Portfolio is now up 62% over the last twelve months against 30% of the S&P 500!

It’s been a wild ride! — This is probably what best explains how I feel about the last few months.

It’s truly been a wild ride, and I am extremely happy and proud of where our portfolio has reached today after this wild ride. I’ll give you the update and talk about the strategies and investment philosophy, but first, let me start by extending my thanks to the great community we have here.

If you are a human with a conscience, writing about finance and stocks is very hard because you know that people make financial decisions based on what you write. The only way to have a clear conscience is to fully align yourself with what you write.

When something you write wins, you win; when it loses, you lose as well.

This is what I have always practiced here. If you have been in the community for a while, you have probably heard me say, “I can’t say buy because I am not holding it.” And not getting a concrete answer might have disappointed you at that point, but that’s the only way to have a clear conscience on my end.

As a result, sincerity has emerged as the default mode of operation in this community, and we have developed a strong sense of friendship. We have exchanged ideas, poked holes in our theses, and restored sanity through discussions when things got volatile.

So, thank you.

When it comes to our performance, I’ll say again what I have said a few times in the past—forget the performance, what’s important is how well your positioning played out.

Indeed, this performance could be 45% today instead of 62%, but this wouldn’t change anything in its impressiveness because what’s important was being able to position to generate this level of return.

If you remember, at the end of last year, I said two things:

• The market is overvalued.

• 2025 was hard, and 2026 will be harder because there is no single asset class that is clearly positioned to go higher.

Yet, despite all this, I also said “stay in the market.” I recommended exiting fundamentally weak or overextended positions, building a nice cash cushion, and leaving the rest in the market and allocating to opportunities as they come.

Then the war happened, and we still didn’t sell a single share.

Because our position was this:

• All our holdings were very strong fundamentally.

• In the worst-case scenario, they would go undervalued, which is never a reason to sell.

In the end, optimism won once again, and parties to the war entered a de-escalation path, so the market has come back. Then, AI optimism peaked as Anthropic revenues tripled in Q1 and compute demand skyrocketed again, driving further optimism.

Result? Most of our positions are now at or near all-time highs.

What’s even more valuable is that we have achieved this without any leverage or mass buying speculative stocks.

Fundamentals of all our positions today justify either the current valuations or way more than the current valuations, but not less. Having said this, I have to talk a bit about a dangerous trend I see on Substack and X (Twitter), under the name of “bottleneck investing.”

This is pure speculation, given a fancy name to make it look more reasonable. Bottlenecks, of course, exist, and it’s often productive to see the bottlenecks ahead and invest in them. If the bottleneck proves legit, companies operating there will likely have strong pricing power, and this will be enough to boost earnings by a lot.

Many positions in our portfolio are there because we believe they address certain bottlenecks as well. There is nothing wrong with this.

But… This doesn’t mean every price is justified or that the stock should be priced as if the bottleneck will persist forever. When this happens, the result is almost always a disaster.

Here is an example:

When you look at Corning stock, you see the last time a price movement like this happened was back in 2000. We know how it ended. This won’t be any different.

In some small-cap stocks, the situation is even more absurd.

You should be careful. If the fundamentals of those stocks were able to take them to highs, those who own them wouldn’t need to shill them every day. Those positions will be dumped on some people at some point.

We have never engaged in such speculative behaviors, and will never do.

Our philosophy is simple. We run a barbell portfolio where:

• We buy companies with durable competitive advantages, i.e., foundational stocks.

• We combine foundationals with growth positions that have strong balance sheets.

• We don’t sell unless fundamentals deteriorate or the price goes off the roof.

This has worked incredibly well for us so far:

Our portfolio is now up by 62% against 30% of S&P 500!

The following transactions took place in our portfolio in November:

• Exited 1 position.

• Opened 1 new position.

In the previous updates, I had provided my outlook for each company in the portfolio.

In this update, I’ll also provide an outlook for our new position and share my overall portfolio commentary and strategy going forward.

Let’s dive in!

🚨Our portfolio is proprietary to the members supporting the publication🚨

Members get portfolio updates every month!

Here is a 25% discount to celebrate new readers!

Valid only until Monday!

Get 25% off for 1 year

📊Here Is Our Full Portfolio!

As of today, we have 18 holdings in our portfolio.

8 of these positions can be considered foundational.

Read more

We are now up 60% in the last twelve months, against just 33% of the S&P 500.

How have we achieved this? Well, looking back, I think there are three factors:

• We let the winners run.

• We timely opened new positions trading at deeply undervalued prices.

• We waited patiently for our existing undervalued positions to play out.

Before the year started, we had closed three groups of positions:

• Those that went too high above their intrinsic value.

• Those whose fundamentals were deteriorating.

• Those that were too small to make a difference.

We rolled a portion of the proceeds from these exits over to new undervalued positions that could create alpha as they were starting from a lower base, and kept the rest to create a bigger cash pile as the market was obviously overvalued.

But we kept the exceptional companies we owned, even if they appreciated as their fundamentals were improving fast, and we didn’t see a safe selling price. In these positions, we would rather ride the downside volatility than give away the optionality.

Then, within this year, we have only opened three new positions. One of them is up by 30%, and the other two are largely flat. The good thing with other positions is that they tested their lows a few times since we bought them, and bounced back. They have now broken the downtrend and are ready to go up if favorable market conditions continue.

Another driver of alpha was the recovery in one of our most undervalued and biggest positions. We heavily increased our allocation in this name last year after it got heavily undervalued and it finally started to play out, creating alpha while the market has been largely flat YTD.

We had finished all these positional arrangements before the Iran War started. So, at the outset of the war, we didn’t have any “egregiously overvalued” position that could stress us during the downturn. Thus, we managed to do exactly “nothing” while everybody was panicking.

For me, this is the gold standard of positioning.

If you can afford doing nothing during downturns and it looks like the only logical thing is buying more if you are to do something, then the positioning was right.

That’s exactly how it was during the downturn. We didn’t have any position to stress us, and we could only consider buying more if the prices came down to our “no-brainer” zones. In the end, most positions didn’t come down enough for us to buy, given our ~20% cash position, and we ended up doing nothing.

Now, the market has recovered and recorded a new all-time high, and valuations have swung from “somewhat attractive” to “unattractive” again.

Indeed, the market is now trading at around 21 times earnings, only 200bps below where it was in late 2025:

And the credit spread is still below 3%, meaning the money is still abundant, and optimism dominates the capital allocation decisions.

This is not an environment where you find one fat pitch after another.

In this kind of environment, the best opportunities come when the qualitative conditions that led to the discounting of particular stocks start changing, and the discount looks increasingly unjustified. The market is sometimes slow to recognize these changes and to bump up those stocks to their full expected value.

I think we have one such opportunity, and it’ll be an obvious one in hindsight.

Yet I still want to keep my guard up, as the market's valuation makes me nervous, and I don’t want to reduce my cash position. This is why I will fund that position by closing a position that is probably the lowest conviction for me in the current portfolio set-up. It’s in the money, but I think the upside scenario is getting weaker.

So today, I’ll be making two transactions:

• Close one position.

• Open one new position.

This way, we’ll be able to get exposure to an additional optionality while eliminating a potential weak link in the portfolio in terms of market power.

You’ll also find the link to the portfolio spreadsheet at the end of the write-up. I’ll be updating it after I execute the trades.

Let’s get started.

Subscribe now

Here are the exact trades I am making:

Let’s start with the exit, as always.

Read more

Founder mode happens when an entrepreneur acts like he is still the business, because this is how it is when companies are young. There is nothing as “business” on the first day; it’s a founder and an idea that would hopefully make people pay.

The founder builds the product, the founder does the marketing and sales, the founder keeps the books, solves problems, and thinks of ways to make the product better.

As the company grows, the founder often assumes a more managerial position. Now he says what needs to be done, and it’s up to his direct reports how to do it.

Founder mode is on when a founder refuses to be confined and stays in the game.

When founder mode is on, impossible becomes at least feasible.

Because all the great success stories start from the impossible. Only 0.0001% of all startups become unicorns. All of them were built by some crazy people working in founder mode. Founders beat the odds.

Some of them do it over and over again. Take Steve Jobs.

Found Apple, fired from Apple, built Next, built Pixar, returned to Apple, built it again to the most valuable company in the world. He beat the odds every time.

Why did I say all these? Because if it wasn’t for the founder mode, even the thought of what Oracle is trying to do would be insane. It’s literally been impossible so far.

No business making above $10 billion in revenue has ever grown 75% annually for 5 years. This is exactly what Oracle is aiming for in its cloud infrastructure business:

It dared to undertake this; it dared to risk it all to pull this off, and it dared to let all of us know this because its founder, Larry Ellison, is still at the helm as CTO and Chairman, and his founder mode is turned on.

It doesn’t take much brains to see that it’ll be an immensely more valuable company if it can pull this off. The market struggles to believe it.

Can he pull this one off? This is what we’ll dive deep into today.

But.. Let me say this for now—before Larry Ellison built Oracle, there was no relational database company called Oracle. With founders, it’s always from 0 to 1.

Let’s get into it.

What are you going to read:

1. 🏭 Understanding The Business

2. 🏰 Competitive Analysis

3. 📝 Investment Thesis & Risks

4. 📊 Fundamental Analysis

5. 📈 Valuation

6. 🏁 Conclusion

Subscribe now

🏭 Understanding The Business

Great companies that last for decades can do so because they are often rebuilt multiple times, catching the next wave each time.

And nobody is better at rebuilding companies than their founders…

Steve Jobs built Apple twice.

He built it around personal computers in 1976, and rebuilt it around consumer electronics in 2001 with the release of the iPod. Apple is great because it has withstood the ravages of time and adapted to changes that have destroyed many.

If he were alive, he would surely be rebuilding Apple for the next wave. But most of the great entrepreneurs born in the 1940s—1950s and rode the post-war investment boom, creating great companies in the 1970s—1980s, are either dead today, like Jobs himself, or have long been on the sidelines like Bill Gates.

But one man is standing, Larry Ellison, and he is building Oracle for the third time.

The first one was in 1977. It was 10 years after Larry Ellison drove to California after dropping out of two colleges. He was academically gifted but temperamentally allergic to authority. He couldn’t do well in a top-down academic hierarchy.

When he drove to California in 1966, he had no degree, no plan, and no money, but had a working knowledge of computer programming picked up during his brief university stints. For the next decade, he jumped from one technical work in the Valley to another.

In 1977, he was working at Ampex, where he helped build a database for the CIA in a project code-named “Oracle.” In mid-1977, Ellison and two colleagues from Ampex decided to create a company. They called it Software Development Laboratories (SDL). It was a generic name because they didn’t know what to do; they didn’t have a product.

Then Ellison came across a 1970 paper by Edgar Codd, a British computer scientist working at IBM's research labs in San Jose. It was titled “A Relational Model of Data for Large Shared Data Banks.”

What he was proposing was mind-numbingly simple, yet novel.

Data was traditionally stored in flat files, where records were organized by category in pure text format. This was basically the digital version of the file cabinets. Data manipulation and retrieval required custom coding.

What Codd was proposing was storing data in relational tables, where each data point is written in a row and gets a unique ID, while columns correspond to attributes. These IDs create relations when used in different tables.

Imagine you have three tables. One defines the products, where each product gets an ID corresponding to its attributes; one defines customers; and one records transactions. In the transactions table, you no longer need to show all the attributes of products and customers who bought them. You just use the IDs, and you have it all:

IBM sat on this. Perhaps they didn’t even notice the importance of the paper, as their research output was pretty large at the time, the same way Google was late to capitalize on their own AI research around transformers.

But Ellison understood it at the moment, and he thought this would be the default way of storing data. IBM had also published the specifications, so there was nothing stopping a small firm from building a commercial product.

His other co-founders, Bob Miner and Bruce Scott, started writing the code, since they were technically stronger, and Ellison began selling even before the product was finalized.

He remembered that the CIA wanted a database, and he himself worked in the Oracle project at Ampex. He leveraged connections he made there and convinced the CIA to give them a contract for a relational database. They made the product and named it “Oracle.”

They went beyond the CIA and commercialized it in 1979. The demand was insane. They changed the company name to Oracle in 1982 to match the product everybody knew, and went public in 1986 at a $270 million valuation.

This was the first time Ellison built Oracle on a relational database.

In the early 2000s, he had to rebuild the business. It was forced by the technological changes underway. The next wave.

At the time, the enterprise software was having its boom. Computers for each officer had become standard in most businesses across the developed countries, so everything was running on enterprise software. Resource planning, customer relations management, bookkeeping, etc., everything.

It was actually a good thing for Oracle because these applications needed databases underneath, and most enterprises were using Oracle as the database.

But there were two problems:

• Database business was maturing while software was growing skyrockets.

• Growing footprint of enterprise software players was a business risk for Oracle.

If those software players fine-tuned their software to work best with another preferred database, they could even build this themselves; Oracle would lose business.

So, vertical integration by expanding to enterprise software was a strategic mandate for Oracle, not an option. If it pulled it off, it wouldn’t just secure its legacy business but also create a new source of substantial revenue.

For Oracle, the limit wasn’t money but time. It had giant cash flows from its database business, but it needed to act fast. Thus, instead of building, they acquired, embarking on one of the most ambitious roll-up sprees in tech history, one that everybody thought it would fail.

It didn’t fail. It worked out phenomenally.

Oracle acquired over 50 companies between 2005-2010, most of which were enterprise software like PeopleSoft (HR & ERP), Seibel Systems (CRM), Hyperion (finance), etc.

As a result, Oracle transformed into a gigantic vertically integrated provider of software and infrastructure:

This was the new Oracle, maybe version 1.5, but not 2.0.

It was vertically integrated, spanning application and database layers, as young in 2010 as it was in 1999. Yet, it had to change again, before it squeezed the juice out of its new structure. This time, it was an absolute necessity.

In 2006, Amazon changed how digital companies were built and run by launching a public cloud service within Amazon Web Services (AWS).

Before AWS, every startup needed to buy its own servers and host their applications and databases within them. This came with a significant hardware overhead. Public cloud service allowed them to rent servers online and host their databases and applications in those servers.

This changed the game not just for newcomers, but also for the incumbents like Oracle, as it brought a completely new distribution model.

Before the cloud, software generally worked on a perpetual licensing model bundled with regular maintenance services. With the cloud, companies could now offer their software on a subscription basis, and customers would get the new versions automatically.

That changed the game. Customers loved it, and cloud native software competitors like Salesforce and Workday started to eat away at Oracle’s market share.

That wasn’t the only threat Oracle was facing. Its legacy database business was also under threat because all the new big companies like Airbnb and Uber were built on the cloud, so they were using native database services that their cloud players were offering.

If Oracle couldn’t adapt, it would have 0 exposure to new giants and would slowly die.

They built Oracle Cloud in 2014, which got a complete overhaul in 2016. The new version was tightly integrated with the Oracle database, offering the ability to rent physical servers (bare metal), not just virtualized containers, and it was priced lower than AWS. In 2019, Oracle also implemented the multi-cloud strategy, which is basically the distribution of the Oracle database through other cloud providers as well.

By 2020, Oracle had largely adapted to the cloud. Its software was being distributed both through traditional licensing and its own cloud, while its database was distributed through traditional licensing and its own cloud, and Microsoft’s Azure.

This was Oracle 2.0—the survived and thrived through two decades of transformation, first adapting to software and then integrating cloud.

Adapting to the cloud didn’t just secure Oracle, it also put it in a position to capitalize on possibly the biggest technological revolution since the internet—AI.

Enter Oracle 3.0.

Oracle was a latecomer to the cloud. That generally means disadvantage, but sometimes, it’s an advantage, especially when something happens, and incumbents will need to adjust to capitalize on that. In that case, latecomers, with their newer stack, can act faster and capitalize on the opportunity.

This is exactly what’s happening with Oracle Cloud Infrastructure (OCI).

To his discredit, Larry Ellison was late in seeing the transformative effect of cloud computing. In 2008, he said:

“The interesting thing about cloud computing is that we've redefined cloud computing to include everything that we already do... I don't understand what we would do differently in the light of cloud computing, other than change the wording of some of our ads.” — Larry Ellison, 2008

This is why Oracle was late to the cloud; it launched OCI 10 years after AWS.

