness is no longer solely determined by how efficiently an operator produces Bitcoin. Low-cost electricity and efficient ASICs remain essential, but financial engineering and treasury management increasingly determine who survives difficult parts of the cycle.

This is especially relevant because miners now have another competitor for their most valuable input: electricity.

AI and high-performance-computing operators are competing aggressively for power infrastructure, while some miners are simply turning their sites into AI data centers.

Hut 8, for example, signed a $9.8 billion, 15-year AI data-center lease in July covering hundreds of megawatts at its Texas campus.

A megawatt now has multiple competing uses.

That raises the cost of making poor capital-allocation decisions.

And it makes the BTC sitting on a miner’s balance sheet increasingly important.

The Miner May Be Showing Us What Comes Next

CoinRabbit’s own mining cohort provides another useful data point.

According to the company’s 2026 research, approximately 33% of CoinRabbit Private Program clients in its mining cohort use Bitcoin collateral to support short-term operational needs.

That means some miners are already using BTC as working-capital infrastructure.

Bitcoin is no longer simply their output.

It is becoming part of their financing stack.

The same research found significant demand for something even closer to traditional corporate revolving credit: more than 90% of miners interviewed said they would try an open-ended credit structure backed by a single collateral pool if CoinRabbit offered it.

That is particularly interesting to me.

Because once you move from individual loans toward a continuously accessible credit line, you are no longer talking about a crypto-native novelty.

You’re starting to recreate familiar corporate treasury infrastructure around a new kind of collateral.

And I suspect miners may simply be adopting it earlier because their need is more obvious.

An entrepreneur holding BTC has the same basic problem.

So does a family office.

So does a company with Bitcoin on its balance sheet.

So does an individual sitting on a concentrated position acquired years ago.

They may want liquidity.

They may not want to sell.

Bitcoin Is Becoming a Balance-Sheet Asset

We’ve already watched Bitcoin move through several stages of financialization.

First, exchanges made it liquid.

Then institutional custodians made it easier to hold at scale.

Derivatives made it possible to hedge and express more complex views.

Spot ETFs made Bitcoin accessible from a conventional brokerage account.

Now we’re starting to build the credit layer.

And this is generally what happens as an asset class matures.

The value of real estate isn’t simply that property can appreciate.

It also supports an enormous credit system.

A stock portfolio isn’t merely a collection of investments.

It can support portfolio-backed lending.

Treasury bonds aren’t valuable only because governments repay them.

Their collateral properties make them foundational to global financial markets.

The more credible and financeable an asset becomes, the more things can be built around ownership of that asset.

Bitcoin appears to be moving in that direction.

And we now even have examples outside crypto-native lending platforms.

The Better/Coinbase mortgage structure is a good illustration: Bitcoin collateral supports a separate loan that funds a home down payment, while the borrower avoids having to sell the Bitcoin simply to generate the cash.

This doesn’t mean Bitcoin suddenly behaves like Treasuries.

It clearly doesn’t.

Its volatility alone creates radically different collateral requirements.

But the direction of travel is worth paying attention to.

What Happens to Bitcoin’s Sell-Side?

This is where the macro implications become more interesting.

Suppose a long-term holder needs $100,000.

Historically:

Need $100K → Sell $100K of BTC

That creates $100,000 of potential spot supply.

Now add a functioning collateral market:

Need $100K → Sell BTC OR Borrow Against BTC

Some holders will still sell.

Some absolutely should sell rather than add leverage.

But the second option means that a liquidity requirement no longer automatically creates a liquidation requirement.

Over time, that could reduce one category of discretionary selling.

Miners make this obvious.

If a miner needs $5 million for electricity and has no financing, the BTC treasury becomes the financing source.

Sell the Bitcoin.

If the same miner can obtain appropriately structured secured financing instead, the operating expense doesn’t necessarily create immediate BTC supply.

I would be careful about taking this too far.

Crypto-backed lending is not mechanically bullish for Bitcoin.

In fact, leverage can make markets substantially more violent.

The interesting trade-off is this:

Collateralized credit can reduce voluntary selling during normal conditions while creating additional forced selling riskduring extreme conditions.

That brings us to the biggest problem with the entire thesis.

The Ghost of 2022

You can’t discuss crypto lending seriously without discussing 2022.

Celsius. BlockFi. Genesis.

The failure of centralized crypto lenders made one thing painfully obvious:

The collateral itself is only one part of the risk.

What happens after you hand it to someone matters just as much.

One of the concepts that moved to the center of the conversation after 2022 is rehypothecation.

At its simplest, rehypothecation means collateral pledged by a borrower can itself be reused by the lender for another purpose.

Your BTC backs your loan.

The lender then lends or pledges that collateral somewhere else.

That creates another counterparty.

Potentially another loan.

Potentially another source of leverage.

Potentially another point of failure.

The CoinRabbit × GoMining research argues that post-2022 due diligence has increasingly shifted away from simply optimizing interest rates toward minimizing counterparty and collateral-management risk.

I think this distinction is essential.

The lesson from 2022 shouldn’t be reduced to:

“Crypto lending doesn’t work.”

Traditional credit markets fail too.

The more useful question is:

What does the collateral architecture actually look like?

This is also where CoinRabbit’s product design becomes relevant to the story rather than just appearing as a sponsor.

Its current terms explicitly state that CoinRabbit does not re-lend, rehypothecate or otherwise reuse collateral. Its Bitcoin loan page says pledged collateral is kept in cold wallets with multisig access rather than lent out to generate interest.

That doesn’t remove every risk associated with using a centralized lender.

But it removes one specific layer of risk that became impossible to ignore after 2022.

CoinRabbit has a more detailed explanation of its no-rehypothecation framework here. Read CoinRabbit’s rehypothecation research

There Is No Free Liquidity

The second risk is simpler.

Bitcoin is volatile.

Debt is not impressed by your long-term thesis.

If you place $100,000 of BTC against a $50,000 loan, Bitcoin can decline substantially before the collateral value approaches the loan value.

If you place the same $100,000 against $90,000 of debt, your margin for error becomes dramatically smaller.

That is Loan-to-Value, or LTV.

Higher LTV:

More liquidity today.

But also:

Less protection against a drawdown tomorrow.

CoinRabbit currently offers LTV configurations between 50% and 90%, and its own product page explicitly warns that higher LTV means receiving more loan funds but facing a higher margin-call risk.

Interestingly, its research with miners found demand for LTV as high as 95%.

CoinRabbit says it currently caps maximum LTV at 90%, because a 95% structure would leave only a 5% collateral buffer before reaching margin thresholds.

I find that data point fascinating because it captures the basic tension of financial engineering perfectly:

Capital efficiency and financial fragility often increase together.

The more of an asset’s value you extract today, the less volatility you can tolerate tomorrow.

So the correct conclusion isn’t that holders should always borrow instead of selling.

Sometimes selling is the right answer.

