The Making of a Too-Big-to-Fail Bubble

AI's $2 Trillion Problem—And Why Taxpayers Will Pay For It

During a recent call with Phyle members—our private community where we exchange ideas and investment insights—Matt Smith made an observation that perfectly captures the paradox we’re facing in AI markets today.

“The technology is real,” he said. “It’s genuinely transformative. But the market is in a bubble.”

Matt is, of course, absolutely right on both counts.

And that distinction matters, because it’s the line between recognizing a true technological revolution and getting swept up in a speculative mania. Right now, we’re watching both unfold at the same time.

Let me show you what I mean.

The Valuation Problem

Did you know that about 40% of every dollar you invest in a basic S&P 500 index fund now flows to just ten companies: Nvidia, Microsoft, Apple, Alphabet, Amazon, Broadcom, Meta, Tesla, Berkshire Hathaway, and JPMorgan?

Now, not all of these are pure AI plays—Berkshire and JPMorgan certainly aren’t. But the tech giants dominating this list are the ones pouring hundreds of billions into AI infrastructure. And if you look at Nvidia—the poster child of the AI boom—it alone captures nearly 8% of every dollar going into the index.

And that’s where the valuation problem really starts.

According to Bain & Company—one of the world’s three largest consulting firms—AI companies need to generate around $2 trillion annually in new AI-driven revenue by 2030 just to justify their current stock prices.

Let me put that number in perspective.

Here are the approximate 2024 revenues for the biggest tech players—the same companies pouring the most money into AI:

• Apple: ~$391 billion

• Amazon: ~$638 billion

• Alphabet (Google): ~$350 billion

• Microsoft: ~$245 billion

• Meta: ~$164.5 billion

• Nvidia: ~$90 billion

Combined, these six giants—Apple, Amazon, Microsoft, Meta, Nvidia, and Google—generated roughly $1.9 trillion in 2024 from everything they do. iPhones. Cloud services. Search advertising. E-commerce. Social media. Gaming chips. All of it.

Think about that for a moment.

These aren’t small companies struggling to gain traction. These are the most dominant technology firms on the planet, with decades of infrastructure, billions of users, and near-monopolistic market positions. And even with all those advantages, the revenue AI needs to generate would exceed everything they currently do combined.

Bain also notes that even assuming significant productivity gains from AI adoption and full reinvestment of on-premises IT savings into cloud infrastructure, the industry would still face an $800 billion revenue shortfall. In other words, AI is already priced for a future that even the most optimistic projections can’t support.

The main players know this is a problem. Overpromising and underdelivering could collapse the entire AI narrative—and the valuations built on top of it.

So naturally, they turn to the government.

Sam Altman, the CEO of OpenAI and creator of ChatGPT, said something interesting recently in an interview with economist Tyler Cowen. He suggested that when a technology becomes “sufficiently huge,” the federal government ends up acting as “the insurer of last resort.”

Translation: he’s positioning AI as a national-priority project—something the government (meaning taxpayers, since the government has no money of its own) will be expected to fund when the economics stop making sense. In other words, he’s framing AI as too big to fail.

The Money-Go-Round

And here’s the thing. The economics of AI already don’t make sense.

You see, the current AI boom isn’t being funded by profits. It’s being fueled by debt—much of which doesn’t even appear on corporate balance sheets.

How are they hiding it? Through something called circular financing—a mechanism where money flows in a loop between companies, creating the appearance of revenue and growth without generating real economic value.

Here’s how it works:

Nvidia invests billions in OpenAI. OpenAI commits to massive cloud contracts with Oracle and Microsoft to run their AI models. Oracle and Microsoft then spend tens of billions buying Nvidia GPUs to fulfill those contracts. Nvidia takes that revenue and invests more back into OpenAI.

The loop closes: Nvidia → OpenAI → Oracle → back to Nvidia.

Everyone reports growth. Everyone looks successful. But the money being spent is borrowed, and no real economic value is being created—just the appearance of it.

This spending runs through private credit facilities, special purpose vehicles (SPVs), joint ventures, and circular financing arrangements that keep much of the debt off official balance sheets.

The whole “operation” pushes their stock prices up and makes the overall market look strong.

But it only looks strong because these companies are passing the same money around to each other, amplifying it at each turn.

The scale of this charade matters enormously. According to Deutsche Bank, if it weren’t for AI-related capital spending right now, the U.S. economy would already be in recession. In other words, these AI expenditures—largely funded by borrowed money being passed in circles—have become a meaningful driver of GDP growth, which means Washington now has a vested political interest in keeping this bubble inflated.

So if AI spending slows and investors start questioning the sector, companies will fail and the economy takes a hit. And you can bet that because AI has been framed as strategically critical to America’s competition with China, there will be enormous political pressure to bail these companies out with taxpayer money.

That’s why Sam Altman can openly suggest the government will serve as the “insurer of last resort” without anyone batting an eye.

So yes—the technology is real. AI will transform industries, eliminate jobs, create new ones, and reshape how we work and live. But make no mistake—there’s a real bubble wrapped around it.

And if I were a suspicious person, I’d say there’s a bit of a con being pulled here too. Spend enough money, grow large enough, become “critical infrastructure,” cozy up to the government, and you’ll never have to face the consequences of your own bad bets. The taxpayers will.

Regards,

Lau Vegys

P.S. Of course, none of this means the bubble won’t burst. Even with Washington in the mix (and they always pick winners and losers), a shakeout is inevitable. To drive that point home, later this week I’ll show you why trying to pick the AI survivors is a losing game, and where the real opportunities actually are.