But one of the common properties of great men is that they draw lessons from their mistakes and use them to improve their vision. Ellison had drawn his lessons from underplaying the cloud, so he didn’t repeat it when ChatGPT was launched in 2022.

He immediately envisioned that LLMs would be incorporated into almost every technology, and we would need a lot of compute power to run them. We would need:

• Enormous GPU clusters

• Ultra-fast networking

• Massive power and cooling

That kind of computing required specially designed data centers and cloud infrastructure to allow customers to manage it.

He knew that the legacy players like AWS, Azure and Google Cloud were built for general-purpose computing. Also, these companies were now run by professional managers, whereas Ellison still directly controlled Oracle and owned over 40% of shares. So, he knew he could act and deploy the capital faster than them.

This was his opening to catch the next edge.

Oracle started aggressively deploying capital and doubling the capex in 2023, which led to its first OpenAI deal. This signalled to the whole industry that Oracle was capable. Over the next two years, it expanded its partnership with OpenAI and added new customers like XAI, Cohere, and META.

Result? Capex skyrocketed, of course. It now expects to spend over $400 billion through 2029. Cloud revenue expectations, naturally, surged as well:

For reference, Oracle’s whole revenue in 2023 was $50 billion. This is probably the most ambitious transformation project a company of the size of Oracle has ever embarked on.

This is Oracle 3.0, the third time the business is being built by its founder. At the age of 81, his founder mode is still turned on. This is why this article opened with reference to founder mode. Do you think the management would bet the whole company on this if it weren’t for the vision and pushing of Ellison? I don’t think so.

Step by step, he turned a database company into an ecosystem of complementary tech businesses that are also complementary to each other—database, enterprise software, and cloud infrastructure.

Can he pull off Oracle 3.0 as well and make Oracle into one of the largest AI infrastructure businesses in the world? It’s a very narrow corridor.

Oracle’s ability to pull this off depends on being able to actually spend and build the infrastructure, which depends on them keep generating cash flows. But these cash flows, especially the part that comes from software, are under threat as AI is eroding software moats. Then it needs to outmaneuver the competition, other cloud providers.

Can it? It’s hard, but there is a path. Let’s dig.

🏰 Competitive Analysis

Chris Hohn and his hedge fund TCI broke the hedge fund earnings record last year. He is one of my favorite investors, and once asked about his strategy, he said “very few things matter, 90% of the things are noise.” When the interviewer asked what matters, he said, “competition matters.”

He then opened it up: “Competitive advantage has two elements—resistance against disruption and resistance against competitors.”

Indeed, if the company has these two, it’s very hard to lose money on it, provided that you don’t overpay for the stock initially. If it has ample growth opportunities on top of this, you can make significant money over time.

So, we need to evaluate Oracle’s business ecosystem from this perspective. Let’s look at the snapshot first:

The majority of Oracle’s revenue comes from enterprise software applications and database franchises. Services and hardware are byproducts of these businesses, and they aren’t critical to the broader business at this stage. Software and application cash flows fund the cloud infrastructure unit, which is the current growth engine.

Database segment is perhaps the most straightforward one and doesn’t require much deliberation. Oracle is the most popular relational database in the world, and it controls 30% of the market according to DB-Engines Rankings.

Databases are deeply embedded in enterprise workflows, and they are very sticky. There are very few things a company can gain by changing its database provider, and it comes with a nasty process. Migration could spawn many problems, and data loss during migration constitutes a strategic risk.

As almost all businesses will need some type of database regardless of how AI transforms the application layer, we can say database cash flows are safe, and we could assume it to grow at least as fast as the market, given Oracle’s dominant position. Growth will be around 10% annually by 2032, according to Market Research Intellect, and it’ll be driven by surging enterprise data volumes thanks to AI.

Enterprise software segment has a bit more nuanced situation. The root cause of the complication is the risk of AI making software companies redundant, eating away at their margins first and then completely commoditizing them.

Here, we have to make a distinction between consumer and enterprise software. Yes, AI will commoditize consumer software, and it’s already happening. This is because there is generally no embeddedness and system dependencies in consumer software, so switching barriers almost don’t exist, and customer inertia is low.

Enterprise software, however, tends to be embedded in workflows. Especially the foundational level enterprise software like ERM, the kinds that Oracle sells, have many system-level dependencies, thus switching barriers and customer inertia are way higher.

This prevents each company from developing its own foundational enterprise software, for now. However, it surely undermines the industry’s pricing power. Incumbents can’t price as aggressively as they used to because development costs are plummeting.

So, the basic proposition here is that enterprise software will be safe as long as the cost of development, maintenance, and migration doesn’t drastically drop below what they are already paying. This won’t happen soon because inference is still costly, and you still need to employ top-level engineers to build, deploy, and maintain the systems. Inference expenses alone would be way beyond what many customers are now paying for their software.

Companies are already struggling to stay within their inference budgets:

Thus, though possible, mass replacement of already embedded foundational enterprise software won’t be economically feasible soon. Given high switching barriers stemming from system dependencies, staff education, and inertia, we can also say Oracle’s software stack is protected against competitors. So, cash flows from this segment are also safe for the short and medium term.

Now, let’s turn to cloud infrastructure.

This is the best type of market you could ever find:

• It’s growing very fast thanks to the heavy compute demand of AI.

• It’s an inherently sticky business.

It’s sticky for two reasons.

First, demand is recurring. AI labs always need the capacity to train new models and serve existing models on the inference side. Thus, if your cloud provider is satisfying your needs, there is almost no reason for you to dump it. If you fall short of computing for some reason, you’ll lose market share to competitors.

Second, it comes with workflows and system dependencies. Once you build the business, workflows, and dependencies, switching means operational risk. Even if you are natively multi-cloud, dumping mega-watts, or even giga-watts of instances and reallocating that capacity to another provider is an operational hustle. Why would you do that as long as it satisfies your needs, provided the price is more or less the same?

You wouldn’t. Thus, competition in this market becomes pretty much muted in this market after you land the customer. Much of the competition occurs to land the customer. Given that cloud business is not a mystery and all the primary providers can satisfy the needs of the customers, the decisive factor in customer acquisition is how fast you can bring large volumes of capacity online and offer your customers.

This is where founder mode becomes relevant again.

Ellison is basically betting the company on this business, so they are moving way bolder and faster in responding to demand. As long as there is demand, they are in the pole position with Amazon, Google, and Microsoft triopoly to capture it, and maybe a step ahead because their decision process is faster, as the founder is still at the helm. And once they capture it, competition becomes muted.

In short, we are looking at a business that doesn’t have much trouble with competition.

• The database business is dominant, and it’s not under a threat of disruption.

• Software business is inevitably threatened by AI, but the short- to medium-term is safe.

• On the cloud side, the competition happens for customers and becomes muted after landing them. However, the demand is so strong right now that nobody will feel competition for customers in the medium-term. They’ll sell as fast as they build.

So, we don’t need to worry about competition, and when the competition is out of the way, it all comes down to two things—the market development as expected, and execution.

Share

📝 Investment Thesis & Risks

Oracle has always been a good company, but we wouldn’t have been talking about it as an investment opportunity if it weren’t for the cloud business and the immense demand for compute. This is where the whole investment thesis depends on.

Market research firms estimate that the global cloud computing market will reach $5 trillion by 2035:

Thus, the thesis is simple:

AWS targets $600 billion in revenue in 2035. Even if Oracle can reach half of AWS’s size, we are looking at $300 billion in revenue only from OCI, which generated just $18 billion last year.

Up until recently, there were two main risks to this thesis:

The first one is the reliability of the demand. Most of this demand will be driven by AI, the training and inference needs of AI models. If you have been following this publication for a while, you know that I have been pretty skeptical about AI labs’ ability to scale revenue as they forecasted. I no longer think that.

The latest generation of models has reached a set of capabilities where they can now be deployed as full-time workers, instead of being used just as productivity boosting assistants. They are now so good at some tasks, like coding, that the productivity barriers that we used to face no longer exist. The only limit is imagination.

As a result, the demand has exploded. Anthropic’s revenue ramp from $10 billion in January to $30 billion in March proves this:

This is because they are now able to do things that they weren’t able to do before, so people can actually be productive with them now. Not just enterprises, everybody.

And this is just the beginning. We are looking at a general-purpose technology, which means it will help spur further innovations and products. Indeed, almost all electronic devices we are using today can be made smarter by incorporating AI. This will be made gradually, and the demand will jump at every step. They’ll also enable new products that we can’t even imagine today, and they’ll drive the demand even further.

Take the internet as an example. It’s been 23 years since it was commercialized, but internet traffic still grew 19% in 2025. It’ll be the same with AI; we’ll see decades of demand growth through capability expansion and new product discovery. So, I think broader demand risk no longer exists. There’ll be more demand than we can imagine.

The second main risk to the thesis was the concentration of the $553 billion backlog in commitments from OpenAI.

Oracle expects to spend over $400 billion in capex through 2029 to expand its cloud business. Given that its free cash flow before the investment supercycle began was around $11 billion, the vast majority of this buildout will be funded by debt. And its ability to pay this debt depends on OpenAI’s ability to meet its commitments.

OpenAI’s ability to pay, on the other hand, depends on two factors:

• OpenAI’s own revenue ramp.

• OpenAI’s solvency.

Both of these were in question until recently because OpenAI guided for $200-$250 billion in revenue in 2030. Despite this growth, it projected $115 billion cash burn through 2030.

This scenario got largely derisked over the last three months.

First, Anthropic’s growth showed that OpenAI’s projections aren’t bullshit. Demand is there. Second, OpenAI recently raised $122 billion and will likely do an IPO in the next 24 months, and will likely raise at least another $50 billion. The solvency risk no longer exists if OpenAI doesn’t fail miserably.

So, the cloud growth thesis now has the best outlook ever since Oracle kicked off this pivot.

This is why I said “the time has come” in the title. Cloud growth thesis has never been as derisked as it is now. Main external risks have largely vanished, and the faith of the thesis now depends mostly on Oracle’s own execution.

The time has truly come.

📊 Fundamental Analysis

➡️ Business Performance

We need to see Oracle’s performance before and after they doubled down on cloud growth in 2023.

Its performance was already solid before 2023 for a software+database company at its size. It was consistent, but the growth was, of course, problematic. Post-2023, we see a meaningful acceleration in growth, thanks to the cloud pivot:

The revenue jump in 2023 was largely driven by the accelerating distribution of enterprise software through Oracle Cloud. Though this was largely a one-time bump as enterprises sought to modernize their digital infrastructure post-COVID to adapt to changing work environments, such as the rise of remote work.

Thus, we saw that the applications cloud services revenue jumped by 32% in 2023, driving the overall growth in the business. Cloud infrastructure revenue grew by only 6% in that year. However, post-2023, cloud application services rapidly went back to single-digit growth while cloud infrastructure accelerated every year:

In short, we are looking at a company that is already strong in its core and aggressively capitalizing on a new opportunity. Cloud is not a distant story for Oracle; it’s become the company’s main growth engine since 2023, and it’s just the beginning. There are no red flags here, only hope.

➡️ Financial Health

This is where most people who consider investing in Oracle turn back. The headline numbers look scary. Oracle has over $108 billion in financial debt while it generates just above $27.7 billion in EBITDA and carries $38.5 billion in equity:

However, when we lean into details, it’s much less scary. Oracle’s principal debt, as of the last 10-K, was as follows:

• Fiscal 2026: $7.3 billion

• Fiscal 2027: $5.7 billion

• Fiscal 2028: $10.1 billion

• Fiscal 2029: $2.0 billion

• Fiscal 2030: $7.3 billion

• Thereafter, all the way to 2065: $60.5 billion

The average coupon rate on this debt was roughly 4.5-4.7%. So, Oracle’s interest payments will look as follows:

• Fiscal 2026: ~$4.0 billion

• Fiscal 2027: ~$3.9 billion

• Fiscal 2028: ~$3.7 billion

• Fiscal 2029: ~$3.3 billion

• Fiscal 2030: ~$3.2 billion

• Thereafter: roughly $55–60 billion

Of course, this doesn’t include ~$15.5 billion in new debt raised since May 2025, but the overall picture is simple. At the current state, Oracle’s total debt service won’t likely exceed, on average, $15 billion annually by 2030. So, if the cloud revenue ramps as expected, assuming 30% EBIT margin, which is similar to AWS’s margin profile, debt repayment won’t be an issue.

In the last earnings call, the management said they don’t need to raise any more money in 2026 to fund the year. Bloomberg expects FY 2027-FY 2029 to create another ~$40-$50 billion funding gap, and free cash flow will be around +$20 billion in 2030.

Even if that gap is bridged solely by raising additional debt, it’ll add only around $2.2-$3 billion in annual interest payments, assuming similar coupon rates to its existing debt, and it won’t need to raise more beyond 2030 as +$20 billion starting FCF is enough whether they want to leverage down by paying the principals as they hit maturity or refinance and roll over.

So, numbers aren’t as risky as people think. The fact that Fitch also affirmed Oracle’s credit note as investment grade (BBB) in February illustrates this. The main risk to its financial planning was never the numbers themselves; it was whether the revenue would ramp as expected.

As I said above, given the current demand outlook and OpenAI’s improved fiscal visibility have largely secured Oracle’s expected revenue ramp going forward.

To sum up, we are looking at a business that is in the middle of arguably the biggest corporate pivot in history. Performance numbers show that the plan has worked so far. The balance sheet is leveraged indeed, but it’s actually less risky than it looks on the surface. Risks were never in their numbers, but in their thesis and dependencies. These risks have substantially abated lately, so their position is also stronger than ever.

📈 Valuation

I see many detailed valuation projections for Oracle, and they all suffer from the same defect—the numbers are just too speculative.

Some assume a $110 billion peak capex, and some others assume $65 billion, some assume an EBIT margin around 20%, and some others around 35%, etc. I don’t think it makes sense to create a detailed model with that much speculation. It only makes something too speculative look much less speculative, which is meaningless.

As Aswath Damodaran says, the best valuation is the simplest one with the least number of dependencies. In the end, valuation is a scenario, not an objective work. We create a scenario and value it.

If you have foreseeability in the business and financials are strong enough to guarantee survival, the only thing you need to do is create a conservative scenario and run the math.

What would be a conservative scenario for Oracle?

The management is guiding for $224 billion in revenue in 2030 and expects $166 billion of it to come from OCI. To be conservative, I assume it’ll be 20% less than the management’s target, which gives us $180 billion revenue in 2030.

Microsoft has the most similar business mix to Oracle, with software and cloud. Their net margin has been around 35% for the last decade. I assume Oracle will have a lower net margin due to its faster and recent cloud ramp and will end up with around 25% net margin.

This gives us $45 billion in net income in 2030. Assuming a 20x exit multiple, we’ll have a $900 billion company. Discounting it back to today at 10% annual rate gives us around a $615 billion company.

It’s currently valued at $503 billion, implying 22% discount to our fair value.

🏁 Conclusion

Oracle is trying to pull off arguably the biggest corporate pivot of all time. No public company generating above $10 billion in revenue has ever grown as they are trying to do with their cloud business.

I have been very skeptical about their efforts initially, not because of how the financial projections were, but because of the assumptions those projections relied on.

The biggest risk to Oracle's thesis has been the overall demand for AI models and services and OpenAI’s ability to ramp revenue fast enough to meet its obligations to Oracle. These two risks have now substantially abated as Anthropic’s recent revenue ramp proved that the demand is there, and the frontier AI labs can grow revenues incredibly fast if they have the compute.

Valuation is also attractive even if we assume substantially lower targets than those of the management. In short, we are now at a point where the risk/reward of betting on Oracle’s pivot is at its most attractive point since it embarked on this path.

Is it at the “take a large position” point? Of course it’s not. The risk is still serious. Something may happen tomorrow, and the demand projection for AI could change completely. But, it’s attractive enough to take a 3-5% position. After all, if the management’s guidance materializes, the current price implies 8x earnings in 2030 for a hyperscaler business that will still be growing by +20% beyond 2030.

Add all these above the fact that this pivot is driven by its founder, who managed to grow the business through major technological transformations for the last 6 decades, and the odds suddenly look better.

I’ll start with a 3-4% position, and observe the execution and demand. If they give further confidence, I’ll grow the position.