The more important conclusion is that Bitcoin holders increasingly have access to the same decision framework wealthy investors already use with other assets:

What should I sell?

What should I continue holding?

What should I finance against?

And what level of leverage can my balance sheet actually tolerate?

Crypto Wealth Is Starting to Look More Like Private Banking

There is one final part of CoinRabbit’s research that initially seemed separate from this story but actually fits surprisingly well.

Their report with ChangeNOW looks at financial privacy.

As crypto wealth becomes larger, simply owning the assets creates problems that smaller holders rarely have to think about.

Public blockchains are transparent.

Once an address can be associated with an identifiable person or entity, transaction history, counterparties, balances and transaction patterns can potentially become visible to outside observers.

That is very different from private banking.

Your neighbors cannot open a website and inspect your brokerage account.

They cannot see your bank balance.

They cannot follow every wire transfer your company makes.

With public blockchains, the underlying financial infrastructure was designed around verifiability rather than discretion.

And as the amount of wealth stored on-chain grows, that trade-off becomes more important.

CoinRabbit’s internal HNW research found that roughly 50% of surveyed high-value holders had experienced at least one targeted scam or social-engineering attempt during the previous three years. More than 70% said they distribute assets across multiple wallets or custody providers.

Independent security data suggests the physical risk is also becoming harder to dismiss.

CertiK documented 52 verified physical “wrench attacks” in the first half of 2026, with approximately $124.1 million in recorded financial exposure. Europe accounted for 39 of the 52 cases, while France alone accounted for 33.

This is why I think the evolution we’re watching is ultimately larger than crypto lending.

As digital wealth matures, the infrastructure around it starts recreating some of the functions private banking has provided wealthy clients for decades:

Liquidity.

Collateral management.

Custody.

Risk management.

Discretion.

CoinRabbit’s Private Program is explicitly aimed at clients with $500,000+ in capital and combines customized lending terms with private-manager support and other wealth-management services.

For readers managing larger digital-asset positions, you can learn more about CoinRabbit’s Private Program here.Explore CoinRabbit Private

The Next Question for Bitcoin

For most of Bitcoin’s history, the question was simple:

Should I buy Bitcoin?

Then it became:

How much Bitcoin should I hold?

I think the next question is increasingly going to be:

What can I do with Bitcoin once I own it?

That is a much more mature question.

And the infrastructure is beginning to answer it.

ETFs allow Bitcoin to sit inside traditional investment portfolios.

Institutional custody allows larger pools of capital to hold it.

Mortgage products are beginning to recognize it as collateral.

Miners are experimenting with using BTC reserves to finance operating expenses.

Platforms like CoinRabbit are building dedicated credit infrastructure around it.

And actual borrower behavior suggests that once some crypto holders begin using their assets as collateral, many return to the mechanism again.

None of this eliminates Bitcoin’s volatility.

It doesn’t eliminate liquidation risk.

It doesn’t eliminate lender risk.

And it certainly doesn’t turn borrowing into free money.

But it changes something fundamental.

For a long time, Bitcoin wealth had one obvious exit valve:

sell the Bitcoin.

Now another financial layer is being built.

Hold the asset.
Use it as collateral.
Access liquidity elsewhere.

That is how many mature assets already function.

And if Bitcoin is going to continue evolving from a speculative asset into a genuine reserve asset for individuals, companies and institutions, its ability to support credit may ultimately be almost as important as its ability to support ownership.

The next stage of Bitcoin’s institutionalization may therefore have less to do with who is buying it...

...and much more to do with what they can do with it once they refuse to sell.

This edition was produced in partnership with CoinRabbit. CoinRabbit provided proprietary data and research used throughout this analysis. Editorial conclusions and opinions are those of Macro Notes.

Crypto-backed borrowing involves material risks, including collateral volatility, liquidation, counterparty and custody risk. This edition is for educational purposes only and does not constitute investment, tax or financial advice.

Learn more about CoinRabbit

Go deeper with Macro Notes Premium

Macro Notes Premium gives you access to all of our premium research — but it now includes much more than the newsletter itself.

Your subscription also includes Altis Terminal, our investment research platform where you can explore 100+ live portfolios published by Macro Notes, Public Markets, and 10+ other contributors.

Track positions, trade history, performance, and the evolution of each investment thesis over time.

Altis Terminal normally costs $499/year as a standalone subscription. It is currently included with Macro Notes Premium at no additional cost.

Subscribe to Macro Notes Premium to unlock the full edition and access Altis Terminal.

Subscribe to Macro Notes Premium

I’ve never really been that interested in it.

There is obviously a lot of money to be made trading oil, but for me it’s too dependent on events, too short-term and too speculative.

That’s not really how I like to invest.

But the ships transporting that oil?

That’s much more interesting.

Tankers are a much more long-term and cyclical business. Ships get old. Regulations change. At some point, fleets need to be replaced. And because building new ships takes time, these cycles can become relatively predictable.

So today, we’re going to talk about tankers — the ships that transport crude oil and refined petroleum products around the world.

And when I say this business is cyclical, the basic idea is actually very simple.

When a lot of oil needs to be transported and there aren’t enough ships available, freight rates go up.

When too many new ships arrive, freight rates collapse.

But something unusual is happening to this cycle today.

The global tanker fleet is getting old.

Across the major tanker categories, around 18–21% of the fleet is already more than 20 years old.

Looking at the broader tanker fleet, more than 21% of tankers above 25,000 tonnes are at least 20 years old, while another 28% are between 15 and 19 years old.

Normally, many of these ships would gradually be scrapped and replaced.

But the war in Ukraine and the sanctions that followed created something very unusual:

a second global tanker fleet.

Hundreds of older tankers are now being used to transport Russian, Iranian and Venezuelan oil outside traditional Western markets.

S&P Global estimates that this so-called shadow fleet now represents around 22% of the global tanker fleet.

And this is where things get interesting.

We effectively have two markets:

GLOBAL TANKER FLEET

        ┌── Compliant fleet
        │   Western insurance
        │   Major oil companies
        │   Normal financing
        │
        └── Shadow / sanctioned fleet
            Russia / Iran etc.
            Older vessels
            Limited Western access

Sanctions can therefore paradoxically reduce the effective supply of tankers available to the traditional market, even though those ships still physically exist.

At the same time, oil shipping routes have become longer and less efficient.

Imagine a tanker that used to need 25 days to complete a trip and now needs 45.

That ship is occupied for much longer.

You haven’t removed a single tanker from the global fleet.

But you have effectively reduced the amount of transportation capacity available.

This is the basic idea behind something called ton-mile demand.

What matters isn’t only how many barrels of oil need to be transported.

It’s also how far each barrel has to travel.

So let’s summarize all of this very simply.

Imagine you have a fleet of 100 tankers.

A significant number of them are getting old. Another part of the fleet is now being used to transport sanctioned oil. And at the same time, shipping routes are getting longer.

You don’t necessarily have more oil to transport.

