The Two-Company Economy: Why the Entire AI Boom Suddenly Feels More Fragile

Salutations, Olio aficionados! 👋

Welcome to the 258th edition of Weekly Olio. We’re thrilled to introduce a fresh new twist to your Sundays: Publisher Parmesan, our hand-picked, thoughtfully crafted edition designed to spark inspiration and insights for the week ahead.

It’s the perfect way to unwind, recharge, and prepare for the week with something truly worth savoring.

If you’re new here and you’re looking for more long-form, crispy writing, click the link to subscribe under this GIF 👇

Those requests came from non-traded BDC investors in Q1 2026, and most got back roughly half of what they asked for. Moody's U.S. BDC sector outlook: Negative.

On Percent's marketplace that same quarter: new issuances, scheduled payments, 0.44% lifetime net loss rate on asset-based deals since inception.† The difference is structural: concentrated corporate loans with redemption windows that close at manager discretion vs. asset-based finance with 6–24 month deal terms. 14.6% net ABS returns LTM after losses (3/31/26).† Starting at $500.

See how private credit works when it's built for individual investors.

_{Alternative investments are speculative. No assurance can be given that investors will receive a return of their capital. †Past performance is not indicative of future results. Terms apply.}

For the last two years, the artificial intelligence boom has looked unstoppable. Nvidia became one of the most valuable companies in history. Microsoft, Google, Amazon, and Oracle committed hundreds of billions of dollars toward AI infrastructure. Startups reached valuations that would have seemed absurd even during the peak of the internet era.

But beneath all the excitement sits an uncomfortable reality: an enormous portion of the AI economy now depends on just two companies — OpenAI and Anthropic.

That concentration is beginning to change how sophisticated investors think about the market. Is this the early formation of a new technological operating system, similar to the rise of cloud computing and the internet? Or are we watching a circular capital structure where venture money, cloud commitments, and infrastructure spending are feeding one another in ways that look larger than they really are?

The more interesting question is not whether AI is real. It clearly is. The question is whether the current financial architecture supporting the AI boom is sustainable at its current pace.

And increasingly, some of the smartest people in technology and investing are beginning to ask the same thing.

Silicon Valley usually changes slowly until suddenly.

The internet took years before Google emerged as the dominant search engine. Social media took almost a decade before Meta became untouchable. Even cloud computing evolved gradually, with Amazon Web Services building its lead over many years.

The AI market has behaved differently.

Just months ago, OpenAI looked untouchable. It had the strongest consumer brand, the fastest adoption curve in technology history, and a near-monopoly on public attention. If you asked investors in late 2024 who would dominate generative AI, almost everyone gave the same answer.

Then something unusual happened.

Anthropic stopped acting like a research lab and started acting like an enterprise software company.

Within months, the narrative shifted. OpenAI remained the cultural leader, but Anthropic became increasingly viewed as the preferred platform for businesses, developers, banks, law firms, and enterprises looking for dependable AI infrastructure.

That kind of transition almost never happens this quickly.

In venture capital, number-two players rarely become number one overnight. Markets typically reward incumbency, ecosystem depth, and distribution. Yet Anthropic managed to reposition itself with extraordinary speed.

The reason matters.

Anthropic understood earlier than most that large language models alone are not the product. The product is the workflow built around them.

That distinction may end up defining the next decade of AI.

The market’s enthusiasm around Anthropic is not difficult to understand.

While much of the AI ecosystem remained focused on demos, image generation, and consumer excitement, Anthropic moved aggressively toward business utility. It built tools that could slot directly into professional environments where people already spend money.

Law firms. Banks. Consultants. Software developers. Large enterprises.

This sounds obvious in hindsight, but most transformative technologies initially fail because they remain impressive rather than useful.

Anthropic’s real breakthrough was not necessarily technical superiority. It was operational positioning.