Hope this helps.

Share

Subscribe now

That’s all friends!

Thanks for reading Capitalist-Letters!

Please share your thoughts in the comments below.

👋🏽👋🏽See you in the next issue!

“As fast as we put the capacity in, it's being consumed.” — Andy Jassy, Amazon CEO, said this in Amazon’s Q1 earnings call last year.

Almost a year later, after he said the words above, we read in his annual letter earlier this week that:

• Tarinium 2 capacity was sold out.

• Trainium 3 is almost fully subscribed.

• A part of Trainium 4 is already reserved.

Not just the capacity they have connected, but also the capacity they are planning to connect is being bought out.

It’s not that different for other could players as well. Nebius missed the consensus revenue estimate last quarter because it ran out of capacity to sell. Satya Nadella said Azure will double its capacity over the next 2 years, and Sundar Pichai said in a podcast that they could spend even $400 billion in capex if hardware supply weren’t constrained.

They know they could sell the capacity as much as they put it out. When you turn to buyers, you see pretty much the same lingo. OpenAI’s CFO, Sarah Friar, recently said: “I do spend a lot of time trying to find any last-minute compute available here in 2026.”

Normally, I would be very skeptical of this much alignment in an industry and think that it should be hype-craft. It’s not, that’s the reality on the field.

Last week, we got one perplexing data—Anthropic ARR increased from $9 billion in early January to $30 billion at the end of March. It basically tripled in just 8 weeks.

Anthropic’s increased API outages following the release of Claude 4.5 show how exponential the demand was.

The demand was so excessive that Anthropic had to reduce the default thinking level of the model to be able to serve everybody and solve frequent shortages. This is what caused the retardation of Claude that we felt over the last few weeks; it was dictated by a structural shortage that couldn’t be solved quickly.

The lenses we see AI through suggest that this is just the beginning. We are in the early innings of a demand shock that General Purpose Technologies experience when they reach minimum viable capabilities. This stage determines the direction and scale of the medium-term infrastructure buildout, shapes business strategies, and picks winners from losers.

We’ll explore what this means, where it leads us, and what the implications are for investors.

Subscribe now

This Is Just The Tip Of The Iceberg

The coming demand is so insane that what we see today is nothing compared to it, maybe not even the tip of the iceberg.

The economic concept I have been using to make sense of AI and where it’s heading is General Purpose Technologies (GPT—funny coincidence, it’s not the same GPT as in ChatGPT). The concept basically suggests that there are a handful of technologies that are inputs to an extraordinarily large number of other innovations.

Think about the fridge and chips.

Fridges, regardless of how transformative they are in human life, don’t contribute many other innovations. Chips, on the other hand, are the backbone of the modern world. The vast majority of all electronic products have chips, and they power many digital technologies like the internet, software, etc.

As a result, the long-term demand curve of GPTs looks way different than other technologies. While other technologies flatline in demand, GPTs keep rising since their use cases expand as they develop, spurring innovations and developments in the existing ones.

Again, if we follow the microchip example, we see that as chips themselves developed, other technologies that use them, like computers, also developed, and new versions of the computers required even more chips. This development also enabled new use cases, like mobile phones, further increasing the demand.

Our thinking is that AI is one of those general-purpose technologies that will infiltrate many products we are using today and spur the development of new ones.

Thus, the demand for AI will increase secularly over the long term. AI itself will also develop, enabling new use cases and new products, which in turn increase demand again. A self-perpetuating demand cycle, just as we see in electricity, chips, and the internet.

This is intuitive, as we can easily imagine that most of the electronic devices we are using today will integrate AI. Computers, mobile phones, cars, home systems, etc. This is where it will eventually go.

Each new generation of GPTs comes with enhanced capabilities that spur investment in complementary innovations, increasing the demand for a GPT. Not each generation is equally powerful in unlocking new use cases, innovations, and thus demand.

There is a point where a GPT reaches a set of capabilities that makes deployment into many areas and products feasible. I call this “minimum viable capabilities.” When a GPT reaches its minimum viable capabilities, it’s deployed beyond the early use cases that were natural and easy fits, i.e., the transformative power of a GPT hits at scale.

This is where demand explodes.

Think about chips.

Jack Kilby created the first working integrated circuit in 1958, which was roughly the size of a paper clip. It contained just a single transistor.

The first commercial device based on this concept, the Type 502 Binary Flip-Flop, was introduced in March 1960 and priced at $450 each, about $4,800 in today’s money for a single chip that did almost nothing by modern standards. It was a narrow application that was a natural fit for the narrow capabilities of the first version of a GPT.

Early applications remained very limited and were focused on the cost-tolerant industries like aerospace and military. Chips reached their minimum viable capabilities for mass deployment somewhere between 1971 and 1975, as Intel developed its commercial microprocessors.

The first one was the Intel 4040, released in 1971. It was a 4-bit processor that could only run from ROM and was intended for a single-purpose use by a Japanese calculator company. Intel 8080, launched in 1974, was probably the main “minimum viable capabilities” moment, as it was small, cheap, and programmable enough to be embedded in everything.

Total semiconductor sales jumped from $2 billion in 1970 to $10 billion in 1979 as chips were now going into everything.

Look at all transformational GPTs, and you’ll see a similar pattern. Steam engines spread after the invention of the Corliss Steam Engine, which allowed automatic control of steam admission under varying loads; the electrification took off after Tesla discovered the A/C transmission; the modern internet emerged in 1983, but it took off after Tim Berners Lee introduced the World Wide Web, etc.

All those milestones expanded the capability set of those GPTs to a degree that large-scale adoption and following productivity growth were made possible.

We are at a similar “minimum viable capabilities reached” moment for AI models.

Early versions of the models we are using today were impressive compared to chatbots before, but they were actually useless in creating something of economic value.

They worked best when they were asked a question already in the dataset, but were pretty useless for real creation. They hallucinated a lot, and even the code written by them was largely broken.

This is no longer the case.

The tool use has improved so much that even a standard model today can use several tools from search to code execution, which enables them with tasks that require going above and beyond their training set. The core capabilities have also developed substantially as we focused on things that AI can do very well, like coding.

As a result, models today can execute complex tasks agentically, performing hours of complex work like coding at an elite level. Thus, we have moved from having those who could be considered as “assistants” at best to having “virtual employees.”

Most people don’t understand how much of a demand this unlocks. It’s not just skyrocketing software new creation and deployment, it’s also forcing firms to rewrite all the software.

The vast majority of our software installed base was created by humans over the last 30-40 years. They are filled with security weaknesses, bugs, etc. If these weaknesses aren’t patched, new models will become a pure cybersecurity threat as they’ll be able to easily break most of the software we have now.

Coding is just the most obvious example. A whole ecosystem of tools and connectors has now been developed, allowing AI models to use applications autonomously. It’s not just basic software like Excel, Word, and PowerPoint; they are using complex apps like Figma, Adobe, etc., and they can come up with satisfactory work.

Marketing another area that is being quietly transformed. Most talking head UGC videos you see today on Instagram ads are created by AI, static ads are 90% AI generated, so are the copies of all those ads.

We have reached the minimum viable capability set, and the demand is exploding. The fact that Anthropic’s revenue tripled after the release of Opus 4.5 in late 2025 illustrates this:

This level of demand clearly exceeds Frontier Labs’ projections as Anthropic had to dumb down Claude and OpenAI had to discontinue Sora to repurpose the compute into higher demand areas, like Coding. They can’t serve everybody, and even those they can serve don’t get the full potential. It’s just insane.

This is just the beginning, as we are just at the first or second generation of models that could be considered to have minimum viable capabilities. When we get the next generation models, deployment will again accelerate as new capabilities will be put into productive use.

The demand shock we are experiencing has one substantial consequence for the frontier AI labs, and thus the whole ecosystem: Those who secure the compute will take market share.

Thrive Or Die By Compute

All general-purpose technologies are doomed to be commoditized. This is the nature of the concept. Such a large-scale diffusion can only happen with commoditization.

Electric is a commodity, the internet is a commodity, most of the chips are commodities, etc. Look at all that could be considered a GPT, and you’ll see that it’s a commodity.

The natural result of this is that when one provider is supply-constrained, the demand easily shifts to another. This is exactly what we are seeing now. Anthropic has overgrown OpenAI thanks to its narrow focus on coding and enterprise use. The demand exceeded their ability to serve while running the models at full potential. They dumbed down the models to serve everybody. People felt it and started to switch to OpenAI’s Codex from Anthropic’s Claude Code.

This will always be the case until full commoditization, where the market forces will force every supplier to offer similar performance and usage at more or less the same price. Thus, the provider that has more compute has a significant advantage at this stage.

It’s not just in terms of the ability to serve customers and capture market share. Development of new products, thus capturing new demand, is also limited by the computing power you have.

Take a look at this leak:

Anthropic is developing a Lovable-like vibe-coding feature. That’s the most natural thing in the world, as Lovable also runs Claude under the hood. They’ll come for all those AI software that are basically wrappers of their foundational models. The only thing that limits them is compute.

So, securing more compute is a strategic priority for all frontier labs now. They will live or die by their ability to secure compute. The one that can secure more compute can distribute its frontier model more widely and bring new products to market faster, accelerating its revenue growth, which enables it to access more compute.

Those who fall so much behind will lose their ability to catch up unless we see decisive flattening in model performance soon. They won’t be able to sell models as they’ll be way behind the SOTA, so they won’t get more compute to improve, and this cycle will eventually lead to their downfall.

So, basically we are looking at the following picture:

• The capabilities have hit a milestone that enables deployment at scale and scope.

• Demand is already skyrocketing, and we are seeing just the tip of the iceberg.

If you are an AI lab, the strategic implication of this is that you should buy every bit of compute you can find. You know you’ll sell, the demand is there, and you are only limited by your ability to serve. At this stage, underestimation of the compute you need is costlier than overestimation.

While their demand trajectory looks vertical, the compute supply is drying up. Half of the data center projects planned for this year will be cancelled or delayed as electrification won’t be possible due to the bottlenecks in electric equipment supply.

Meaning? If you have already secured the site, power, and construction is on time, you are golden. You’ll be able to sell faster than you bring the capacity online. This is the most important implication for us, investors.

Share

AI Infrastructure Trade Is Now Derisked

This is the most important consequence for investors.

Up until recently, we have been skeptical of near-term overbuilding as all hyperscalers have been pressing hard to expand their capacity, and most Bitcoin miners have pivoted to capitalize on the opportunity. Oracle is a striking example because it has transformed itself from a software company to a hyperscaler and put itself under a lot of debt to pull it off.

So, the faith of these investments was effectively tied to how much demand frontier models would generate. Oracle currently has around $553 billion of cloud backlog, and OpenAI is responsible for over $300 billion of it:

As a result, CDS on Oracle debt spiked and reached record highs, sending the stock down by 50% from its highs. This is why it’s currently trading at just 15x 2028 earnings, making it the cheapest hyperscaler in the market.

Earlier this year, Oracle CEOs said OpenAI wasn’t as big a risk as the market thought because they would be able to sell to other players. That didn’t satisfy the market because we were still suspicious of the broader demand as well.

It’s no longer the case as I explained above.

The technology has hit an inflection point, and we are just in the early innings of the deployment at scale and scope. Demand is skyrocketing, and AI labs’ ability to generate revenue is only limited by the amount of compute they have.

Thus, AI labs will race to secure more capacity. As they turn capacity into revenue, their ability to raise money also enhances, which further derisks investments like Oracle.

They still have to show the execution at the unit economics level, but that’s a more manageable risk profile than the one that is tied to external demand factors. However, the market is yet to acknowledge this, and many of those compute plays are still trading at a significant discount to their potential, most notably Oracle.

Thus, I think the infrastructure plays, Oracle in particular since it already has the commitments, are attractive and investable. I don’t know when, but I know it’s a matter of time because the market recognizes this enhanced risk profile and bumps up those stocks.

Closing

I see the AI through the lens of GPT literature.

Every GPT hits a critical capability milestone that accelerates its diffusion, and thus skyrockets the demand. I think we have been waiting for this moment for AI since ChatGPT was launched in 2022.

We didn’t know whether AI would ever get there, fulfill its promise, and become a transformative GPT. That also created skepticism for the demand and investments that relied on the assumption of skyrocketing demand.

With the latest generation of models, we reached that milestone.

Models can now be deployed in many tasks and create real economic value at scale, which was something they were incapable of before, probably by late 2025. Buyers feel that, and the demand is skyrocketing. This is just the beginning. We have just got there, we’ll see an insane amount of demand ahead as AI is deployed at scale and scope. AI labs’ ability to serve this demand is limited by their computing capacity, which is getting scarcer due to bottlenecks in the supply chain.

As a result, it’s a strategic imperative for them to buy and reserve every bit of compute they could find. If they can’t serve, demand will shift to competitors that have the compute capacity and thus can serve.

This derisks compute plays that invested a lot in building infrastructure, like Oracle. The market still keeps this discount as it’s always slower in acknowledging qualitative changes in the environment. I don’t know how long it will take, but at some point, the market will recognize this and bump these stocks up.

We are just seeing the early signs of demand shock that is coming and we’ll need a helluva lot more compute. I am sure of this.

Subscribe now

Share

That’s all friends!

Thanks for reading Capitalist-Letters!

Please share your thoughts in the comments below.

👋🏽👋🏽See you in the next issue!

Steve Jobs once said, “You can't connect the dots looking forward; you can only connect them looking backwards.”

This doesn’t mean you can’t have any reasonable prediction about the future, but it means that you can only track the flow going backwards. Then, you can build a reasonable narrative of the future based on the flow you see in the past.

I have avoided writing about the Iran War exactly for this reason. Everybody was so focused on what would happen the next day. But to understand what is actually happening and what may be the implications, we needed a breather.

A breather without constant news flows, so we could look backwards and connect the dots. We got it with the ceasefire.

Connecting the dots and answering these questions isn’t easy. It requires going through the roots of modern power dynamics, economic order, the modern world order and its challenges, and finally, the changing power dynamics.

What the birdview makes me think is that the world might be on the edge of historical changes in the balance of power and the economic order. How could this order change, and what will be the implications are the two most important questions today. Will the US stay as the undisputed global hegemon? What does Europe’s and the Gulf’s drift away from the US mean? The roots of the world order and its transformation so far will be illustrative.

Let’s connect the dots and understand what’s happening and where we are headed.

Subscribe now

The World Order

“World order”… This is a term that is literally everywhere today. Every news channel, every commentator, and basically everybody monitoring the situation in Iran is using this term, though most of them don’t have any idea what it means.

What’s the world order that everybody talks about?

It’s funny because no such thing as the world order actually exists.

Once asked, “How’s the world order?”, Henry Kissinger famously replied, “Who asks?”

The idea of the world order is an illusion, but it’s an ancient and powerful illusion. In terms of its objective, it isn’t much different from religions. Maybe one can even go as far as suggesting that religion is, in its essence, a search for an order.

We see this in the earliest sources that gave rise to the idea of a global order. The earliest trace of the idea is the Sumerians’ concept of “me” (pronounced as "may"), from their legend Inanna and Enki. “Me” referred to a set of divine decrees that governed everything, from spiritual to how beer should be made. That was the original “order of everything.”

The Egyptians had a similar concept of Ma’at, which basically meant truth, justice, and balance. Similar to the Sumerian concept, the order of Ma’at wasn’t a human construct; it was a divine obligation.

In the ancient world, these concepts could never converge because the contact and understanding among civilizations were very limited. Most Sumerian villagers didn’t even know that another civilization called Egypt existed. Thus, the early contact between civilizations was characterized by skepticism, not understanding.

In Ancient Sumerian texts, Amorites (western nomads) were described as people who “know no grain,” “know no house,” and “know no city”, essentially defined by what they lack relative to Sumerian civilization. The Gutians (mountain peoples from the Zagros) were described as barbarous agents of chaos.

One looking at these records today may not be sure whether Sumerians were objective or whether every other civilization that they couldn’t understand was barbaric to them. If you don’t understand, strange means the same thing as enemy.

It’s certainly the latter as we see the same pattern in other ancient texts as well. The ancient Chinese concept of Tianxia put China at the center of civilization, and the peripheries were occupied by barbaric states. The civilization was described in terms of familiarity.