But suddenly, you need a lot more tanker capacity to transport the same amount of oil.

That’s the important part.

And when the utilization of the remaining ships gets high enough, freight rates don’t necessarily rise gradually.

They can explode.

And that’s basically what we’re seeing today.

Despite what is happening with the price of oil itself, geopolitical disruptions and longer shipping routes have kept tanker freight rates at extremely high levels.

That’s what I like about this thesis.

We don’t really need to predict whether oil will trade at $60, $80 or $120.

We just need to understand how much oil needs to move, how far it needs to move, and how many ships are actually available to move it.

But if you’ve followed the thesis this far, you probably already see the obvious problem.

You might be thinking:

“Okay Pierre, if there aren’t enough tankers, shipowners will simply order more of them... or they probably already have.”

And that’s exactly what’s happening.

In 2026, tanker orders exploded.

By the end of July, shipowners had already ordered 96.4 million deadweight tonnes of new tanker capacity this year.

That’s more than the previous full-year record.

And if you look at the total tanker orderbook today, it represents more than 30% of the existing fleet.

That’s a lot.

Actually, when we first started looking at these numbers, this almost seemed to kill the entire thesis.

If almost one new tanker is already on order for every three tankers currently operating, where exactly is the shortage?

But this is where the story gets more interesting.

Because that 30% number tells us how many ships are coming.

It doesn’t tell us what they’re replacing.

Remember what we saw earlier.

A huge part of the tanker fleet is already old.

More than 20% of the broader fleet is over 20 years old, and another large part is quickly approaching that age.

So a significant part of this enormous orderbook isn’t necessarily creating new capacity.

It’s replacing capacity that eventually has to disappear.

And there’s another problem.

Those old ships aren’t disappearing as quickly as you might expect.

Why?

Because freight rates are high.

If you’re the owner of a 20-year-old tanker that can still generate a lot of cash, you’re not necessarily going to send it to a scrapyard tomorrow.

You’re going to keep operating it for as long as you reasonably can.

And this creates a strange situation.

Tanker supply gets tight
        ↓
Freight rates increase
        ↓
Old tankers become more profitable
        ↓
Owners delay scrapping them
        ↓
The fleet gets even older
        ↓
More replacement will eventually be needed

So now we have two forces moving in opposite directions.

On one side, a massive wave of new tankers is coming.

On the other, we have one of the oldest tanker fleets in decades, a large shadow fleet absorbing older vessels, longer shipping routes increasing ton-mile demand, and relatively little scrapping because old ships are still making money.

And that’s where I think the thesis becomes much more interesting.

Because the question isn’t really:

“Is there a global tanker shortage?”

That’s too simple.

The real question is:

Which parts of the tanker market are actually facing a shortage — and which ones are about to get too many ships?

Because not all tankers are the same.

VLCCs, Suezmaxes, Aframaxes and product tankers have different fleet ages, different orderbooks, different delivery schedules and different exposure to global trade routes.

And once you start looking at the market this way, the picture becomes much more interesting.

So we went through the tanker market segment by segment.

We looked at the age of the existing fleet, how many new ships are coming, when they’re actually going to be delivered, how much capacity could eventually need to be scrapped, how sanctions are affecting available supply, and which parts of the market benefit the most from longer shipping routes.

And most importantly:

which public companies are actually positioned to benefit from all of this.

Today, we’re launching a new live portfolio on Altis Terminal built around this thesis.

In the Premium section, we’ll go through the complete research, identify where we think the most interesting supply-demand imbalances are, look at the listed companies exposed to them, their valuations and risks, and explain every position we’re adding to the portfolio.

Macro Notes Premium members can also follow the portfolio live on Altis Terminal as the thesis evolves.

Get Macro Notes Premium + Altis Terminal

Macro Notes Premium includes full access to Altis Terminal, which normally costs $499/year as a standalone subscription.

On Altis, you can access 100+ live investment portfolios and theses, follow every trade in real time, see the complete trade history, and track how each thesis evolves over time.

Macro Notes Premium is currently $300/year, but the price will increase to $400/year at the end of this week.

If you subscribe today, you lock in the $300/year price — existing Premium subscribers will keep their current price and won’t be affected by the increase.

→ Upgrade to Macro Notes Premium today and lock in $300/year.

Upgrade to Macro Notes Premium

The Real Opportunity Starts Here

So far, we’ve established something important.

The tanker market looks tight.

The fleet is old, sanctions have removed a large number of vessels from the traditional market, trade routes are getting longer, and freight rates have exploded.

But at the same time, shipowners have already ordered a huge number of new tankers.

And that’s where the easy version of the thesis ends.

Because the real opportunity isn’t simply:

“There aren’t enough tankers.”

Some parts of the market could actually end up with too many ships.

Others look very different.

In the Premium section, we’re going to break the tanker market down segment by segment — VLCCs, Suezmaxes, Aframaxes and product tankers — to understand where the real supply-demand imbalance is hiding.

We’ll look at:

• how old each part of the global fleet really is;

• how many new ships are coming and when they will actually be delivered;

• which segments have the biggest replacement problem;

• why modern tankers can already be worth more than brand-new ships still waiting to be built;

• how sanctions and the shadow fleet are changing effective supply;

• why longer Atlantic-to-Asia oil routes could structurally increase tanker demand;

• which parts of today’s record freight rates are structural, and which are simply driven by current geopolitical disruptions;

• and the biggest risk to the thesis: the enormous wave of newbuilds coming in 2027–2029.

Most importantly, we’ll identify the five public companies we believe are best positioned for this cycle, explain why we chose each one, look at their current valuations, balance sheets, fleet exposure and risks, and show the exact weights we’re using.

We’re also launching a new Tanker Replacement Cycle portfolio on Altis Terminal, where Premium members can follow every position, trade and future update to the thesis in real time…

Read more

GPUs.

Companies wanted far more computing power than Nvidia and the semiconductor supply chain could provide. Prices rose, hyperscalers rushed to secure capacity, and investors focused almost entirely on the accelerator layer.

But over the past year, another bottleneck has quietly become much more important.

Memory.

And the reason is structural.

GPUs perform the calculations, but an AI system needs several additional layers around them to continuously feed, process and retain data.

A simplified AI infrastructure stack looks like this:

AI accelerator
      ↓
HBM
Feeds enormous amounts of data directly into the GPU
      ↓
Server DRAM
Holds active workloads, model states and temporary data
      ↓
Enterprise SSDs
Provide fast access to datasets, embeddings and inference context
      ↓
Mass-capacity HDDs
Store huge volumes of checkpoints, logs, video and historical data cheaply

Each layer solves a different problem.

HBM is the fastest and most strategically important.

Modern accelerators need extraordinary memory bandwidth because having a powerful GPU is useless if data cannot reach it quickly enough.

And each generation of AI accelerator is using more HBM.

That creates an unusual problem for the memory industry.