The company recognized that enterprises care less about novelty and more about reliability, controllability, compliance, and workflow integration. In other words, businesses wanted AI that behaved less like magic and more like infrastructure.

That strategic shift changed investor perception almost immediately.

At one point, private market conversations reportedly floated valuation expectations approaching $1 trillion for Anthropic. Whether those exact numbers materialize or not is almost secondary. What matters is that sophisticated investors are even willing to entertain them.

And that tells you something profound about where capital markets believe AI is heading.

The excitement around AI infrastructure spending has produced extraordinary numbers.

Microsoft, Amazon, Google, and Oracle are all committing massive amounts of capital toward data centers, GPUs, networking infrastructure, and long-term compute capacity.

On earnings calls, the growth rates sound astonishing.

But buried inside those numbers is a structural concern that increasingly sophisticated investors are beginning to focus on.

A surprisingly large portion of the AI infrastructure economy depends on just two customers: OpenAI and Anthropic.

In some cases, commitments from those two companies reportedly account for nearly half of cloud providers’ AI-related backlog growth.

That concentration changes the risk profile dramatically.

Because once enough money flows through a small number of entities, the market stops looking like broad-based technological adoption and starts looking like financial dependency.

This is where some investors become uncomfortable. Not because they doubt AI.

But because they begin asking whether the revenue is truly diversified, durable, and independently generated.

If a large portion of hyperscaler growth ultimately traces back to two AI labs heavily funded by venture capital and strategic partners, then the ecosystem begins to resemble a circular capital structure.

Capital flows into AI labs. AI labs commit enormous spending to cloud providers. Cloud providers report explosive AI revenue growth. Markets reward cloud providers and infrastructure companies. Those gains fuel even more investment into the ecosystem.

At a certain point, investors begin asking a very uncomfortable question:

How much of this growth is organically self-sustaining, and how much is internally reinforced by the same pool of capital?

Most churn blindsides the CSM in renewal week. Champion left. Usage dropped. NPS slid months ago.

A colleague in Slack watches the signals around the clock. Your CSMs catch every risk months before renewal.

11,000+ teams use Viktor daily. SOC 2 certified. Your data never trains models.

Get Started for Free

To understand why this cycle became so extreme, you have to understand Nvidia.

For most of its history, Nvidia was viewed primarily as a semiconductor company serving gaming and specialized computing markets.

Then AI transformed GPUs into the most important industrial asset in technology.

Suddenly, every major company needed compute. And Nvidia became the only company capable of supplying it at scale.

That dynamic created something rare in modern markets: a genuine infrastructure bottleneck.

Historically, transformative technological periods tend to create one foundational layer that captures disproportionate value. During the internet era, it was networking and operating systems. During mobile, it was smartphones and app ecosystems.

In AI, it became compute. This is why Nvidia’s rise has been so violent. The company effectively became the arms dealer for the entire AI race.

And once markets realized that every ambitious AI company would require extraordinary amounts of compute, capital allocation across the ecosystem accelerated dramatically.

The problem is that infrastructure booms often create excesses before they create stability.

Railroads did. Telecom did. The internet did. AI may be following the same pattern.

One of the hardest things in investing is distinguishing genuine demand from financial demand. Real demand comes from customers paying for products because those products solve meaningful problems. Financial demand comes from capital availability.

For now, AI contains both.

There is clearly genuine utility emerging. Enterprises are adopting coding assistants. Customer support is changing. Research workflows are accelerating. Legal and financial analysis are becoming partially automated.

That part is real.

But there is also an enormous amount of infrastructure spending happening ahead of proven monetization. The market is essentially pricing AI as though it will become a foundational layer across every industry simultaneously.

Perhaps it will. But historically, technological revolutions rarely monetize as quickly as infrastructure gets built.

The internet changed the world, but many early internet businesses still collapsed. Cloud computing became enormous, but plenty of cloud-era companies disappeared. The smartphone transformed consumer behavior, yet most mobile startups failed. Technological inevitability does not guarantee financial inevitability.