This self-centric understanding of the world in ancient empires had a few important implications:

• The perception of no order in other tribes turned them into security threats.

• Disagreements in commerce or borders could quickly turn into wars.

• They wanted to impose their order on everybody to feel secure.

They were naturally driven to extend the coverage of that order, as the lack of order in the neighbours meant a security threat to them. Thus, for the empires of pre-modern ages, the world order meant their order. When something was out of their control, it lacked order.

The term “Pax Romana” reflects this overarching sense of order.

When the Romans extended their control over the whole Mediterranean and the Levant, they called it Pax Romana, i.e., the Roman Peace. It isn’t a modern term. Seneca the Younger used it first around AD 55. For them, the Roman order was the equivalent of peace, security, and prosperity in the world. Tribes of Gaul wouldn’t agree.

This sense of an Empire-centric global order dominated until the modern age. Even 1,400 years after the height of the Roman Empire, Ottomans, when they conquered a substantial part of the historic Roman Territory, named it Pax Ottomana, reflecting the continuity of the imperial understanding of world order. The tradition lives.

Thus, when we are talking about the world order today, in the eyes of the imperial traditions, it still means their order because they are essentially governing over different people. They have to reflect their order on different subjects.

Think about Russia. It’s a modern empire. There is no ethnic unity. It rules over 190 ethnic groups in its territory. To govern Chechnya and St. Petersburg under the same authority, it needs to reflect a concept of a global order.

Iran is also an imperial state.

They have a long history of the Persian Empire, and today they are ruling over many different ethnicities. Only around 50% of the population can be considered as pure Persian today. Thus, they also have to radiate an imperial understanding of world order, one that their current regime finds in Shia Islam.

This understanding of the world order, i.e., an imperial peak secured by natural borders, has persisted in Asianic nations until today, as the abundance of land gave rise to great empires rivaling each other’s world vision, instead of smaller sovereign states.

This is why Russia and China, as centers of power, challenge the current order. They have deep imperial traditions that urge them to reflect their sense of world order over the globe and impose it on others. Until that happens, they don’t feel secure.

On the West, however, the concept of world order evolved differently post-Medieval.

Europe was especially well-suited for a different concept. There were many small states within a small continent; they shared the same religion, though with twists, and even languages were similar to some extent.

After the Thirty Years‘ War from 1618 to 1648, when they convened in Westphalia, they were torn and willing to negotiate an order rather than trying to impose it on each other. They were similar enough to negotiate and different enough not to merge and become another empire.

Westphalia Peace accepted states as the units of international relations and recognized the fundamental right to exist. Thus, self-interested sovereign states emerged as the recognized actors of global politics, and the world order meant the balance of power between them.

The problem of the Westphalian order was that it was fragile. It didn’t have a mechanism to protect the balance against an aggressor. After Napoleon’s imperial ambition was defeated by a coalition led by Britain, they addressed this fragility in the Congress of Vienna in 1815, where the Great Powers agreed to maintain the order together against an aggressor.

Though Austria, Prussia, Russia, Great Britain, and France were referred to as Great Powers, it was actually Britain that underwrote the whole system thanks to its naval and economic dominance. Thus, Britain ascended to the role of the world police as the hegemon of the time.

That position provides some privileges to the hegemon.

Some contradictions in the order are tolerated because of the sovereign-equality principle, but others defer to the hegemon on critical matters and align with it. Its pressing interests make up the red lines and take priority over others’ interests. This is like the tax others accept for the continuity of the order. It’s sort of the position the Romans called “Princeps”, i.e., the first in order among equals.

Britain played this role starting from 1815. When Russians tried to expand to the South in 1853 by invading the Danube and set their eyes on Constantinople and the Dardanelles straits, Britain fought alongside Ottomans in stopping Russia, not because it cared about a declining Muslim empire, but because Russia’s expansion would upset the order of the West.

For a hundred years, from 1815 to the start of World War I in 1914, Britain, as the hegemon, played a key role in maintaining the Western order. Pax Britannica.

At the end of the WWI, the British power had eroded substantially. They were unable to prevent WWII and were in an even weaker position at the end of it. The Western order needed a new police, a new hegemon.

The US rose to the role.

Since the end of WWII, this role requires more than internal policing, as China and Russia (first as the Soviet Union) have detached themselves from the Western order and are imagining different world orders rooted in their eastern imperial traditions, as explained above. They are joined by some other imperial traditions as well, notably Iran.

The rise of challenging visions of order required the US to go beyond policing and lead the Western order against the competing visions by projecting power, which it did through nuclear deterrence and economic dominance.

So far, it has worked, but the challenges are growing.

Pax Americana: The Post-Modern World Order

It was August 1945, almost three months after the Germans surrendered. Allied forces met in Potsdam to plan the post-war peace. This is a frequently told story from it.

During the conference, Truman casually walked over to Stalin and mentioned that the United States had developed “a new weapon of unusual destructive force.” He was basically letting Stalin know that America would be running the show from then on and that it would be the anchor of the new order. Stalin nodded, showed no surprise, and simply said he hoped America would make “good use of it against the Japanese.” Truman walked away thinking Stalin hadn't understood.

At the time, nobody knew, but today we know why Stalin was so calm. Because he knew. There were not just one, but several Soviet spies in the Manhattan Project; they had passed several important secrets to the Soviet Union, and the Soviet atomic bomb project was already underway.

By 1949, the Soviets had their own atomic bomb. This put the US in a strange position. It was a global superpower, but it was much more limited in its ability to resort to brute force compared to prior superpowers, because the rival was also equally destructive.

Imagine the 15th-16th century. Ottomans didn’t refrain from using their giant Basilic Cannon. They wouldn’t refrain even if their rivals also had it. The maximum extent of damage was not humanity-ending. But the possibility of a nuclear war threatens the very existence of humanity.

Thus, America’s ability to resort to brute force against its rivals was limited from almost the beginning. So, the US had to come up with new ways to defy the Soviet vision of World order.

It used two things—deterrence and economics.

Once the US understood that the Manhattan Project was compromised and the Soviets would eventually make the atomic bomb, it decided to build permanent bases overseas, as nuclear deterrence cut both ways. They couldn’t nuke the Soviets, but the Soviets couldn’t nuke the US either. So, deploying permanent bases overseas would mean that Soviet aggression against the respective land would mean aggression against the US, which the Soviets wouldn’t dare to do. It formalized this through the establishment of NATO in 1949.

The second tool of the US empire is its grip over the global economy.

Before World War I, global currencies were directly pegged to gold. The US exported massive amounts of weapons during the two world wars, and buyers paid in gold. As a result, the US had most of the world’s gold reserves by the end of WWII.

Between two wars, some countries had dropped the gold standard and devalued their currencies to escape deflation. This led to tariff wars and the collapse of global trade. A new economic order was needed to prevent this from happening again in post-WWII.

Countries convened in Bretton Woods in 1944 and adopted the gold standard again. The medium of exchange was also needed, as gold isn’t easily transferable. There wasn’t any real alternative other than the USD, since most of the gold reserves had already accumulated in the US. They made the USD the medium of exchange and pegged it to gold at at $35 an ounce. Thus, it became the official reserve currency.

This arrangement had two consequences:

• The US would have to run a trade deficit as it had to send more dollars out to fund the allies and their recovery than it took in, so that they could develop, and the US could have long-term trade partners, which would in turn fuel the US growth.

• Other countries would absorb the US’ trade deficit as the best way for them to make use of their USD reserves was recycling it back to US treasuries. Calling the gold was seen as an unfriendly act and was logistically hard.

This gave the US great legroom to ramp up spending.

It could print and spend more than it actually had without facing the catastrophic consequences any other country would face. The thinking was that the more it spent, the more the Soviets would have to spend, and this would eventually break their economy, as they didn’t have the reserve currency.

The idea was correct, but the Soviets saw the vulnerability of the system. If the US were running trade deficits, its increasing spending would distort the system at some point, as it would have more dollars out than gold at Fort Knox. Then, if other countries redeemed their gold at the same time, the system could break. Thus, they also provoked the US to spend more.

In the late 1960s, the US had substantially increased spending due to the Vietnam War and expansionary policies at home like Medicare and Medicaid. Other countries grew increasingly skeptical of the safety of the gold backing their dollar holdings and started to call in their gold. This forced Nixon to drop the gold standard.

America knew that the USD wouldn’t rapidly lose its reserve status, as it had become deeply embedded in the global financial system by then. But they also knew that this could change in the medium to long term without a mechanism that would push constant demand for the dollar. This would restrain the US’s spending in the long-term, and may lead to losing its edge against the Soviets.

Kissinger, the Secretary of State at the time, took the risk of losing reserve currency very seriously. As an avid student of history and international relations, he knew that a superpower without a reserve currency was simply naked.

The most striking example of this in post-medieval history is the Ottomans.

The empire was at the height of its military power in the mid-16th century. It controlled all of Anatolia, a significant portion of the Balkans, all populated areas in Northern Africa, the Eastern Mediterranean, and the Arabian Peninsula.

Yet, the reserve currency of the era was the Venetian Ducat, the currency of a small city-state, because it controlled the trade in the Mediterranean despite the Ottoman control over seaways, and avoided everything that would require them to devalue their currency.

As a result, Ottomans had to devalue their currency constantly to fund their endless military campaigns, and by the end of the 16th century, their economy was already breaking, which also undermined their military power over time.

This was exactly what was happening to the USD after it dropped the gold standard. It was devalued twice (1971 and 1973) and was losing value against every major currency. And then, in October 1973, the Yom Kippur War triggered the OPEC embargo, quadrupling oil prices overnight. The American economy was simultaneously experiencing inflation, recession, and an energy crisis.

When the US started negotiations with Saudi Arabia to lift the embargo, Kissinger saw an opening to use oil as the new anchor for the dollar. Saudi Arabia had sheer security concerns. The rulers were moderate about the Arab-Israeli conflict, but they were surrounded by hardline Arab states and people. The regime was vulnerable to these threats despite the massive wealth they were sitting on. America offered security guarantees to the regime in exchange for Saudi Arabia recycling its excess oil revenue to US treasuries, perpetuating the USD demand, and basically funding the US deficit.

This created what’s called the petrodollar.

The system expanded as other Gulf countries followed suit. By 1975, the rest of the Gulf states also agreed to price their oil in USD, and the US extended the security umbrella to cover them as well. The dollar found a new anchor and constant demand that allowed the US to ramp up spending basically indefinitely without facing the strict limitations of being pegged to a hard asset like gold.

The petrodollar played a pivotal role in the US winning the Cold War. Soviets could mobilize their economy for military purposes, but at enormous cost to civilian living standards. Every ruble spent on missiles was a ruble not spent on consumer goods. The Soviet public bore the full weight of military spending directly.

The US faced the same tradeoff, but the reserve currency softened it enormously. Reagan used this advantage as a strategy by increasing the buildout, because the US could finance it through deficits, the Soviets couldn’t. The Soviets tried to match, but they had to divert more resources from the public, which led to their economic disintegration in the late 1980s and ultimately the collapse of the Union.

This is the mechanism that built Pax Americana. Nuclear deterrence limited the direct conflict, shifting the fight to the economic domain. The reserve currency gave the US the upper hand in the economic race, and the petrodollar made the US's triumph inevitable, as it could ramp up the spending until it won, since the deficit would be funded by oil revenues of the Gulf, and nobody could stop needing oil.

Therefore, for other great empires with different visions of world order, mainly Russia and China, the way to challenge the system is to challenge the USD, which requires an attack and an alternative to the petrodollar.

This is why the current conflict with Iran is critical.

It’s an oil-producing country, and it controls a waterway that is indispensable to global oil flow. If this current conflict results in a way where Iran becomes more integrated into the world economy without adopting the petrodollar, it’ll likely be the greatest damage to the system since its beginning. Second order effecst can also unravel to further undermine the system.

Let’s find out how.

Share

Iran: The Fault Line Of Pax Americana

It’s an axiom that Russia and China have long been trying to undermine the dominance of the USD, as it’s the anchor of American dominance.

Though the idea is obvious, de-dollarization is only recently observable. Challengers of Pax Americana, China and Russia, have only recently become powerful enough to press for de-dollarization. Russia had to complete its restoration after the collapse of the Soviet Union, and it took decades after Deng Xiaoping's reforms for China to become big enough to threaten the US dominance.

Thus, de-dollarization took off only after Russia invaded Crimea in 2014.

Europe and the US imposed heavy sanctions on Russia that included travel bans, asset freezes, and restrictions on key Russian state entities from raising long-term dollar debt on Western markets. That eroded the confidence in Russian stability, and investors moved their money from Ruble to USD, triggering a collapse in Ruble.

This was the wake-up call for the Kremlin.

They understood what could happen in further sanctions, like getting cut off from the SWIFT system, which ultimately happened after the Ukraine War. So, they started openly calling for de-dollarization.

China saw this as well and woke up to how its dollar exposure was making it a hostage.

This led to Russia and China partnering up to spearhead the de-dollarization. In 2014, they signed a three-year currency swap deal worth 150 billion yuan and committed to reducing the USD exposure of their reserves. Between 2013 and 2020, the Russian central bank halved its dollar-denominated reserves, and China cut its US treasury holdings by 50% from its highs in 2013:

Their efforts for de-dollarization accelerated after Russia invaded Ukraine in 2022. Major Russian banks were cut off from the SWIFT system, and Russian oil was sanctioned. But they were prepared. They extended the currency swap agreements multiple times, and the Bank of Russia established a standing swap facility in January 2023 for CNY/RUB swaps, limiting volatility and ensuring ample yuan supply for trade settlements.

Thus, Yuan quickly became the dominant trade currency between Russia and China. The trade between the two countries was now de-dollarized.

China expanded this model by establishing direct swap facilities with 40 of its big trade partners. As a result, more than half of China’s cross-border trade was in Yuan in 2025. Also, you see the jump in Yuan-denominated trade financing in 2022? That’s ramped up de-dollarization after Russia invaded Ukraine in 2022.

Against these challenges, the petrodollar system is something like a Great Wall that protects the reserve status of the USD. You can de-dollarize your bilateral trade lines, but the global oil trade runs on USD. This is why the Iran conflict is so important.

Iran has long been restricted from the US banking system and hasn’t used USD in its oil trade at least since 2012. This means that if you include Russia as well, around 14% of global oil trade is off the dollar rails now. This is a significant challenge to the petrodollar, and it could become a bigger challenge depending on the outcome of the war.

Since the conflict started, China has done its best to help Iran keep the Hormuz closed, even though that directly hits China itself in the short-term as the largest oil exporter in the world. It did this because the stakes are larger.

Iran gained significant leverage by keeping the Hormuz closed, and they ask for three things in exchange for opening the strait and giving up their nuclear programme:

• Release our assets

• Unsanction our oil and economy

• Recognize our control over Hormuz

The first one is immaterial in its long-term effects, but the second and third ones are worth dwelling on.

If Iranian oil is unsanctioned, it will almost certainly sell its oil in a currency other than the USD. It’ll generally be yuan, but can also be another currency depending on the other party of the trade. It can certainly accept euros from the EU, etc. This will make the de-dollarization of the oil trade official, though to a limited extent.

It also wants to charge $2 million worth of service fees to the tankers going through Hormuz. It wants the payment in yuan or cryptocurrency, but the former is more convenient for all. This won’t have a direct structural effect on the oil trade, but it’ll have substantial psychological effects since the yuan becomes normalized as a currency you must hold for energy logistics, and the leap to pricing oil itself in yuan becomes smaller.

There are also second-order effects that are beyond Iran but equally threatening to the petrodollar system.

The most notable one is America’s failure to defend the Gulf countries when Iran targeted them in retaliation. This created in them the perception that their security comes after Israel for the US. Accurate or not, this is the perception they have now, and thus they are increasingly willing to hedge their bets. As WSJ reported, Gulf Countries are now rethinking their security ties with the US.

Saudi Arabia has already created the infrastructure to sell oil in Yuan. They signed $7 billion currency swap agreement in 2023. In 2024, they agreed that the Shanghai Gold Exchange would create a physical vault in Saudi Arabia. This way, they could turn their Yuan revenues into gold in Shanghai and store it in their own land, mimicking the petrodollar’s recycling system. As of early this year, this trade started to take place.