HBM is manufactured using many of the same underlying semiconductor resources as conventional DRAM, but it consumes substantially more capacity per unit of usable memory.

So when manufacturers such as Micron, SK Hynix and Samsung allocate more production toward HBM, they are not simply creating additional AI memory.

They are also reducing the effective capacity available for conventional server and consumer DRAM.

That is the first bottleneck.

Then comes storage.

AI models continuously create and consume enormous datasets.

Training requires data.

Inference generates new data.

AI agents retain context.

Video models produce enormous files.

Models create checkpoints that need to be saved.

Not all of this information needs expensive ultra-fast storage.

Frequently accessed datasets are increasingly stored on enterprise SSDs, which creates demand for NAND flash.

But the enormous volume of colder data still needs the lowest possible cost per terabyte.

That is why high-capacity hard drives, a technology many investors expected SSDs to eventually replace, have unexpectedly become another beneficiary of the AI infrastructure boom.

The result is that AI demand is now pulling on almost every major layer of the memory and storage supply chain at the same time:

More AI compute
      ↓
More HBM per accelerator
      ↓
Less effective DRAM capacity
      ↓
Higher server memory demand
      ↓
More active data → Enterprise SSD demand
      ↓
More retained data → HDD demand

And unlike software, none of these industries can instantly add supply.

Memory fabrication plants cost billions of dollars.

They take years to build.

New processes need to be qualified.

Enterprise SSDs must pass demanding customer tests.

Hard-drive manufacturers need years to transition toward new technologies such as HAMR.

So demand can move much faster than capacity.

That imbalance was the core observation behind our Memory & Storage Shortage portfolio.

We first started investing around this idea in September 2025 through Public Markets and documenting the positions on Altis Terminal.

At the time, the thesis was relatively simple:

AI infrastructure was creating data and consuming memory faster than the physical supply chain could expand.

Since then, the portfolio has returned +732%.

But something more interesting is happening now.

The shortage is beginning to escape the semiconductor industry.

Nvidia’s latest results showed that AI demand remains extraordinarily strong, with data-center revenue more than doubling from a year earlier.

Yet Nvidia is now also dealing with rising memory and component costs that are expected to pressure margins.

SK Hynix, one of the world’s largest memory manufacturers, now believes the global memory shortage could persist through 2030.

And despite enormous new investments in capacity, memory prices continue to rise.

TrendForce currently expects conventional DRAM contract prices to increase another 13–18% quarter over quarter, while NAND prices are expected to rise 10–15%.

Those increases are much slower than earlier this year.

But prices are still rising from already extraordinary levels.

And that matters because DRAM and NAND are not only used inside AI servers.

They are inside:

• PCs.

• Smartphones.

• Cloud infrastructure.

• Networking equipment.

• Cars.

• Industrial systems.

• And increasingly almost every device with meaningful computing power.

This is where the thesis begins to change.

Initially, AI created an investment opportunity for the companies supplying memory.

Now, rising memory prices are becoming an input cost for everyone else.

PC manufacturers either absorb the higher costs and lose margin, or increase prices.

Smartphone manufacturers face the same decision.

Cloud providers need more expensive infrastructure.

Networking equipment becomes more costly.

Even Nvidia — arguably the company with the greatest pricing power in the entire AI supply chain — is beginning to feel the effect.

We may therefore be entering the next chapter of the AI infrastructure boom:

AI is no longer simply increasing the price of computing. It is beginning to increase the price of everything that needs memory.

And that raises a much more difficult question for investors.

The shortage remains real.

But after memory and storage stocks have already appreciated several hundred percent, is the investment opportunity still as attractive as the underlying industry thesis?

This week, we are going back through the portfolio that originally positioned us for this shortage to understand what has changed, which parts of the supply chain remain structurally constrained, where new capacity is beginning to appear, and where we would allocate capital today.

Premium members can follow the thesis in real time

Macro Notes Premium is not limited to the editions we publish here.

Your subscription also includes access to Altis Terminal, where you can follow the Memory & Storage Shortage portfolio in real time — including its positions, trade history, performance and future updates.

Altis Terminal currently includes more than 100 live portfolios covering investment theses published across Macro Notes, Public Markets and our other research newsletters.

Premium members also receive:

• Access to the complete Macro Notes archive.

• Every future Premium edition.

• Access to our live portfolios and their ongoing updates.

• More than 100 investment theses organized and tracked inside Altis Terminal.

Macro Notes Premium is currently $300 per year until the end of this week.

We will be updating the subscription price after that.

If you subscribe before the increase, your $300 annual price will remain locked for as long as you stay subscribed.

Subscribe today to unlock the full edition and access the Memory & Storage Shortage portfolio on Altis Terminal.

Subscribe to Macro Notes Premium

In the Premium section, we will look at:

• Why the memory shortage could remain structural for several more years.

• How HBM production is affecting conventional DRAM and NAND supply.

• What Nvidia’s latest results tell us about memory costs and AI infrastructure margins.

• Why the shortage is now beginning to affect PCs, smartphones and other hardware markets.

• A complete update of our Memory & Storage Shortage portfolio, including the positions that drove its +732% return.

• The trades we have already reduced after exceptional reratings.

• Which companies still appear best positioned to benefit from the shortage.

• Where valuations and expectations may now have gone too far.

• And most importantly, where we would allocate new capital today.

Read more

Here is the podcast version created with NotebookLM. The goal of this podcast is to provide additional insights and perspectives on the topics covered in our published editions.

This first episode explores sovereign AI, the subject of this article. Feel free to read it before or after listening to the podcast.

Read the post

As a reminder, if you want to go beyond the analysis we share in the podcast and our publications, you can now explore more than 100 portfolios on the platform we launched a few days ago: Altis.finance/terminal.

Altis Terminal brings together all our investment theses in the form of live portfolios, complete with trade histories, performance tracking, updates, and more.

The platform normally costs $499 per year, but as a Macro Notes Premium subscriber, you receive access at no additional cost. It is included in your $300 annual subscription, along with all previously published premium content and every future edition and research report.

The price of Macro Notes Premium will increase soon. Subscribe today to lock in the current $300 annual rate for both Macro Notes Premium and Altis Terminal. As long as you remain subscribed, this price will never increase for you.

Subscribe to Macro Notes Premium

Subscribe to Macro Notes Premium

Today, I want to step away from data centers and semiconductors to examine a different kind of infrastructure, one that could allow physicians to detect cancer when conventional scans still see nothing.

We previously explored the investment thesis around GLP-1 drugs. It attracted enormous attention for a simple reason: obesity affects an exceptionally large population. Approximately 40.3% of American adults are living with obesity, and the commercial opportunity became large enough to help turn Eli Lilly into the first pharmaceutical company valued at more than $1 trillion.

The thesis we are examining today addresses a different problem.

Cancer does not create the same kind of mass-market opportunity as obesity treatments, but the cost and complexity of treating each patient can be much greater. Direct cancer-care costs in the European Union reached an estimated €120 billion in 2023, representing approximately 6.9% of total healthcare spending.