That distinction matters right now more than almost anything else.

Despite all these concerns, capital continues flooding into AI. Why?

Because investors understand something equally important.

If AI does become the next operating layer for business productivity, then the winners could become some of the largest companies in history.

And when markets believe a platform shift is underway, valuations stop being anchored to present cash flows and start being anchored to future control.

That is what happened during previous technological transitions.

Google looked expensive. Amazon looked expensive. Meta looked expensive.

In hindsight, all of them were probably underpriced relative to the scale they eventually achieved.

This is why investors continue paying extraordinary multiples for companies like OpenAI and Anthropic.

Not because the numbers necessarily justify it today.

But because if either company becomes the default intelligence layer for global enterprise workflows, today’s valuations may eventually look conservative.

The market is not valuing current revenue. It is valuing potential centrality.

One of the biggest misconceptions in AI is that the competition is primarily about building the smartest model.

That phase is already ending.

Increasingly, the market is shifting toward something much more familiar: distribution, workflow integration, enterprise trust, and ecosystem control.

In other words, AI is starting to behave less like a research competition and more like a traditional software platform war.

This is precisely why Anthropic’s rise matters.

The company demonstrated that enterprises value predictability and operational usefulness at least as much as raw intelligence.

Over time, the companies that win may not necessarily be those with the best models. They may be the ones best positioned inside daily workflows. That distinction sounds subtle.

It is not.

Historically, platform companies become dominant not merely because they are technologically superior, but because they become difficult to remove from the surrounding ecosystem.

Microsoft mastered this. Google mastered this. Amazon mastered this.

AI is now entering the same phase.

The entire AI ecosystem currently assumes continued acceleration.

More spending. More infrastructure. More adoption. More demand.

But if growth slows even modestly, the concentration risks become much more visible.

If OpenAI or Anthropic reduce spending commitments, cloud providers feel it immediately. If enterprise adoption takes longer than expected, infrastructure utilization weakens. If monetization lags investment cycles, margins compress across the ecosystem.

None of this means AI collapses.

In fact, AI could still become one of the most transformative technologies in modern history while simultaneously producing periods of financial dislocation.

Those two outcomes are not contradictory. They are often historically connected. The internet changed everything. It also produced the dot-com crash. Railroads transformed economies. They also created speculative bubbles.

Sometimes the technology is right while the timing of capital is wrong. That may ultimately become the defining question of this cycle.

The most important technological revolutions often look irrational while they are happening.

Valuations feel disconnected from reality. Infrastructure spending appears excessive. Capital floods into a small number of companies at extraordinary speed.

Then, years later, historians look back and realize the technology itself was real all along.

The difficult part is determining which companies survive the transition from excitement to durable economics.

Right now, the AI economy increasingly resembles a two-company system sitting on top of one massive infrastructure layer.

That structure can produce extraordinary outcomes.

It can also produce extraordinary fragility.

And somewhere between those two possibilities lies the real story investors are trying to understand.

Interested in learning more about AI? Check out our previous coverage here:

Connect your email, calls, product data and more, and Attio instantly builds your CRM with enriched data and complete context. Whether you’re running product-led growth or enterprise sales, Attio adapts to your unique GTM motion.

Then Ask Attio to plan your next move.

Run deep web research on prospects. Update your pipeline as you work. Find customers and draft outreach emails. Powered by Universal Context, Attio's intelligence layer, Attio searches, updates, and creates across your data to accelerate your workflow.

Ask more from your CRM.

Ask Attio

That’s all for this week. If you enjoyed this edition, we’d really appreciate if you shared it with a friend, family member or colleague.

We’ll be back in your inbox 2 PM IST next Sunday. Till then, have a productive week!

Peeyush and Mohit

Disclaimer: The views, thoughts, and opinions expressed in the text belong solely to the author, and not necessarily to the author's employer, organization, committee or other group or individual.