It wouldn’t be surprising to see other exporters in the Gulf follow the same suit to hedge against the weaponization of the USD and the US’s preference for Israel when it comes to regional interests.

Oil exports from the Gulf countries, excluding Iran, to China make up around 8% of the global oil trade. If you add it to 14% of the oil trade that is already off the dollar rails, we may see +22% of global oil trade off the dollar rails in the next few years.

This is already an unprecedented challenge to the petrodollar and thus to the USD’s reserve currency status. If exporters expand the direct currency swap and gold recycling system for other buyers as well, like Europe, we can see an unprecedented de-dollarization of the global oil trade over the next 5 years.

The end of the petrodollar could bring the end of the USD as the reserve currency of the world, endangering American global dominance and the post-modern world order.

This is why the current conflict in Iran is the fault line of Pax Americana, and challengers are rooting for the dawn of a new order.

Technodollar: The New Anchor

Pax Americana has many shortcomings, or even shitty parts, but it’s impossible to imagine a better world dominated by Russian and Chinese imperial visions, as they don’t face the internal limitations on power that the US government faces for being a liberal democracy, however imperfect it is.

So, can Pax Americana survive or perhaps modify itself even if the challengers succeed in breaking the petrodollar?

It will need to keep the USD as the reserve currency. How can it do that? There is a path, and the name of that path is Technodollar.

When you look at the corporate profits, it’s already here:

If Kissinger asked today what could be the next anchor for the USD, something everybody wants and needs, his answer could be technology.

And the US is the world’s #1 technology producer. What’s even better for the US is that this is an internal consequence of the US system, and it can’t be taken away from it. The American system has proved to be exceptional in creating entrepreneurs who, in turn, produce groundbreaking technologies.

Other imperial dreamers of this time, Russia and China, lack this.

They lack this because innovation requires the protection of personal property and freedom of ideas. However imperfect it is, the US’s liberal-democratic capitalism provides this way better than the authoritarian regimes of China and Russia, and the idyllic but inert welfare states of Europe.

Authoritarian regimes can mobilize resources very fast, so they can copy and follow innovations rapidly, but they are very bad at natural innovation. Look at China. It’s a very strong challenger in everything. They have been catching up even in AI. But how many groundbreaking technologies of the last 50 years came first from China? None.

This systematic advantage is supported by the biggest, unrivaled venture capital industry in the world, almost guaranteeing that the US will have an upper hand in the next big thing.

This is the US’s inalienable advantage.

Does the world want Nvidia chips? Pay in dollars. Foundational models? Priced in dollars. You want robots? Bring in dollars. And whatever the next thing is, it’ll be in dollars.

Imagine oil disappears tomorrow, and we bring in a stranger from a different planet to rationally pick the reserve currency of the world. He’ll look and see that the US has more of the things that everybody wants and will naturally pick the USD again.

This is the way Pax Americana survives, and perhaps further thrives even if the petrodollar breaks. It’s a narrow path, but it’s a path nonetheless.

Closing

There is one objective reality—the current world order is being challenged.

It’s being challenged by China and Russia as Great Powers, and other traditions that align themselves with these actors for their disturbance of the current order, like Iran.

They know that the USD is the linchpin of the Pax Americana, and they have been attacking it for more than a decade now, eroding its power.

However, the USD is guarded by the petrodollar system, and challengers lacked a transformational event that could deeply shake this system.

Iran could be that event.

It’s unimaginable that the US may face some kind of military defeat in Iran. The asymmetry of power is just immense. But this doesn’t mean the US will be the winner in all these.

There is a path where the war ends with Iran’s reintegration into the global trade while pricing its oil in Yuan. Second-order effects may lead Gulf countries to trade oil in domestic currency first with China and then with other buyers, as they have witnessed the limits of the US security guarantees and are looking to diversify their bets now.

Will the world enter this path? Impossible to know.

But, even if the petrodollar breaks at some point, I don’t think this will directly mean the end of the USD and thus US dominance. The US produces technologies and innovations that can match neither the authoritarian regimes of Russia and China nor the aging populations of Europe. Technodollar is capable of replacing petrodollar as the perpetuator of global USD demand.

In short, we are going to see a significant evolution of the world order, in one way or another, post-Iran War. It could be one where the challengers raise their influence; or one where the petrodollar declines and the technodollar rises, transforming the US from a bittersweet world police to a more benign dominant power; or one where the US behaves less restrictively in using force to maintain the current system as it is, in which case the system will look similar but the order will be wildly different.

In many senses, we are witnessing the dawn of a new world order.

Share

Subscribe now

That’s all friends!

Thanks for reading Capitalist-Letters!

Please share your thoughts in the comments below.

👋🏽👋🏽See you in the next issue!

In all times, it’s only second to inventions like writing, the wheel, or discoveries like fire that made humans endure and thrive over the course of centuries. If you leave these aside, nothing comes close to the internet's impact.

What the internet did differently from all other groundbreaking human inventions that changed the world, like the steam engine, was that it made the transfer and spread of knowledge infinitely faster. That led to shorter cycles of groundbreaking discoveries, accelerating human development.

The historical development of internet traffic illustrates this.

Tim Werner Lee invented the World Wide Web in 1989 while he was working at CERN. That was the ground zero. The internet traffic was essentially nonexistent.

CERN opened the World Wide Web to the public for free in 1993. Later that year, the global internet traffic reached 8,715 gigabytes per month. By 1997, this number reached 5 million, more than a 500x increase. What do you think it currently is? It’s around 733 exabytes. 1 exabyte is 1 billion gigabytes. This is how far we have come.

In many aspects, 2026 is to AI what 1997 was to the Internet.

This is even though AI has been adopted pretty rapidly compared to other technologies so far, including the internet.

I think this very rapid adoption led many people to underestimate AI’s capabilities. We suddenly forgot that this technology is still in its infancy and started to assess its impact, drawing on its relatively high adoption rate.

The simple assumption was as follows: If the curve goes like this, it’ll hit the internet level adoption in a few years, and it’s unlikely to lead to an internet level transformation in this period, based on what we are currently seeing and what we can imagine for the foreseeable future.

This led many of us to question the data center buildout, thinking that we’ll have a massive amount of excess capacity if the AI doesn’t fulfill its promises.

What most people miss is that demand is a function of three things:

• Adoption

• Capabilities

• Spillover effects

Most people are focused on the adoption and are completely overlooking capabilities and spillover while they assess AI’s potential and future demand.

This is what we are going to lean into today. I’ll try to explain why I believe the demand will far exceed anything we can build in the foreseeable future by discussing:

• How adoption, capabilities, and spillover determine demand.

• How AI’s capability improvements are driving demand.

• How large is the spillover area for AI?

Finally, I’ll briefly discuss our exposure to compute demand and how I am looking to increase this exposure over time.

Let’s get started.

Subscribe now

Demand = Adoption X Capabilities X Spillover

Let’s take a closer look at the adoption curves I have just shared above.

What do you think is striking here?

To me, the most striking thing is that television actually has a steeper adoption curve than the internet. Within almost 10 years after the introduction, television reached 80% of the American population. For reference, this was just around 45% for the internet.

Yet, despite exponential early demand and adoption, we know that the demand for television had been pretty flat even before streaming became mainstream.

Why? Because, despite its contribution to the human experience as a standalone invention, it’s not what is called a “General Purpose Technology” or GPT.

You may ask what the hell is a General Purpose Technology?

The term was coined by two economists, Timothy Bresnahan and Manuel Trajtenberg, in their seminal 1992 paper called “General Purpose Technologies, Engines of Growth?”

Their seminal insight was that most of the industrial and economic growth in human history stems from a handful of technological developments/inventions, while the rest of the gadgets contributed only marginally to our development.

When they dug deeper into what makes a GPT, they found three properties:

• Pervasiveness: This is adoption. GPTs are adopted by the majority of humanity.

• Improving Capability Frontier: Capabilities of GPTs don’t stay static; they improve over time.

• Complementarities and Innovation Spillover: GPTs spawn other innovations as the productivity of R&D in downstream sectors increases as a consequence of innovation in the GPT.

Adoption is straightforward; for something to become a GPT, it should reach wide adoption. But let’s dig a bit deeper into capability frontiers and innovation spillover.

I am using the term “improving capability frontier” to describe the expanding usage of technologies; this isn’t the original term used by the writers. The original term Bresnahan and Trejtanberg used was “technological dynamism.”

They found that:

“continuous innovational efforts about GPTs, as well as learning, increase over time the efficiency with which a GPT performs its generic function. This may show up as reductions in the price/performance ratio of the products, systems, or components in which the GPT is embodied, or as multidimensional qualitative improvements in them.”

General Purpose Technologies, 1992, p.5

Their initial focus was on the cost of performance by a GPT. They observed that, unlike other technological advancements, the cost of performance by a GPT plummets over time as wide adoption leads to higher innovation efforts on GPTs, which inevitably leads to price reduction over time.

So, what they understood from “dynamism” was increasing efficiency, not capability expansion. They didn’t dwell on the idea that a GPT could, over time, become able to perform a function that it initially wasn’t able to. The transistor was their example.

They argued that the original semiconductor was performing binary logic. Developed chips essentially do the same thing. They do it faster, at a larger scale, which leads to significant cost reduction, but they essentially do the same thing.

This is essentially true, as the very reason they are faster can scale better is that it’s become cheaper and cheaper to produce transistors at smaller sizes, so we can pack more of them in a chip, which delivers better performance at a lower price. This is a direct result of widespread innovation efforts on transistors, as the writers suggest.

However, Bresnahan himself revised this idea in 2010 in his solo chapter for the Handbook of the Economics of Innovation and wrote that “a GPT may have attributes other than cost, and that the transition from one GPT to another or the creation of an important new version of a GPT may involve changes in attributes other than cost.”

He pointed to the possibility of capability expansion by admitting that a new version of the same GPT may have different non-cost attributes than the original.

In short, based on where the literature has reached today, the ability to improve both in terms of costs and capabilities is an integral property of a GPT.

The second distinguishing factor of GPTs that requires closer attention is Complementarities and Innovation Spillover. Again, this is the term I picked for better clarity. The original term Bresnahan and Trejtanberg used was Innovational Complementarities, or simply IC.

Their powerful insight was as follows:

“GPT's are characterized by the existence of innovational complementarities with the application sectors, in the sense that technical advances in the GPT make it more profitable for its users to innovate, and vice versa.”

General Purpose Technologies, 1992, p.5

In a more straightforward language, GPTs are critical inputs to downstream R&D processes. Thus, the cheaper a GPT becomes, the cheaper all the downstream R&D processes become, leading to a higher volume of innovation that can reach more people. Those innovations can also be inputs to further downstream R&D processes.

They again illustrated this by using semiconductors as an example:

Their insight was that these characteristics of GPTs create a virtuous cycle that leads to ever-increasing demand for a GPT:

• Step 1: A GPT improves both in cost efficiency and capability.

• Step 2: An improved GPT makes R&D more viable and profitable for downstream sectors.

• Step 3: Downstream sectors innovate, developing new and better products. This increases their demand for the GPT input.

• Step 4: The increased demand and the expanding set of application sectors make it more profitable for the GPT provider to invest in further R&D.

• Step 5: Return to step 1. GPT improves again, triggering a new cycle.

The natural consequence that Bresnahan and Trejtanberg arrive at is that the demand curve for GPTs reminds us of increasing returns to scale, where a unit of input turns into more than a unit of output.

This is because when we are looking at a GPT, demand doesn’t closely follow adoption as it does with other gadgets, but it’s a function of adoption, improvement in the capability frontier, and innovation spillovers.

This is why television has a very different demand curve than the internet, despite its faster historical adoption:

Television is highly limited when it comes to capabilities and spillover. Its capabilities haven’t expanded much since its invention, and it’s not an input to many downstream R&D sectors. It’s not a GPT. Thus, its demand curve closely follows adoption. When adoption flatlines, so does the demand.

The internet, on the other hand, is essentially a GPT that is input to many other sectors, its capabilities improve by the day, and it leads to other innovations. Just last year, the internet traffic grew by 15% according to Cloudflare. If you look at the adoption curve, you’ll see a maturing technology; if you look at demand, you’ll see sustained growth as it’s a GPT.

This is why the seemingly “already high” adoption of AI doesn’t mean much for the future demand, as it’ll be a function of capability improvements and innovation spillover, because it squarely fits the definition of GPT. This is what most people miss.

Imagine how the demand for microprocessors has increased since Bob Noyce invented them in 1959. Then go beyond it for AI because it’s one of the rare GPTs that can experience a capability expansion, not just cost reduction. And remember, one of the parameters of the demand for a GPT is capabilities.

Share

Improved Capabilities = Higher Demand

This is the core feature of GPTs.

Capability expansion is one of the main drivers of demand for GPTs, together with adoption and innovation spillovers. Because of this, the demand for a GPT can grow for decades or even centuries after the widespread adoption is achieved for generic uses.

Bresnahan and Trejtanberg only observed this in their 1992 paper, but didn’t provide an empirical background. They later did this by tracking how an upgrade from an early steam engine to what’s known as the Corliss Steam Engine expanded the use cases of the steam engine and increased demand.

Some basic context about steam engines first.

In the simplest terms, heat turns water to steam, which expands with enormous force. Channel that force against a piston, connect the piston to a wheel, and you have continuous rotary motion, which is basically what a steam engine is:

The problem with the early steam engine was that it didn’t have a mechanism to control the power delivery. Steam flooded the cylinder at full blast regardless of what the engine was actually doing.

Imagine you have 20 looms connected to that engine via belts and shafts. Someone starts up 5 more looms. Then someone shuts off 10 looms for maintenance. The resistance drops, and the engine speeds up. The engine has no way to automatically adjust how much steam it's getting; it just takes whatever comes.

This worked fine for work where jerky, uneven power delivery didn't matter much. Sawmills, for instance. You're just pushing a blade through wood, and if the speed fluctuates a bit, who cares? Pumping water out of mines, just straight up-and-down motion, no precision needed. Basic grain milling, steamboats, and locomotives, etc.

But it wasn’t helpful for things that required precision, stability, and even power delivery. Delicate textiles, for instance. Cotton thread snaps easily, and an engine that lurches when the load shifts means constant breakage. So textile manufacturers were stuck producing only coarse, low-grade fabrics with steam, but not fine textiles.

This is where a new version of a GPT, the Corliss Steam Engine, stepped in.

It changed the valve design. Previous engines used a single slide valve that opened and closed in a fixed pattern. Steam always entered for the same duration regardless of conditions. Corliss replaced that with separate intake and exhaust valves that could be opened and closed independently, and connected them to the governor through a trip mechanism that could release the valve at any point in the stroke.

That meant the governor could decide, mid-stroke, that the steam was enough and snap the valve shut. Instead of a fixed pattern that delivered the same amount of steam regardless of what the engine was doing, the Corliss engine could sense its own speed and adjust its fuel intake cycle by cycle, in real time.

This had two important implications:

• 30% higher fuel efficiency than conventional fixed-cutoff steam engines

• The steam engine became suitable for textile mills, factory lines, and electricity generation.

Because of this new capability, the use cases expanded substantially, and demand for steam horsepower (hp) kept increasing.

As you see below, the steam hp grew pretty fast between 1838 and 1850. The Corliss engine was commercialized in the early 1950s. Then, from the 1850s to the 1890s, number of steam engines increased by another 8x:

We know that Corliss contributed substantially to this growth, as a rapidly growing industry, textiles accounted for almost 50% of all Corliss engines. Before Corliss, steam hp wasn’t that widespread in textiles due to the limitations explained above.

Thus, if it weren’t for Corliss, steam hp wouldn’t be that common in textiles. This is a concrete example of a new version of a GPT that has improved capabilities, creating demand that didn’t exist previously.

This is exactly the case for AI, and we are at the very early innings of it.

Take a look at this:

In July 2024, Claude had around 15 million monthly active users. This number doubled to ~30 million by July 2025. Yet, token usage increased by 50x.

This was because the demand is a function of adoption and capability expansion. Every user was now consuming, on average, 25x more tokens because Sonnet 4.0 was able to do many things that Sonnet 3.5 wasn’t able to.