But the most interesting part of this thesis begins with a much more personal problem.

A patient can undergo surgery to remove a tumor.

The operation appears successful, and the scans performed afterward show no visible sign of cancer.

Unfortunately, that does not always mean every cancer cell has disappeared.

A small number of cells can remain somewhere in the body, in quantities too low to appear on a CT scan or MRI. They may remain invisible for months before multiplying enough to produce a detectable recurrence.

This is why someone can be treated successfully, declared free of visible disease, and later discover that the cancer has returned.

After surgery, physicians must decide what comes next.

Should the patient receive chemotherapy or immunotherapy? Or should the medical team avoid an unnecessary treatment and continue monitoring?

These decisions still frequently depend on imperfect indicators: the size and location of the original tumor, whether it reached the lymph nodes, and the recurrence rates observed in patients with similar characteristics.

As a result, some patients may receive aggressive treatment even though their cancer has already been eliminated.

Others may receive no additional treatment because nothing remains visible, despite a small number of cancer cells still being present.

The problem is not always that we lack an effective treatment.

Sometimes, the problem is that we cannot measure precisely what remains to be treated.

This is where liquid biopsies enter the picture.

Searching for Molecular Traces

Cancer cells can release small fragments of their DNA into the bloodstream.

By analyzing a blood sample, physicians can search for these fragments and potentially detect cancer activity that remains invisible through conventional imaging.

This is known as circulating tumor DNA, or ctDNA.

The science becomes extremely complex very quickly. But from an investment perspective, the central idea is relatively simple:

Instead of waiting for a tumor to become large enough to appear on a scan, physicians can search for the molecular traces left by cancer cells.

This technology can potentially be used at several stages of the disease.

Before a patient has been diagnosed, it may help screen for certain cancers in people who do not yet have symptoms.

Once cancer has been identified, it can reveal some of the mutations driving the tumor and help physicians select a more appropriate treatment.

After surgery, it can search for molecular residual disease—the microscopic traces of cancer that may still remain in the body.

During treatment and in the years that follow, repeated blood tests can then show whether the molecular signal is falling, remaining stable or beginning to rise again.

This is the investment thesis we are going to examine today.

Cancer care is beginning to move from occasionally observing a tumor to measuring its molecular signal over time.

From a Snapshot to a Timeline

A traditional tissue biopsy provides an extremely valuable snapshot of a tumor at a particular moment.

A liquid biopsy offers something different.

Because a blood draw is easier to repeat than a tissue biopsy, it can potentially create a timeline showing how the disease is evolving.

A first test could be completed after surgery.

Another could follow several months later.

Additional tests could then monitor for recurrence or determine whether a treatment is reducing the molecular signal.

The business model therefore becomes potentially recurring.

This is where the medical thesis becomes economically interesting.

A sequencing machine is sold to a laboratory or hospital. Its manufacturer can continue selling consumables and services, but it remains primarily an infrastructure provider.

The company that owns the diagnostic test operates closer to the medical decision.

It develops the assay, processes the sample, interprets the result, obtains reimbursement, builds relationships with physicians and insurers, and accumulates clinical data over time.

The same test can also be used repeatedly for the same patient.

The economic value could therefore migrate from the companies manufacturing the sequencing machines toward those controlling the tests, the clinical evidence, the data and the reimbursement relationships.

This distinction became one of the most important lessons from our Precision Oncology portfolio.

From Prediction to Medical Action

For several years, studies have shown that the presence of tumor DNA in the blood after surgery is associated with a substantially higher risk of recurrence.

But identifying a high-risk patient is only the first step.

The more important question is whether the result can change a medical decision and improve the treatment that patient receives.

In May 2026, the FDA crossed an important threshold.

For certain patients with muscle-invasive bladder cancer, the agency approved an immunotherapy after surgery specifically for patients whose blood still showed molecular residual disease, as determined by an authorized test.

This is important because the blood test is no longer being used only to estimate whether a patient faces a higher risk of recurrence.

It can help determine whether that patient should receive a specific treatment.

The test is beginning to move from prediction to action.

If this model expands into colorectal, breast, lung and other cancers, the treatment pathway could gradually become:

Remove the tumor
        ↓
Search for molecular residual disease
        ↓
Treat patients who remain positive
        ↓
Continue monitoring the molecular signal

This could allow physicians to direct aggressive treatments toward the patients most likely to benefit, while potentially sparing others from unnecessary side effects.

It also places the companies controlling these tests at one of the most important points in the cancer-care process: the moment when a physician decides what to do next.

The Portfolio Behind the Thesis

In August 2024, we created a locked Precision Oncology portfolio to follow this transition.

The objective was not simply to select a group of healthcare stocks. It was to identify which companies were capturing the most value as cancer care moved toward blood-based screening, treatment selection and molecular residual disease monitoring.

Since its creation, the portfolio has generated a total return of 142.7%.

More importantly, the composition of the portfolio changed as our understanding of the market evolved.

We began with exposure to both the companies developing the diagnostic tests and the infrastructure used to sequence DNA.

We later concluded that the strongest economic value was moving toward the owners of the assays rather than the manufacturers of the machines.

That led us to exit one part of the portfolio and reallocate the proceeds toward a company operating closer to the patient and the medical decision.

Follow the Portfolio Live on Altis Terminal

The Precision Oncology portfolio is available live on Altis Terminal.

You can review the documented trade history, follow how the thesis evolves and receive email alerts whenever a new trade is recorded or an important portfolio update is published.

Macro Notes Premium also includes access to more than 100 live portfolios covering some of the most promising investment themes we are currently following.

Altis Terminal normally costs $499 per year when purchased separately. It is included at no additional cost with Macro Notes Premium, alongside access to all our past and future premium research, for $300 per year.

The price of Macro Notes Premium will be updated soon. Subscribers who join today can secure the current $300 annual rate before the increase.

Subscribe to Macro Notes Premium

The clinical thesis is now stronger than it was in 2024. Test volumes are growing rapidly, reimbursement is expanding, and liquid-biopsy results are beginning to influence actual treatment decisions.

But the share prices have also appreciated substantially.

We therefore need to separate two questions:

Are liquid biopsies becoming an important part of cancer care?

Are the companies exposed to this market still attractive at their current valuations?

The answer to the first question can be positive without automatically making the second one true.

In the premium section, I will open the complete portfolio and examine:

• every position and transaction since August 2024;

• how much each company contributed to the 142.7% return;

• why we exited our sequencing-infrastructure position;

• the companies controlling the most valuable tests;

• the importance of clinical evidence, reimbursement and patient data;

• the medical, competitive and regulatory risks;

• current valuations following the rally;

• the catalysts we are monitoring next;

• and, most importantly, whether I would still initiate or add to these positions today.

Read more

But some of the first companies to capture meaningful economic value are often those supplying the infrastructure that makes the market possible.