Sonnet 3.5 was working, but it was annoying. It was impossible to code a decent app or software end-to-end. Sonnet 4.0 was almost there. We now have 4.6 Opus, and it one-shots basic-to-medium complexity apps.

Open Router data further validates this.

Even though active user counts didn’t move much from July to December 2025 (it, of course, increased, but it certainly didn’t triple), token usage tripled:

It’s a direct result of expanding capabilities. If an LLM can’t vibe-code an entire app, you will use it as a sidekick; if it can, you’ll make it your employee. In the latter, you consume multiples of what you would in the former.

This is easily observable even without data. Think about it. When ChatGPT was launched, we didn’t have an image model or a video model. Now people are generating whole advertisements and short animations/movies using these models. This is a serious capability expansion that drives demand.

We should acknowledge that every new capability expansion will likely require more compute, at least initially. We have had two clear examples of this over the last two weeks.

Anthropic announced computer use in Claude Code and the ability to control it remotely through the Claude app on your mobile:

The moment I saw these, I thought token usage would skyrocket. People using Claude Code have already caused a significant jump in token usage. Now we are talking about AI itself using AI tools through your computer, which gives it the ability to use other internal or external AI applications to achieve a goal.

What do you think this will do to the usage?

Now model this to every function:

• AI uses AI+computer to create end-to-end marketing campaigns.

• AI uses AI+computer to do academic-level research and write reports.

• AI uses AI+computer for end-to-end customer management.

The list goes on and one and this is just one step. This is just the step where AI can use a local computer + other AI tools. There will be a step where AI will be able to do it in every edge device, even in your Apple Watch.

In this sense, it’ll be much like the internet. It’ll infiltrate everything, but it’ll be even more explosive than the internet because token usage can self-perpetuate. AI using AI in a loop. This wasn’t the case for the internet.

And even then, the internet traffic grew 1,000x from 2002 to 2022:

We aren’t even in 2002 for AI. 2026 to AI is probably what 1995-1996 was to the internet. We have seen that every inch of capabilities we unlock drives insane growth in demand, and we are just starting to scratch the surface of it. Demand that’ll be driven just by capability expansion looks beyond comprehension.

Innovation Spillovers Perpetuate The Demand

In their original article, Bresnahan and Trajtenberg avoided listing GPTs, as they were trying to lay down the theoretical basis. Instead, they provided only three examples, thinking that it would be clear that they were not exhaustive: Steam engine, electric engine, and semiconductors.

It was only in 2005 that a broader attempt to list GPTs came. Richard G. Lipsey, Kenneth I. Carlaw, and Clifford T. Bekar listed 25 GPTs in their 2005 book, Economic Transformations: General Purpose Technologies and Long-Term Economic Growth.

This was their full list:

When I saw their list, the wheel instantly caught my attention. It was invented between BC 4000-3000. So, it’s been around for 5 to 6 thousand years, yet the spillover it created keeps expanding, which keeps increasing the demand for wheels.

Here is a very interesting example.

By the 1970s, we used wheels in daily life, in simple and complex machines, in industrial lines, etc. Without wheels, there wouldn’t be a modern society. At the time, it was probably hard for anybody to imagine a new use case that would substantially increase the overall demand for wheels. Yet, we still didn’t have bags with wheels.

It never stops surprising me that we put a man on the moon before we put wheels below our luggages.

It was in 1972 that we put two wheels under a luggage. It had only two wheels on the back. You had to grab the handle at the front, lift it a little bit, so that it could slide on its rear wheels. Isn’t that insane? It took another 20 years for the modern Rollaboard with 2-4 wheels and a retractable handle to be introduced.

Imagine how this little innovation affected the demand for wheels? Every traveller in the world created a demand for 2-4 wheels, a demand that didn’t exist before.

This is a clear example of how innovation spillovers from GPTs can be sustained for a very long time, even for thousands of years, as innovations create new use cases.

We could easily see such a development for the internet as well.

When it was first introduced, streaming didn’t exist; now it’s the norm. How do you think this affected the internet traffic?

This is one of the crucial points the current skeptics are missing when modeling the demand for AI and thus for compute. They look at today and predict the near future, and can’t see factors that’ll substantially increase the demand. Yet, innovation spillovers from GPTs can easily span decades, even centuries.

It takes time to see substantial productivity gains from a GPT as we discover the capabilities and enhance them only slowly. If we could go back in time, we would likely see pretty stable demand for the wheel in its first millennia or two. Use cases would be largely the same.

Indeed, productivity gains from new technologies may take time to strongly manifest themselves:

Thus, when people look at the immediate impact of AI and don’t see any material contribution to growth and reach the conclusion that AI may be a bubble, or we may be overbuilding, they are completely ignoring the very long period of capability expansion for new GPTs and innovation spillovers.

The data below falls into the same mistake. It shows that real GDP growth in the US in the first half of 2025 was essentially 0.1%, excluding investment in processing equipment and software. It implies that AI itself didn’t contribute to the growth:

It’s true, but we are still at the very early innings of capability discovery/expansion and, naturally, spillover is very limited. At this stage, we can’t see the gains at scale. We need to look for the early signals.

And we are seeing many of these early signals. The most obvious one is coding:

It’s a very big deal that the IOS app releases, which were stable from 2022 to 2025, skyrocketed after Claude Code became mainstream and foundational models became capable enough.

I think many people underestimate the importance of this and tend to see coding as an isolated area where AI has a disproportionate effect. They still think the effects outside this will be pretty limited.

This is false. We are seeing the early signals of spillovers.

An amazing story is that an Australian entrepreneur who had no experience in biotech was able to develop an mRNA cancer vaccine for his dog. Yes, he took help from professionals at later stages. Yes, the vaccine didn’t completely kill the tumor and was only able to shrink it.

Qualifications exist. But the main question is whether this was even possible before?

It was nowhere near possible.

Another example is Matthew Gallagher, who basically created the first one-person billion-dollar company. He literally vibe-coded an entire telehealth company, used AI to create image and video ads, and sold through Facebook ads.

Again, the question is simple—was it even possible before?

It wasn’t. Things that weren’t possible before are becoming possible. This is the definition of innovation spillovers from a GPT. We are seeing anecdotes today; we’ll see them at scale in the future.

You can understand how broad the spillover area for AI is by thinking in just two steps:

• Almost everything that works on electricity can be made smarter.

• Then all those things will be used in related/unrelated fields to create new things.

Imagine how this will affect the demand.

We are just scratching the surface now. At this stage, we can’t see the impact at scale. What we can see are signals, and we are seeing very powerful signals.

How To Position For The Coming Demand?

We won’t have enough Claude anytime soon, as we are just scratching the surface of an emerging GPT. Thus, we won’t have enough compute.

Cloud providers are the obvious bets, but there are caveats.

First, demand doesn’t mean execution is granted. Firms still need to execute to capture as much demand as they can without risking their survival. If there is a mismatch between existing demand and their capabilities, the broader demand projections won’t save them.

Second, many hyperscalers have other businesses that can be affected by AI. Thus, valuations should account for other parts as well. If the survival of the software business was guaranteed, Microsoft would be a no-brainer here. But we aren’t sure how the productivity software business will evolve at the age of AI.

Third, Bitcoin miners pivoting to AI infrastructure to capitalize on the power they have already secured are exposed to higher execution risks. Despite this, the market has richly valued them, largely based on their connected power potential. Thus, most of them are out of scope for me.

This leaves me two players: Amazon and Nebius.

We own both of them in the portfolio.

Amazon is attractive to me because its other businesses aren’t threatened by AI. E-commerce and advertising will live, and may even accelerate. On top of these, you are getting the dominant cloud business.

AWS growth lagged other hyperscalers like Google and Microsoft for a long-time due to both its bigger base and lower exposure to OpenAI demand. This is changing now as Claude is growing way faster than ChatGPT. Thus, AWS growth accelerated as it is the preferred cloud provider of Anthropic.

SemiAnalysis predicted AWS growth to reach 24-25% by late 2025 as Anthropic’s gigawatt-scale AWS clusters were scheduled to come online:

We saw this played out as AWS growth reached 24% in Q4 2025. It’s set to further reaccelerate this year as Anthropic keeps growing fast and OpenAI is committed to expanding its AWS consumption by $100 billion.

These will certainly result in acceleration for AWS this year. UBS says AWS growth this year may reach 38%:

I think 38% is a long shot, but it’s possible that it can reach ~30% growth. Despite accelerating AWS growth, Amazon is trading at its lowest valuation at 27x forward earnings. Both the company and the stock are very well positioned now, so we’ll keep holding and increasing exposure over time.

Nebius is my neo-cloud pick.

I don’t like Coreweave’s customer concentration and its debt structures. Nebius is more straightforward; it’s starting from a smaller base, and valuation is cheaper.

They are targeting $7-$9 billion ARR this year with around 800 MW-1 GW connected power. They raised their contracted power guidance for the year from 2.5 GW to 3 GW.

They recently signed a $27 billion deal with Meta and raised $2 billion from Nvidia. According to the terms of the Nvidia investment, they’ll increase their Nvidia capacity to 5GW by 2030.

If they can deliver on this, they’ll be doing ~$50 billion ARR in 2030, based on the current GPU/hour rates. The whole company is now valued at $27 billion. The risk-reward here is still asymmetric.

We already own these two. What I am also considering recently is adding Oracle as well.

There has been a lot of discussion about Oracle’s buildout, and it’s undeniable that they are taking risks. The fact that most of their commitments came from OpenAI made it riskier, as OpenAI isn’t expected to become profitable before 2029.

However, OpenAI recently closed a $122 billion financing round, and we know that it’s eyeing an IPO later this year or early 2027. So, it’ll raise a substantial amount there as well. Thus, Oracle’s plans are now significantly derisked compared to what it was.

I am not buying any Oracle right now. I’ll look deeper at it, evaluate the risk and reward potential in more detail before I make any decisions. But it’s clearly more interesting now, and I am intrigued to dive deeper as it’s at just 19x forward earnings.

Last Words

There is not enough Claude.

This is what I feel whenever I use Claude Code or use it as my research assistant. And this is only the beginning, as AI looks very much like a General Purpose Technology in its infancy.

This means that we are looking at decades of capability enhancements, each unlocking new use cases, leading to innovation spillover and further increase in demand. Yes, adoption may look already decent, but we are just in the early innings of capability expansion and spillover. Demand will exceed anything we can imagine now.

Think about the transistor, another GPT. Today, we have more transistors in a single chip than anything its inventors could imagine in terms of total demand.

We have to see this broader picture, and don’t fall into believing that the capacity we are planning now may exceed demand in the long-term. It won’t.

We should position accordingly, but this doesn’t mean buy everything. Demand doesn’t guarantee execution. The key is getting your hands on the companies that are positioned to benefit from the demand without much execution overhang.

Amazon and Nebius are the two companies that fit this criteria for me, while Oracle is getting increasingly interesting as its biggest customer polished its financial position.

It’s important to pull yourself out of the daily discussions about AI’s contribution to growth, productivity, or efficiency. It’s too early to see developments at scale, but we are seeing the signals and anecdotes. Some things that weren’t possible before are now possible. This is what counts.

We are looking at the development of a new GPT. This is rare. And demand for every other GPT we have had so far has only increased over time.

It’ll be the same for AI.

There won’t be enough Claude, there won’t be enough cloud, there won’t be enough compute. We have to keep this in mind and position ourselves accordingly.

(If you have been thinking whether the “GPT” of “ChatGPT” was a reference to General Purpose Technology, the answer is no. It’s “General Pre-trained Transformer.”)

Subscribe now

Share

That’s all friends!

Thanks for reading Capitalist-Letters!

Please share your thoughts in the comments below.

👋🏽👋🏽See you in the next issue!

Portfolio is now up 21% in the last twelve months against 11% of the S&P 500!

“It’s not going to be an easy year.” — This was how I started the last portfolio update.

I started it this way because it was the theme that I had been emphasizing in several updates before it, and in all the trade updates I had posted before.

In February, we started to see the early weaknesses, and this month the downtrend accelerated as the Iran War weighed on the sentiment. I 100% believe that it would be something else if it weren’t this war, and weak sentiment would find a way to manifest itself. This is because it stemmed from an undeniable fact—the market was expensive.

Because of the elevated market, we stopped allocating fresh capital, did very little buying, and funded almost all the buys by exiting or trimming some other positions.

In short:

• We thought the market was expensive.

• We wanted it to come down to deploy capital.

• We thought it would come down sooner or later and increase our cash position.

This is exactly what we are having today—the market is going down.

This wasn’t market timing, as we don’t get out of the market to get in at some other time. It’s just plain positioning based on where we feel we are in the market cycle.

As Howard Marks says, it’s very hard to tell where we are unless we are close to extremes. If we turn back to late 2025, we could see that we were near the upper end:

Just before the new year, the market was trading at around 23.5x forward earnings, a nearly 50% premium to the 40-year average of 16.

It’s impossible to say when the market will peak and start to decline. However, you could easily say we would likely go down in the medium-term, as whenever the market passed above 20x forward earnings, a correction followed in the medium-term. You could position accordingly by increasing cash position and cutting exposure.

This is exactly what we did, so I am very happy that the market is coming down.

Yet, I see people are feeling the exact opposite. Most people are panicking, feel disappointed, and discouraged from further participating in the market. They started questioning themselves, their positions, and their knowledge.

This is the ultimate worst attitude you can have.

Let me tell you this—even the best investors go down with the market.

One of the biggest dangers I currently see is that people are fabricating portfolios on Substack and Twitter, and those portfolios always go up, somehow.

Don’t fall into this trap.

Some of them are outright lies, and this is very dangerous. I see people posting updates monthly, no granular transaction updates, and somehow they did some transactions within the month, they worked well, and they went up in a down market.

Obvious question—why aren’t you getting a loan and betting it all on your always-up strategy instead of posting your alpha on Twitter or Substack and making it known?

Some others, macro and thematic funds, are creating imaginary indices and baskets, trading this and that, cutting this and doubling down on that, and it all works out.

Well, everybody can become a billionaire with paper trading. Where is the money?

The same question persists—why aren’t you betting it all instead of giving away your alpha?

Because they aren’t real and are not honest. Best investors, best portfolios also go down, and they almost always go down with the market.

The quality of a stock is independent of the stock price. The quality can be undeniable, and it may still go down. If you hold it, you go down as well. Yet, that’s still the best thing you can do. That shouldn’t bother you at all.

Amazon’s quality is undeniable. It’s down ytd. If you are holding it, you are down. And it’s fine. We are down by 13% YTD, and it’s fine. We don’t trade; we want to own the high-quality businesses forever. We manage risk not by trading in and out of them but by adjusting our posture between defensive and aggressive, and by position sizing.

You’ll never make it unless you make your peace with a down market.

Take a look at this, I am obsessed with this picture lately:

As of December 31 last year, Buffett was up 44x on his American Express position.

Let’s take a closer look at this, shall we?

Buffett started building his current American Express position in 1990-1991. At the time, the stock traded between $5-$10 per share, adjusted for the splits. Given that his position was 44x bagger as of December 31, his average split-adjusted price is $8.40, meaning he bought some shares at higher prices, possibly above $10.

In June 2007, the stock was trading at a split-adjusted price of $65. His early shares were now at least a 6x-7x bagger for him, potentially higher. What happened then?

It collapsed by 85%. Not 10, 20, 30.. It’s freaking 85%:

It was almost 20 years after he started building the position, and some of his early shares were underwater. What did he do? He didn’t sell a SINGLE share.

And this was his second biggest position going into 2007, only Coca-Cola was bigger than American Express in his portfolio. Imagine what this did to his portfolio?

The best investors go down, and they are exceptionally good at detaching themselves from the stock price. What’s important is buying and holding exceptional businesses at attractive valuations; the market price is and should be irrelevant to you.

He is not the only one. Look at Bill Ackman:

He is probably down by over 20% YTD. What does this mean? Does this mean he is useless? No.

Like him or not, Ackman is a good investor with a long-term track record of delivering above-average returns.

He owns a lot of Uber, Amazon, Meta, and Brookfield. All these stocks are down by 10-20% YTD, so his fund is down by a similar amount. What message should he get from this? He should no longer hold these? It was a mistake to pick them?

If these were the type of messages Buffett got from a stock price, he would never be up by 43x on his American Express position.

• Good investors pick good businesses.