The market usually recognizes the trend before it correctly prices all of its consequences.

Artificial intelligence is perhaps the clearest recent example.

AI existed long before ChatGPT, but the mass adoption of large language models fundamentally changed the amount of computing infrastructure the industry required. We are now entering another chapter as AI moves from answering questions to operating through agents capable of executing increasingly complex tasks.

The first and most obvious infrastructure winner was Nvidia.

In the third quarter of its 2023 fiscal year, Nvidia generated $3.83 billion in quarterly Data Center revenue. By the fourth quarter of fiscal 2026, that figure had reached $62.3 billion—more than sixteen times higher in just over three years.

But GPUs were only the first bottleneck.

Once companies secured the chips, they needed servers capable of deploying them at scale. Dell subsequently closed more than $64 billion in AI-optimized server orders during fiscal 2026, shipped more than $25 billion, and entered the following year with a $43 billion backlog.

Then the constraint moved again.

Thousands of accelerators had to be connected. That increased demand for optical components, high-speed networking, switches and fiber—the infrastructure behind the fiber-optics thesis we have covered extensively in Macro Notes.

More compute also produced more heat, creating demand for liquid-cooling systems, heat exchangers and specialized data-center equipment. The electrical infrastructure surrounding the servers became equally important: transformers, switchgear, copper, gas turbines and access to reliable power.

According to the International Energy Agency, global data-center electricity consumption reached approximately 485 terawatt-hours in 2025 and could rise to roughly 950 terawatt-hours by 2030. Electricity consumption from AI-focused data centers alone is expected to triple over that period.

Amazon, Microsoft, Alphabet, Meta and Oracle collectively invested approximately $427 billion in capital expenditure in 2025. Their combined spending was projected to reach roughly $560 billion in 2026—an increase of around 30% in a single year.

That capital is no longer flowing exclusively toward chips. It is moving through the entire physical chain required to turn those chips into usable computing capacity.

This is the part of emerging markets that interests me most.

And the same sequence may now be beginning again—but this time, the next major buyers of AI infrastructure may not be technology companies. They may be governments.

Start with the market, not the stock

The most effective way to invest in a structural change is often to begin with the emerging market itself, map its entire value chain, and identify the products or services whose demand will increase much faster than the market currently expects.

The objective is not simply to find a company associated with a popular theme.

It is to understand where the next bottleneck is likely to form.

A useful infrastructure supplier generally combines several characteristics:

• Demand for its product is accelerating because of a structural—not temporary—change.

• Its product represents a small cost for the customer but is essential to the final system.

• Supply cannot immediately adjust to demand.

• The company possesses technical expertise, manufacturing capacity or customer relationships that are difficult to replicate.

• Its valuation still reflects the previous market rather than the one now emerging.

Being unknown is not enough. Being strategically necessary while still being valued as an ordinary supplier is much more interesting.

This is how we approached Nvidia and the semiconductor ecosystem. It is also how we identified opportunities in fiber optics, cooling, memory, power generation and semiconductor-testing equipment.

But identifying a promising value chain is only the beginning.

The more difficult question is whether the original thesis remains valid six months, one year or three years later.

Demand can move to a different part of the chain. A company can lose its technological advantage. Management can introduce governance risk. A strong operating story can be undermined by excessive debt or an unsustainable valuation.

A good thesis should therefore be treated as a living hypothesis—not as a permanent conviction.

Turning investment theses into portfolios

For more than two years, this has been one of the central methodologies behind Altis Research.

Instead of building portfolios around broad sectors, we build them around specific investment theses.

Each portfolio begins with a simple hypothesis about how a market is changing. We then identify the companies positioned to capture the resulting value, document the initial positions, and continue updating the portfolio as new evidence appears.

The individual stock is not the thesis. It is one possible expression of the thesis at a particular moment.

That distinction matters.

If a company disappoints but the structural trend remains intact, the correct decision may be to replace the company rather than abandon the idea. If demand migrates from one bottleneck to another, the portfolio should evolve with it. And if the evidence contradicts the original hypothesis, the thesis itself should be closed.

This is also the reasoning behind Altis Terminal.

Every portfolio on the platform is connected to an identifiable thesis. Trades can be documented with their original rationale, while Locked portfolios preserve the exact prices and dates in real time, without editing or backdating.

The goal is not only to show what someone owns. It is to show why the portfolio exists, how the reasoning evolves, and whether the market is validating or invalidating it.

Readers interested in a specific market can therefore follow several portfolios testing different parts of the same value chain. Terminal can analyze new positions, events and market data against the original thesis, making it easier to identify what has genuinely changed.

Over time, this creates something a newsletter alone cannot provide: a continuous record of an investment idea being confronted with reality.

Altis Terminal opens to Premium subscribers tomorrow

The methodology described in this edition is also the reason we built Altis Terminal.

A newsletter can introduce an investment thesis. The Terminal makes it possible to follow that thesis over time, see the portfolios built around it, examine every documented trade, and understand whether new market evidence is validating or weakening the original idea.

Never miss a great investment thesis.

Tomorrow, we will begin opening Altis Terminal to Macro Notes Premium subscribers. We have already started activating access for existing Premium and Founding members, who will receive their invitations progressively.

The platform brings together hundreds of portfolios and investment theses covering AI infrastructure, energy, semiconductors, financial systems, space, robotics and other structural markets.

More than ten active contributors have already begun publishing and documenting their portfolios on the Terminal, and more research teams and independent contributors will be added after launch.

Altis Terminal will normally cost $499 per year as a standalone product.

However, if you subscribe to Macro Notes Premium today for $300 per year, complete access to the Terminal will be included in your subscription—along with all Macro Notes Premium research.

That means access to:

• Hundreds of investment theses and research portfolios

• Transparent, timestamped trade histories

• Real-time portfolio performance

• Detailed thesis updates and market events

• Research published by more than ten active contributors

• New portfolios, contributors and analytical tools added over time

Tomorrow marks the beginning of the public opening. Subscribe today to secure Macro Notes Premium and Altis Terminal access together for $300 per year.

Upgrade to Macro Notes Premium

The next buyer of AI infrastructure

The next chapter of the AI infrastructure build-out is no longer being driven solely by American technology companies.

Governments increasingly view domestic computing capacity as strategic infrastructure—closer to energy, telecommunications or defense than to an ordinary cloud service.

McKinsey estimates that sovereignty requirements could influence between 30% and 40% of global AI spending by 2030, creating a market worth between $500 billion and $600 billion.

This is the emerging market behind what is now commonly called sovereign AI: the effort by countries to secure greater control over the compute, data, models and infrastructure on which their economies will increasingly depend.

France announced more than €109 billion of investment commitments for AI infrastructure projects. The UAE launched plans for a five-gigawatt AI campus in Abu Dhabi, including a one-gigawatt Stargate UAE compute cluster.