• Good stocks can still go down when people sell for who knows what reason.

• Thus, a good investor goes down.

What then happens is that when the sentiment changes, good holdings go up way more than the broader market, thanks to their business development. This is what certifies a good investor, and it takes a long time for this cycle to complete.

In short, don’t get disappointed in the market or in yourself, and don’t be pushed to follow those who seemingly always go up. It’s a scam, and I guarantee you it’s the worst thing you can do for your portfolio.

What you should do is simple:

• Own good business.

• Build a cash position when the market cycle is near highs.

• When the market starts going down, be chill. Go down with your good businesses.

• When prices are right, buy more of the good businesses and increase your stake.

When the sentiment changes and the market recovers, you’ll be way better off.

The hallmark of a good investor is being able to hold onto their quality businesses without getting their conviction shaken and buy more when the time is right.

So, don’t be disappointed, and use this as an opportunity to do three things:

• Test your conviction.

• Buy more of the businesses you like.

• Unfollow those who are UNABLE to go down as they are 100% scammers & liars.

If you are holding good businesses bought at attractive prices, I guarantee you it’ll work out very well over time. This is what we have always done here and keep doing regardless of what the market does.

💥Our portfolio is now up 21% in the last twelve months against 11% of the S&P 500!

In the previous updates, I provided my outlook for each company in the portfolio. In this update, I’ll also provide an overall commentary about what I expect from the portfolio and the market going forward.

I’ll also talk about the stock that I’ll be giving a priority to buy and why.

Let’s dive in!

🚨Our portfolio is proprietary to the members supporting the publication🚨

Members get portfolio updates every month.

Here is a 25% discount to celebrate our new readers!

Valid only until tomorrow!

Get 20% off for 1 year

📊Here Is Our Full Portfolio!

As of today, we have 18 holdings in our portfolio.

8 of these positions can be considered foundational.

Read more

Buffett made 10x his money on his Apple position that he opened in 2016. When you say it this way, it sounds natural, but it’s not, as it was already the biggest company in the world when he bought it.

This is probably the biggest alpha ever created in a mega-cap by any investor.

And it’s not a one-time event for Buffett; it’s a pattern.

He has made 22x on Coca-Cola, 44x on American Express, and a mind-blowing 51x on Moody’s. The next time Moody’s doubles, he’ll have a 100-bagger, American Express goes up by 130% in the next 5-6 years, and he’ll have his second 100-bagger.

I am thinking a lot about this lately because of how hectic the market is, how tired every investor I see feels from trying to stay on top of everything going on, from the war to oil prices to inflation to AI, you name it, and trying to position accordingly, while in fact it’s this simple.

Now compare Buffett’s simple picture above with this:

85% of active large-cap managers underperform their benchmark in both the short and long term. What’s even worse is that 85% of all managers across all strategies underperform their benchmark in the short, medium, and long term.

The question is why? What does Buffett do that they don’t do?

It’s especially interesting to think about large-cap managers because they are operating in exactly the same space as Buffett. And in the large-cap space, there aren’t that many alternatives, so everybody knows almost all the alternatives, and many of those managers probably had many overlapping positions with Buffett at some point.

So, what was his alpha?

The stock-picking ability alone can’t be Buffett’s alpha. Indeed, he bought Apple when it was the largest company on earth. Because of the sheer size of his portfolio, he couldn’t buy hidden market gems. He bought Apple, and many managers bought it too.

What does Buffett do that others don’t?

It’s simple. Buffett didn’t try positioning based on macro headlines, geopolitical events, or try to predict the next bottleneck in an emerging technology and stocks to benefit from it. He didn’t care about short-term catalysts, 12-month price targets, etc.

Thus, a more interesting and deeper question is, why do people still do it? Why are both professional and retail investors almost maniacally trying to understand what’s going on now and forecast what’ll happen next?

For Wall Street, the structure of the industry is said to be the root cause. If you don’t do what others are doing and underperform, you are the problem, and clients withdraw their money, you lose management fees, etc. But if you do what others do and still lose, it’s the market.

That’s correct, but it’s only half of the picture. Why? Because retail investors don’t have this concern, but they are still trying to position according to headlines.

The decline of the average holding period is the living proof of this:

What has changed in the last 50 years that has pushed us to hold stocks for shorter?

It’s the technology. Information now disseminates at the speed of light. In the old days, you had to read a newspaper, do research by looking at manuals, etc., and think before you called your broker and put an order. Now we see the headline in news feeds a minute after the event happens, people instantly scroll on X to see the reaction, and can react within minutes after the event happens.

Naturally, the holding period declines.

This is the real question—why do we feel the need to position according to information, and whether it creates a genuine alpha?

This is what we’ll discuss today. We’ll try to answer three questions:

• Why do we care so much about information?

• Does information really create alpha?

• The source of real edge in the market?

I hope this will give everybody a freshness of mind at a time when everything feels hectic. As Charlie Munger says, when everything goes insane, staying sane is your competitive advantage. We’ll achieve this together.

Subscribe now

1️⃣ Why Do We Crave Information?

The short answer is that humans are suckers for information.

But why? It’s simple—because we are smart and we know it.

This is what distinguishes humans from all other species. We know this, and from our early days, we get an education that conditions us to make sense of the world by using our brains.

Thus, information gathering is indispensable for us. Without it, we wouldn’t survive. However, our information-gathering system is flawed because it creates a very early tendency to rely on information.

Why? Because the early information we gather is physical and based on practice. Thus, it’s almost 100% correct, and you always gain from relying on information.

Think about a little child touching a hot stove.

He gets burnt, and that information is solid, physical, and 100% correct. He knows that he’ll burn his hand again if he repeats the same action. He never touches it again and never gets burnt the same way.

The information is 100% correct, and it always provides an upside.

Early life is full of similar experiences, so we develop a very early tendency to rely on information. When we start a formal education, let’s say first grade, this tendency is almost absolute. How many of you remember questioning the things your teacher said in the first grade? No, we were taking her every word as 100% correct information.

The questioning is taught to us later in the formal education process.

Teachers push us to think about the questions they ask, they organize debates, and we are encouraged to ask questions and raise counterarguments to develop an independent brain.

So, most of the questioning is developed artificially in modern life. Our early information doesn’t leave much room for questioning because they are almost always correct.

As a result, the tendency to rely on information is very strong unless it’s mitigated by formal education. The problem is that formal education is naturally limited. Many people have only so much capacity to develop a deep understanding of a few topics through formal education.

Thus, whenever the topic is out of our expertise, the tendency to rely on information dominates. This blurs the chain of causation and leads to narratives often being perceived as correct information. This increases our confidence in the information and willingness to act on it, even though it was actually a narrative.

Thus, when we see hundreds of headlines in a day, we naturally tend to take them as information, as we can’t have expertise in all of the topics. Narratives replace truth, so we find ourselves urged to react to headlines.

It’s a chain that conditions us to crave information to act:

• We don’t know the world when we were born, and we need information to live.

• Our early experiences produce almost 100% accurate information all the time.

• Accurate information creates a tendency to rely on what’s given as information.

• Questioning is built later with formal education, and it’s strong only when we have depth of knowledge.

• Out of expertise, causal chains are hard to distinguish from narratives.

• When we can’t differentiate, we act with a tendency to rely on what’s given as information.

This is why we crave information.

Our life starts with uncertainty, and we get out of that uncertainty by gathering 100% accurate information. These conditions lead us to crave information when we face uncertainty again, but we lack the ability to distinguish real information defined by causal links from mere narrative in areas where we don’t have deep knowledge.

Thus, our tendency to crave and rely on information makes many narratives appear to be causation-based. Now, account for the fact that all the people feel and act in the same way. What happens? Narratives overwhelm accurate information by a magnitude.

Can you generate alpha relying on this and succeed in the market?

Let’s discuss.

2️⃣ Can You Generate Alpha Through Information?

There is a very dangerous and unfortunate trend I am currently seeing everywhere.

Substack is overwhelmed by deep macro research, thematic research, etc., and people are hopelessly paying for them to generate alpha and succeed in the market.

This idea is insane to me.

Let’s get to the basics, shall we? Macro funds posted their best returns since 2008 last year, and it still looked pretty miserable:

They achieved only around 16% after fees in a year where the S&P 500 returned 18%. If you invested with them since 2015, you are underperforming all major indexes by a margin.

Here is the question—don’t you think all these funds already had access to the top-quality research by Phds in relevant areas?

Of course, they had. The same applies to all thematic funds as well. They have always had the best research, best data, and best tools. They underperformed anyway.

Don’t get me wrong, the problem isn’t the researchers or the writers themselves. The problem is that both the writers and readers are falling into the same trap as I described above.

Writers are creating narratives and confusing them with causation because we all suffer from the same tendency to crave information in the face of uncertainty. Readers are taking what they wrote as accurate information due to the natural tendency to rely on information when the knowledge isn’t deep.

Result? Most macro-research is just narratives, and people are relying on it as if it were accurate information. Normally, it doesn’t work.

Daniel Kahneman provides one of the most solid examples of this phenomenon in his book “Thinking Fast and Slow”:

On December 13, 2003, after U.S. forces captured Saddam Hussein, Bloomberg News reported on the market reaction with the following headline:

• 13:01 Headline: “U.S. Treasuries rise; Hussein capture may not curb terrorism”.

• 13:31 Headline: "U.S. Treasuries fall; Hussein capture boosts allure of risky assets".

What Bloomberg did was just create narratives, not provide accurate information. As the headline incident suggests, you can’t generate any alpha from this, as narratives change all the time with perception.

If you bet believing Saddam’s capture wouldn’t curb terrorism, you would have lost money 30 minutes later as the sentiment changed and narratives evolved.

Accurate information is rare, and it tends to be priced in very quickly; all else is narratives that don’t reliably generate alpha.

In the example, Saddam’s capture was the accurate information, but once it was out, the market positioned itself accordingly. Thus, there are only two ways you can generate alpha from real information:

• You are an insider, so you know it before others.

• You speculate on the events.

In the former case, you are legally prohibited from acting on the knowledge. In the latter, I guarantee you’ll be more wrong than right and end up losing money.

So, the knowledge trap is real.

We build our knowledge mostly on narratives, like the Bloomberg headline above or macro-articles. Our evolutionary tendency to crave information in the face of uncertainty pushes us to take narratives as accurate information in the absence of real, deep knowledge.

The turkey problem illustrates this.

A turkey is fed for a thousand days before it gets slaughtered for Thanksgiving. Until the 1000th day, it thinks that he’ll be fed by the owner the next day, which looks like accurate information, probably the most accurate information in his life. But it’s actually wrong, and he understands this on the 1000th day:

The problem isn’t that accurate information doesn’t exist. It does; the owner knows the turkey will be fried on Thanksgiving, but the turkey neither knows it nor has any means to access that information; it’s within the owner’s mind.

It’s similar for the market movements.

Take Bloomberg’s headline incident above. There is, of course, a reason the market moved the way it did. If they could get into the minds of everybody active in the bond market, they would be able to present a clear picture, but this isn’t possible. So they drew narratives between phenomena.

This could be pretty harmless as the bond market may turn 1% up or down, but it could also be devastating, as in the case of turkey.

Thus, most of the information is useless in generating alpha.

Accurate information is useless because it’s already very well-known and priced, and narratives are useless because they just present a perception and are vulnerable to the turkey problem.

You can understand the uselessness of information by just looking at central banks.

They have all the accurate information, but they need to create the correct links between them so they can correctly speculate about the future, and we don’t get crashes or crises. Yet, they regularly fail to do it.

In a space where the institutions created literally for this job fail, you shouldn’t expect to generate alpha by reading macro-research or diving deep into geopolitical research to predict what comes next. This is illustrated by the macro funds themselves failing to generate alpha. How can you beat them by reading their research while they can’t beat the market?

A thematic AI basket may be doing very well right now, but it’s vulnerable to the turkey problem. What will happen when AI investments dry up? When will it happen, and what will trigger it? Nobody knows, but acting like it won’t happen makes you the turkey.

If you think anybody can see the next theme after AI, and the ones after that consistently, you are again being the turkey. Nobody has been able to do it so far, so why would this change know?

Michael Burry called the 2008 crisis correctly, but he also called the last 8 crises that never happened. He literally posted this in 2023:

The market is up 60% since then. This is not to throw a stone at Burry; it’s just to illustrate the impossibility of predicting market movements and themes based on research and the so-called “knowledge.” Because what it creates is not accurate information, just narratives.

I am sure Burry did a lot of research before calling a crash in 2023, but the links he created in his mind were wrong.

In short:

• Accurate information is rare and tends to offer no alpha.

• Most information is just narratives between phenomena.

• Some links may seemingly work, but believing they’ll keep working makes you a sucker. You become a turkey.

So, if it’s not the information, what creates alpha?

Share

3️⃣ Where Is The Edge?

If you get yourself into situations where you constantly need accurate information to navigate successfully, you’ll fail at some point because information space is dominated by narratives, not by accurate causations. This flows naturally from what we discussed so far.

Thus, the secret of success is getting yourself out of the situations where you need a lot of information and start operating where you need as little information as possible.

The less information you need, the better your ability to find accurate information and eliminate narratives that may or may not be correct.

When do you need as little information as possible?

It’s when you are buying something with a durable competitive advantage.

It has two components:

• Durability: If you are sure that the business will be around 20 years from now, this eliminates a huge burden of information.

• Competitive advantage: If you know that competitors won’t be able to take its business for some reason, you are unburdened from constant market analysis.

Thus, the real edge is not the information, but being able to rely on the lack of information. This is what Buffett is really good at.

Just listen to what he said about Apple:

He doesn’t know how the iPhone works or even its unit economics. What he knows is that people love their iPhone so much that they would give up on their second car rather than their iPhone.

This tells him that as long as we use mobile phones, it’s very hard for others to take Apple’s business. At that point, if you also believed that we would keep using mobile phones for the next 10-20 years, the only extra information you needed to invest was the fair value of the business. Then you would invest or not, depending on the price.

He basically eliminates almost all the need for information and makes decisions based on very little information.

Compare this with the information needs of a macro investor. He probably needs a daily flow of information on currencies, geopolitics, inflation, and employment numbers, etc. That creates an even bigger cloud of information as people interpret it and create narratives. It’s impossible to pick the right links consistently.

Buffett, on the other hand, needs no daily information from Apple, and even quarterly information doesn’t mean much given his mode of operation:

• Durability

• Competition

• Price

Only price can meaningfully change substantially from quarter to quarter, while durability and competition require many more than one quarterly print to allow accurate observation.

As a result, he naturally starts with a longer initial holding period, which is prolonged if his thesis on durability and competition stands. His current average holding period for the top 5 positions that make up 70% of the portfolio is 20 years:

Don’t fall into the mistake of thinking Buffett is an exception in this.

What was the hedge fund that broke a profit record last year? It was Chris Hohn’s TCI. His average holding period is 8 years.

What allows him to do this is doing the same thing as Buffett—he eliminates the need for information by focusing on what matters.

What matters? Hohn says himself that 90% of the things don’t matter and what matters is competition and disruption, i.e., durability:

If you have that type of business, your need for information plummets, so you can think more clearly, hold longer as you are trying to observe very few things from prints that are 3 months away from each other, and sometimes 6 if you are investing outside the US.

That’s the real edge. Getting yourself into situations where the lack of frequent and accurate information doesn’t matter.

For example, two weeks ago I wrote a thesis about a fast-growing discount retailer in Mexico called TBBB. It’s simple to understand, has durability since retail business has been here for ages, and it has a competitive advantage as well, thanks to local economies of scale.

Compare it to drone defense businesses that are recently hyped, thanks to Israel having to use million-dollar missiles to take down $30k Iranian drones. A real asymmetry. People saw this asymmetry, of course, and reasoned that it shouldn’t be this way, and started to invest in companies that develop drone defense solutions.

One company that develops such systems is Aerovironment, which develops laser weapons to counter swarms. Now, put it next to TBBB and compare:

What company do you think requires more information to hold and manage the position? I am not saying Aerovironment is a bad business. It could be a really good business, but investing in it demands a constant flow of information and tinkering:

• You should research what the counterdevelopments are.

• You should understand what the alternatives are already here and in development.

• You should research who really needs and buys it, and who develops their own etc.

These are just a few examples. It requires multiples of accurate information than a discount retailer, and it’s really hard to get accurate information to rely on, eliminating narratives dominating 90% of the information space.

In short, the edge is not access to more and more information and research; it’s distancing yourself away from situations where you need more information and getting into situations where you need very little information.

That’s how you succeed in the market, not by trying to consume every bit of information to stay on top of things.

🏁 Final Words

Most investors think that more information and knowledge are the key to success.

It couldn’t be further from the truth. If it were true, central banks would be very successful in preventing crises and shocks, but they haven’t been.

If the information were the key, researchers would make a bank, and Warren Buffett-Chris Hohn would be street sellers.

What you need is investing in things that require less information, not trying to get more information on hyped stuff.

This is really hard to do. Especially when there is a ton of competent-looking research getting published every day, one thematic investor winning one day and another the next day, and there is a constant stream of people who look like they have made money capitalizing on the information flow.

All those are illusions. The reality is this:

All these funds already have access to the best research and fastest information flow. The result is disappointment.

This is because information is not the edge. The edge is getting into things that don’t require too much flow of information.

Of course, the investment world will try to get you to believe the opposite. This is how the money is made. If you get into things that require more information, of course, you'd better have more and faster information. Otherwise, your returns would look even worse than the funds above. If you have cancer, you'd better get the chemo.

But the best thing you can do is to avoid cancer by eating healthy and doing regular exercise. In real life, that may not be enough, and you may still get cancer. But in investing, you always have the choice. If you don’t get on that path, nobody can push you onto it.

Always remember when you are bombarded with a constant flow of information and research: The real edge is not more information, but not needing a lot of information.

Subscribe now

Share

That’s all friends!

Thanks for reading Capitalist-Letters!

Please share your thoughts in the comments below.

👋🏽👋🏽See you in the next issue!

I often think about this question, especially when the times get hard. Of course, one can come up with many answers like:

• Deeper knowledge.

• Better research.

• Or even luck.

I don’t think it’s any of these.

How, you may ask, “how come deeper knowledge isn’t a factor?”

I have always said …

Read more

Take Mercadolibre as an example. There is a reason it’s called “the Amazon of Latin America.” It took the model and implemented it in an underpenetrated market.

There are many others like MercadoLibre because this is a very fruitful soil for entrepreneurship. Look at what’s working in other markets, and bring it to your own.

These companies tend to be very attractive investment opportunities, as the business model is already derisked. When this is combined with low penetration in their respective market, they can easily grow at an above-average rate, at least for a decade.

The downside of this is that it tends to attract a lot of competition.

As the model is proven and a blueprint already exists, it’s easy for others to follow once somebody starts it.

However… In some cases, the business model and industry allow you to create a durable competitive advantage, like in Mercadolibre’s case. In those cases, the business is well-positioned to exploit the underpenetrated market without getting decimated by the competition, and it can become a giant.

This is one of those opportunities.

This company takes a proven business model and implements it in a deeply unpenetrated market.

Result? Booming business. Take a look at this:

• Revenue has quadrupled in the last 5 years.

• Operating cash flow has almost quintupled since 2020.

• EBITDA margin is expanding despite the hyper-growth phase.

What’s even better is that it’s trading only around 12x forward EV/EBITDA despite the fact that it’s still a very small ~$3 billion company with decades of runway ahead.

Plus, it has one of the most promising emerging moats I currently see in the market.

So, let’s cut the introduction and dive deeper into this gem.

Subscribe now

🏭 Understanding the Business

Private equity insiders are one of the best-positioned people to become successful entrepreneurs. They observe many portfolio companies and investment opportunities, thus they have a high chance to spot winning models early.

Read more

The legendary CEO of General Electric, Jack Welch, puts this eloquently:

This principle directly applies to investing as it’s essentially a game.

It’s a game that we play to make money. The problem is that there ar…

Read more

Portfolio is now up 27% in the last twelve months against 16% of the S&P 500!

“It’s not going to be an easy year.”

This was the theme I emphasized in the last several portfolio updates and almost all my trade updates since September.

One of the strengths of our community is that it goes beyond the writer-reader relationship. We are always in very close contact through our Discord community, and I feel like we have come to see ourselves as a group of friends.

We have done very good trades together, made money, and also naturally had some losers. All these have resulted in a strong sense of friendship in the community.

This environment enabled us to be very straightforward and act with a sense of responsibility.

This is why I have been very clear for the last few months that:

• 2025 was not an easy year, as 75% of large-cap managers underperformed.

• We outperformed because we were exceptional in stock picking and also lucky.

• 2026 will be an even harder year as everything has peaked, including uncertainty.

Toward the end of 2025, the picture was this: Equities, metals, and commodities were stretched, and economic-geopolitical directions were uncertain.

This is why I said a downward move in 2026 was more likely than an upward move.

This is what history taught us. Whenever the forward P/E of S&P goes above 22x, it drastically comes down at some point over the next two years:

To be prepared, I laid out a three-pillared strategy:

• Cut risk assets that do not have much legroom for further appreciation.

• Diversify away from the US and into the emerging markets.

• Keep an eye on opportunistic purchases in the US market.

This is why we began growing our cash position last October and went from 10% cash position to almost 20%.

We started to see these predictions play out. Forward P/E has come down below 22x as investors rotated out of the more expensive tech stocks with a higher weight in the index to more defensive sectors.

This is why many growth investors saw their portfolio down between 10-20% while the index stayed stable near its all-time high.

Though we prepared for this by exiting the riskiest assets in our portfolio, we weren’t fully immune, as we always prefer to stay invested at all times. As a result, our portfolio was down by 7% YTD, which is still pretty solid given that the Nasdaq is down by 4% and many growth portfolios are down by +10%.

Another piece of good news is that our drawdown is mainly due to a pullback in two of our biggest and permanent positions—SoFi and Amazon.

We see these two as permanent positions, so we ride ups and downs in the positions, not trade them, and consider pullbacks as buying opportunities, so our YTD drawdown doesn’t bother me at all. Quite the opposite, it excites me as betting on these stocks at the current levels will be the drivers of the alpha in the future.

I want to remind everybody that this is how you should feel about the downturns.

I have been trying to explain to people who are coveting my returns for the last three years that the seeds of these returns were planted at times nobody wanted to invest in the markets.

When nobody wanted SoFi back in 2023, I was buying. When nobody wanted Nebius last April, I tripled down at around $20 per share.

Yet, I still see some people trying to position themselves according to what may come next in the markets. Please don’t.

Peter Lynch puts it best:

Investing in stocks is simply owning pieces of the business. Can you imagine being a sole owner of a business and trying to sell when you think the market will go down and buy again when you think it’ll go up? Sounds insane, right? Doing the same in stocks is equally insane; they are basically the same things.

If you are holding exceptional businesses bought at attractive prices, you have nothing to be afraid of. You should stay in the position, be unemotional, and do nothing.

Yes, you’ll go down with the market, but this is part of the game. If the companies you are holding are exceptional, they’ll keep increasing their intrinsic value even in down markets, and the stock price will rise back stronger when good times come again.

And even beyond staying in position, ideally, you should be buying at those times.

This is why we grew our cash position. Not to sit on it, but to buy more of the exceptional businesses we own in case of a downturn.

Why should I be worried? There are signals that I may get these opportunities, as I already positioned myself accordingly as I explained above, and trying to prepare you for months. I should be excited. You should be excited.

Most people in markets are not too stupid to see quality after a while, but 99% still don’t have the emotional strength to buy when they are cheap or hold through downturns. We capitalize on their weakness. When they break and sell us bargains, we buy them, and we are happy. It’s not more complicated than this.

This is how we have always operated, and we’ll always operate this way.

I think this sane and measured strategy will deliver for us now and in the future as it’s delivered so far:

Our portfolio is now up 27% in the last twelve months against 16% of the S&P 500!

The following transactions took place in our portfolio in February:

• Exited 1 position.

• Increased 2 positions.

• Opened 1 new position.

In the previous updates, I provided my outlook for each company in the portfolio. In this update, I’ll also provide an outlook for our new position and provide a general commentary about the portfolio and strategy going forward.

Let’s dive in!

🚨Our portfolio is proprietary to the members supporting the publication🚨

Members get portfolio updates every month.

Here is a 25% discount to celebrate our new readers!

Valid only until tomorrow!

Get 25% off for 1 year

📊Here Is Our Full Portfolio!

As of today, we have 18 holdings in our portfolio.

8 of these positions can be considered foundational.

Read more

Early Buffett & Graham style. I was looking at obscure corners of the market, buying positions with extreme illiquidity and no coverage. Bid-ask spread would be insane, so it would take weeks for me to build positions.

This method works when you do it right, but it is tiring. It’s a constant effor…

Read more

The most successful business ideas emerge from founders’ own problems or from the inefficiencies they observe in the industries they work in.

Shopify is a great example.

Tobias Lutke wasn’t trying to build an e-commerce software company. He wanted to create an online store to sell snowboards. The site would be called Snowdevil…

Read more

Now you may ask—how come you can hate earnings season as an investor?

Well, earnings seasons are like rock festivals.

There are only a few good groups worth listening to, and you only want to listen to specific songs of them. Yet, people around are euphoric, fueled by alcohol, and they take every single group and song with a screa…

Read more

It has become something like a motto among tech entrepreneurs as we watch everything turning into a software-based digital product.

Software made billionaires, software created jobs, sof…

Read more

Investing is very similar to this.

In good times, people get used to the market going up. Because the overall economic environment is friendly, money is abundant, and failures can easily be tolerated due to the abun…

Read more

Portfolio is now up 39% in the last twelve months against 14% of the S&P 500!

It’s been another great year for our portfolio. When I published the portfolio update in December, our portfolio was up around 44% in the trailing twelve months, and this was already better than all of the top hedge funds that publicly disclosed their performance.

Yet, it was not a complete picture because we didn’t have the annual numbers; we only got the numbers for the 9 months ending on September 30. Now we have them:

As you see, our portfolio would trail only Apis Flagship, Melqart Opportunities, and Whale Rock Opportunities funds in 2025.

What’s better is that it would top the list if we compare 2-year performances starting from 2024. If you started from January 2022 and looked at the total return for the 4-year period ending on December 31, 2025, our portfolio would top the list by a margin.

I am not telling all these to brag and praise the sophistication of our strategy. I am trying to do the exact opposite, because if you know me and know the portfolio, you understand that I keep things very simple.

I try to buy exceptional businesses at fair valuations and hold them until I think fundamentals deteriorate. That’s it. What I think creates all the difference with those funds and other portfolios in terms of consistency is a few small but very important points. Let me dwell on those points for a bit, to elaborate on our strategy.

First things first, I am managing my personal portfolio, so I am not under any stress from clients. Nobody asks me why I am not buying the current sexy thing, which is essentially the crux of our performance. We can afford to stay independent-minded.

Second, I understand and care about the base effects. Most quality-oriented portfolios operate with a strict buy-and-hold strategy. I don’t, at least not for the totality of the portfolio. As you know, we run a barbell portfolio:

• At one end, we have market-leading businesses bought at attractive prices, which I call foundational positions.

• At the other end, we have fast-growing businesses with strong balance sheets and some kind of competitive edge, but not necessarily a moat.

Ideally, I never want to sell foundational positions. With them, our goal is to increase our ownership over time whenever they become attractive, using proceeds from growth positions or fresh cash. We expect just a bit above average but consistent returns from this portion of the portfolio.

When it comes to growth positions, we can exit or trim them. Once they reach fair value, further upside can only come from multiple expansion, which extends our downside in those positions. So, if we keep holding them, there are only two options:

• Very limited upside in the following periods due to base effects.

• Extended downside as they are more reactive to downturns.

In these cases, we exit or trim the growth positions and look for other growth opportunities that have larger upside than downside. This rotation requires skilled stock picking, but when done right, enables a portfolio to beat base effects and sets it up for sustained outperformance.

In the last few years, we have been extremely skillful in stock picking.

I say “we” because some of our best positions were pitched to me for the first time by the members of our community. Though I was the one who made decisions, finding the right business was still the majority of the work, because when something is really great, it’s not that hard to see it.

This strategy has worked very well for us, and in the last four years, we beat the market by increasing margins. This is why I said many times last quarter that we were repositioning the portfolio for continued outperformance by adding new positions.

That already started to work for us.

This month, we exited one of the positions we opened just in the last quarter for over 30% gain, and one position for around 10% gain. We added two new positions instead. One of the new positions is already up by 10%, while the other one is up by 5%.

This helped us offset a substantial base effect coming from strong performance in January 2025, and kept the performance close to ~40% levels, despite the marketwide choppiness.

I honestly believe we could have done much better, as I skipped buying some of the stocks I published research about to preserve my cash position. Fluence Energy is one of them, and the stock has doubled since I published my research back in October.

Though I wish I bought it, I don’t care much about it. Looking at our portfolio, I see several growth positions that should drive alpha this year. Plus, none of our positions, except one, is substantially overvalued, so we don’t have much downside as well.

All these make me very confident that we’ll deliver another great performance this year, and our strategy will keep delivering as it’s delivered so far:

Our portfolio is now up 39% in the last twelve months against 14% of the S&P 500!

The following transactions took place in our portfolio in November:

• Exited 4 positions.

• Increased 1 position.

• Opened 2 new positions.

In the previous updates, I provided my outlook for each company in the portfolio.

In this update, I’ll also provide an outlook for our new positions and share my overall portfolio commentary and strategy for 2026.

Let’s dive in!

🚨Our portfolio is proprietary to the members supporting the publication🚨

Members get portfolio updates every month.

Here is a 25% discount to celebrate 23,000 readers!

Valid only until Monday!

Get 25% off for 1 year

📊Here Is Our Full Portfolio!

As of today, we have 18 holdings in our portfolio.

Read more

It’s because private equity has dual drivers of investment return:

• Organic and inorganic growth

• Multiple expansion

Think about it.

You buy a B2B service company, let’s say a cleaning company, which makes $10 million net profit a year. Then you buy another company that provides a complementary service, let’s say security, which also does $10 million a year.

You combine them together under the same corporate umbrella, cut the fat in back-office functions, cross-sell, and each company is now netting, let’s say, $14 million a year. This is the first driver of investment return.

When they were separate, they would get a lower exit multiple because they were smaller companies. Bigger companies make their owners richer; this is why they get a higher multiple.

So, when separate, each would get 4x exit multiple, and their combined value would be $112 million, even if we assume each would still grow its net profit to $14 million. But together, we are looking at a bigger company that gets a higher multiple, let’s say 7x. In this case, the combined entity is now worth $196 million.

The combined entity is worth 75% more than the total value of the separate companies. Effect of the multiple expansion due to size. The second driver of investment return.

The problem here is that private equity funds don’t generally have permanent capital, so they need to exit and pay their investors after 5-7 years and repeat the strategy.

Yet, this exposes the firm to the same risks—investment risk, and execution risk—all over again, jeopardizing the future gains. This is why I don’t own any private equity companies despite their unfair structural advantages.

Had they permanent capital, allowing them to buy and hold exceptional businesses forever, that would have benefited public shareholders much more. It would also be structurally similar to what Berkshire was to Warren Buffett, an investment vehicle to own exceptional companies.

This company does exactly that.

They own one of the most exceptional businesses in the world, which is growing rapidly in its industry, and the parent company is growing alongside it.

Take a look at this:

It grew revenues over 23% annually in the last five years, and the parent company grew alongside it at a similar rate.

I have been following the company for some time, but the stock consistently traded at a premium to its net asset value. Now that premium has eroded, the company is likely valued less than what’ll be its net asset value at the end of this quarter.

This is why I am pulling the trigger now and buying it. It’s hard to find an opportunity to buy such an exceptional business at net asset value in this broadly overvalued market.

So, let’s cut the introduction here and dive deeper into this gem.

Subscribe now

🏭 Understanding The Business

Access to capital has always been a big problem for small businesses (SMEs).

They are seen as riskier by big lenders, so they are often pushed to predatory lenders that charge high interest rates with unfavorable conditions.

Read more

This is obviously true.

Given the ways markets are structured, being right doesn’t mean anything. If everybody else is also right with you, what you get is the market return. Mediocre.

To be really successful in the market, to make the big…

Read more