Japan committed more than ¥10 trillion in public support for its domestic AI and semiconductor sectors, with the objective of mobilizing more than ¥50 trillion in combined public and private investment over the following decade. Singapore has committed more than S$1 billion to AI compute, talent and industry development as part of its National AI Strategy 2.0.

These countries do not all have the resources—or the time—to develop an entirely independent semiconductor supply chain.

But they can finance data centers, secure accelerators, purchase rack-scale servers, create sovereign cloud regions and ensure that strategically important data and models remain within their jurisdiction.

That creates a new category of buyer and a new infrastructure market.

We began documenting this shift on October 17, 2024, through our Sovereign AI Infrastructure portfolio. As of August 11, 2026, the Locked portfolio had generated a total return of +199.3%.

The return is impressive, but it is not the most useful part of the story.

What matters is how the portfolio reached it.

One original position was removed when governance, legal and export-control risks changed the investment case. Its capital was redeployed into another company offering exposure to the same underlying demand with a stronger risk profile.

A cloud-infrastructure position was later sold after the financing required to fulfill its enormous AI backlog materially changed the balance-sheet equation.

The thesis survived, but its expression evolved.

In the rest of this edition, I will open the complete portfolio: why we created it, why each company was selected, which positions generated the return, what forced us to sell or redeploy capital, and whether sovereign AI infrastructure still offers an attractive opportunity after a 199% gain.

Read more

Sometimes we look for investment opportunities that receive much less attention, like the Polish banking thesis I covered in a previous edition.

Today we’re going to look at a very similar idea.

We’re going to focus on a country that rarely makes international headlines, especially compared to its neighbor Ukraine.

We’re going to talk about Romania... and banks.

In other words, two topics that sound about as unexciting as it gets.

The name of the thesis is Romanian Banking Compounding.

At first, it doesn’t sound particularly interesting.

But once you start digging into the data, you realize it’s much more fascinating than it appears.

The first reason is that Romania has long been, and probably still is today, an underbanked economy.

In other words, the country’s financial intermediation ratio was only 22.2% in the first quarter of 2026, meaning that the amount of bank credit relative to the size of the economy remains remarkably low.

That means Romanian banks are not dependent only on:

• gaining market share,

• cutting costs,

• or buying back their own shares.

They can still benefit from something much more powerful: the structural growth of credit as the country continues to develop financially.

That’s the first reason why Romania became an interesting market for us and why the portfolio we built at Altis Research around this thesis has now returned 90.7%.

The second reason is that Banca Transilvania reported a gross loan-to-deposit ratio of just 63.5% in the first quarter of 2026.

In simple terms, the bank had far more deposits than it needed to fund its loan portfolio.

When interest rates are high, those low-cost deposits become incredibly valuable.

The bank can lend or invest that capital at much higher yields without relying heavily on wholesale funding or bond markets.

The third reason that strengthened the thesis this year is profitability.

Banca Transilvania generated a return on equity of 19.54% in the first quarter of 2026.

That’s probably one of the most important numbers in this entire thesis.

A bank capable of generating nearly a 20% return on equity while maintaining a relatively liquid balance sheet can compound its book value surprisingly quickly while continuing to distribute dividends.

The market didn’t need to assign these banks aggressive valuation multiples.

Their earnings and book value simply had to keep growing.

The fourth reason is Banca Transilvania’s history as a consolidator.

The bank became the country’s largest lender by acquiring and integrating other banking businesses, then using its scale, deposit base and technology across a much larger customer network.

In other words, the market often viewed Romania as a simple macroeconomic story, while much of the value creation actually came from very specific microeconomic drivers:

• economies of scale,

• improving operating efficiency,

• a growing customer base,

• stronger distribution,

• and the ability to spread technology investments across a much larger business.

The fifth reason is the combination of earnings growth, dividends and valuation expansion, which I’ll explain in much more detail later in this edition.

So to summarize, these are the five reasons why I think this thesis is worth understanding:

1. An economy that is still underbanked.

2. An abundant deposit base with an attractive funding structure.

3. Returns on equity approaching 20%.

4. A consolidating banking industry.

5. The combination of earnings growth, dividends and valuation re-rating.

We’ll go through each of these in much more detail.

But the main takeaway I’d like you to remember is this:

Ordinary businesses can produce extraordinary returns when you understand them correctly.

The biggest opportunities are not always hidden inside futuristic technologies.

Sometimes they’re found in traditional businesses operating in markets that very few investors are paying attention to.

Today I’ll introduce the foundations of this thesis.

As always, you’ll also be able to follow it live on Altis Terminal, where the portfolio is published and updated in real time.

Every time we add a new portfolio, it requires additional integrations behind the scenes.

We connect our event-monitoring engine, AI-powered insights and signal generation system to the most reliable data sources available for that specific market.

That means adding new financial data providers, exchange-specific data feeds and specialized sources depending on the country and industry we’re covering.

Subscribe to Macro Notes Premium for $300/year and receive full access to Altis Terminal.

We’ve already started inviting Premium members into the beta ahead of the official launch on August 14.

Altis Terminal will normally cost $499/year and will include hundreds of investment portfolios published by dozens of Substack publishers.

But if you’re a Macro Notes Premium subscriber, your access is included at no additional cost.

Subscribe to Macro Notes Premium

Macro Notes Premium — Romanian Banking Compounding

In the premium section, we’ll go far beyond the headline return and rebuild the investment thesis from the ground up.

We’ll start by looking at the complete portfolio, every position, every trade and the role each bank played in generating the +90.7% return.

More importantly, we’ll answer the question that really matters:

Why did this work?

We’ll break down the five structural drivers behind the thesis, explain why Romania’s banking industry was different from most European markets, and show how an ordinary-looking sector quietly became one of the strongest-performing investment opportunities of the past two years.

We’ll also examine each company individually—including Banca Transilvania, BRD, Erste/BCR and the other holdings—to understand why they were selected, how they contributed to the portfolio and what has changed since the original investment thesis.

Finally, we’ll look ahead.

Has the market already priced in the opportunity?

Or do Romanian banks still offer attractive long-term returns?

We’ll review today’s valuations, the main macroeconomic risks, the impact of lower interest rates, and whether I would still be buying these companies today—or whether it’s time to reduce exposure and look for the next overlooked market.

Read more

They were excellent.

But today, I want to focus on just one part of the business: Starlink.

Several years ago, I answered a question on Quora:

“Will Starlink make Elon Musk the richest and/or most powerful man in the world?”

At the time, SpaceX had launched only around 400 satellites.

Yet my conclusion was already the same.

I believed Starlink would become far more important than most investors imagined.

Not simply because it could provide internet access to rural areas.

But because it had the potential to become the world’s first truly global internet network built from space.

That conviction was based largely on SpaceX’s planned launch cadence and the number of satellites required to eventually provide global coverage.

Several years later, yesterday’s earnings tell a very interesting story.

Starlink is now SpaceX’s largest source of revenue.

The Connectivity segment generated $4.29 billion in quarterly revenue, representing more than 55% of the company’s total revenue.

Revenue grew 66% year over year, while the subscriber base continues to expand at an impressive pace.

But I actually find another conclusion even more interesting.

For years, most investors viewed SpaceX as a space company that happened to be building an internet service.

Today, you could almost argue the opposite.

SpaceX has become an internet infrastructure company that also happens to operate a space business.

And ironically, its space business is what gives it its greatest competitive advantage.

No other company can manufacture its own satellites, launch them on its own rockets, replenish its constellation at such a rapid pace, and control almost every part of the value chain.

That is exactly what led me, a little over a year ago, to begin documenting my Satellite Internet portfolio.

Interestingly, SpaceX is not one of the positions.

Simply because it wasn’t publicly traded at the time.

More importantly, my thesis was never really about SpaceX.

It can be summarized in one sentence:

Satellite internet will become a new layer of global infrastructure, creating an entirely new ecosystem of publicly traded companies.

In other words, the real question was never:

“Will SpaceX succeed?”

The real question was:

What happens when a new global infrastructure begins to be deployed?

History shows that every major infrastructure creates its own ecosystem.

When railroads were built, the biggest winners were not only the railway companies.

Steel producers, locomotive manufacturers, banks financing the expansion, and coal suppliers all benefited from the build-out.

Building the infrastructure itself is often a very long-term investment thesis.

But the companies supplying the materials, components, and services required to deploy that infrastructure often benefit much earlier.

That is exactly the approach we try to take at Macro Notes.

Rather than simply identifying the future winner, we try to understand which ecosystem is emerging and which companies are likely to benefit before the market fully appreciates the opportunity.

That is precisely what the Satellite Internet portfolio was built to do.

Follow every thesis live on Altis Terminal

The best investment ideas don’t end after a single article.

They evolve over months—and often over years.

That’s why every portfolio on Altis Terminal includes the complete investment history behind the thesis: every buy, every sell, every comment, and every performance update is documented in real time, with prices and timestamps locked the moment each trade is recorded.

At launch, Altis Terminal will cost $499 per year and will provide access to more than 100 live investment portfolios covering macro, technology, public markets, and long-term thematic investment ideas…

However, if you subscribe to Macro Notes Premium before August 14, you’ll receive full Altis Terminal access at no additional cost.

In other words, for $300 per year, you’ll not only receive every Premium Macro Notes edition, but also access to the entire Terminal—including every live portfolio, documented trade history, and future investment thesis we publish.

Subscribe to Macro Notes Premium

In the Premium section, I’ll walk through the thesis in much greater detail.

We’ll look at why I chose AST SpaceMobile, Globalstar, and EchoStar, what worked, what didn’t, and why the portfolio has now returned +258.8%.

Most importantly, we’ll answer the question that matters today:

Most importantly, we'll explore how Starlink's latest results change this investment thesis—and where the next opportunities may emerge.

Read more

If you haven’t read it yet:

Today, I want to explain why I don’t think this thesis is over.

The past 18 months were only the first chapter.

At its peak, the portfolio returned nearly 700%. As of July 28, it was up +593% since inception.

I believe we’re now entering the second chapter.

Here’s why.

Every new generation of AI models requires more computing power.

For the past two years, semiconductors have been viewed as the number
one investment opportunity created by AI. That’s why companies supplying
AI chips have seen their valuations surge.

But there’s another problem developing beneath the surface.

All of these chips need to communicate with one another.

In the early days, training a frontier AI model required a few hundred
GPUs.

Today, the largest AI clusters contain tens of thousands.

Tomorrow, they’ll contain hundreds of thousands.

Every GPU constantly exchanges data with every other GPU in the cluster.

At a certain scale, the challenge is no longer simply producing more
computing power.

The challenge becomes moving data fast enough between all of those
machines.

I believe this is one of the most important shifts taking place across
AI infrastructure today.

It’s exactly why I built the Fiber Optics portfolio more than a year
ago.

This wasn’t just a theory.

It became one of the best-performing portfolios I’ve documented over the
past year.

If you’re a Macro Notes Premium member, you’ve probably followed the
buys, trims and portfolio updates as the thesis evolved.

Back then, almost all of the market’s attention was focused on Nvidia.

Yet the opportunity in optical networking was surprisingly easy to
anticipate if you spent enough time studying how AI infrastructure
actually works.

Many readers spotted the opportunity early.

Some even shared insights that helped me better understand the industry
and refine the portfolio.

The thesis was simple.

AI wasn’t only going to create explosive demand for computing power.

It was also going to create explosive demand for the infrastructure
required to move all of that computing power efficiently:

• Optical fiber

• Optical networking equipment

• High-speed switches

• Networking components

This demand is no longer theoretical. It is now becoming visible directly in company results.

Put simply:

Without networks, even the world’s most powerful GPUs become
bottlenecks.

Until now, I’ve documented these portfolios directly inside the
newsletter.

Starting August 14, you’ll be able to follow them inside Altis
Terminal.

Altis Terminal brings together more than one hundred investment
portfolios covering some of the most promising long-term investment
theses across different sectors and investment styles.

At launch, you’ll be able to follow portfolios from five independent
Substack publishers. Over time, we’ll invite more researchers and
newsletter authors to publish their ideas and documented track records
on the platform.

Access to Altis Terminal is included at no additional cost with a
Macro Notes Premium subscription.

Subscribers who join today will receive access as soon as the platform
opens on August 14.

Subscribe to Macro Notes Premium

But back to the thesis.

I believe the Fiber Optics thesis is now entering its second
chapter, and I think it could prove even more interesting than the
first.

Here’s how I would summarize it.

• AI clusters are becoming dramatically larger.

• Hyperscalers are talking more about networking than ever before.

• Network speeds are moving from 800G to 1.6T.

• Infrastructure spending is no longer focused solely on GPUs.

• Increasingly, it’s about everything required to connect those GPUs
  together efficiently.

This transition is important because growth is not only coming from the construction of more connections. Each new generation of connections is also becoming faster, more complex and potentially more valuable.

The first question was:

Was this thesis correct?

I think that question is now largely behind us.

The more interesting question today is:

If the market is finally beginning to understand this shift, where
does the next opportunity still exist?

In the Premium edition, I explain:

• Why I believe the networking bottleneck is only just beginning.

• Which companies remain best positioned today.

• Which positions I’ve already trimmed and why.

• The next catalysts I’m watching over the coming quarters.

• The biggest risks that could invalidate the thesis.

• How I’m currently positioned across the Fiber Optics portfolio.

Macro Notes Premium isn’t just access to research.

It also gives you access to the portfolios behind each thesis.

With the launch of Altis Terminal, you’ll be able to follow every
documented trade in real time, understand why each position is added or
reduced, and watch the thesis evolve as new evidence emerges.

Macro Notes Premium - The Second Chapter of the Fiber Optics Thesis

Read more