AI Boom's Dot-Com DNA: Leverage, Demand and the Bust Ahead

Henry Blodget, the former CEO of Business Insider who lived through the dot-com collapse, sees the current AI spending cycle as a replay of the same three-act drama: explosive demand, mounting leverage, and a guaranteed bust. His argument, laid out in a recent newsletter, is that the AI boom will end the way most technology manias have ended – with a crash that wipes out the first wave of companies and investors, followed by a longer, more productive second wave.

Blodget's central case is simple and historically grounded. Every speculative bubble needs two ingredients, he writes: real growth and leverage. The internet boom had both. Millions of new users came online each month, and companies borrowed hundreds of billions of dollars to build telecom networks. The housing bubble had the same recipe: low interest rates and mortgage innovation drove a building boom, and leverage from lenders and homeowners amplified the cycle. In both cases, the music stopped when supply caught up with demand and the financing dried up.

The AI boom has the same structure, Blodget argues, but at a vastly larger scale. He points to Anthropic, the AI startup whose revenue exploded from zero to an annualized run rate of $65 billion in roughly two years. That is a 7x increase from last year, he notes. The growth is so fast that no company in history, to his knowledge, has ever matched it. The Wall Street Journal also reported that Anthropic is now profitable, which Blodget says "nukes" the widespread claim that AI economics don't work. He contrasts this with OpenAI, which he calls "the Netscape of the AI era" – the company that kicked off the boom but is now losing ground. OpenAI's sequential revenue growth slowed to 18% in the second quarter, he writes, citing the Journal, a sharp deceleration from earlier periods.

The demand side of the ledger looks strong, but the leverage side looks terrifying. Blodget singles out Alphabet, which in the second quarter burned cash for the first time in its history as a public company. He points to its "off balance sheet" commitments, which increased by nearly $500 billion in three months. Alphabet's free cash flow was about $25 billion a quarter last year, he notes, meaning the company would need five years of that cash flow to pay off the commitments it made in a single quarter. Alphabet itself is an extreme example, he says, but the broader pattern is similar. The tech giants – Alphabet, Meta, Amazon – have shifted from generating tens of billions in free cash flow to burning cash on AI infrastructure.

Then there are the circular financing deals that Blodget says amplify the entire ecosystem. Nvidia announced an agreement to invest $1.5 billion and provide $105 billion in credit to build a data center in Ohio. The data center owner will presumably use some of that money to buy Nvidia chips. Separately, Nvidia agreed to invest $30 billion in OpenAI, which will use some of that cash to lease data center space. None of this is illegal or even unusual, Blodget says. But it turns Nvidia's cash and borrowing capacity into financing for future purchases of Nvidia chips. It allows OpenAI and other AI providers to buy far more chips and compute than they otherwise could. In short, it is leverage.

Blodget does not claim to know when the bust will hit. He cites the lesson of both the internet collapse and the financial crisis: even the best investors miss the top. He names Stan Druckenmiller, who held on too long in 2000, and the bank CEOs who over-levered into mortgages in the early 2000s. They knew everything that could be known, and they still got it wrong. The same uncertainty applies now. The AI boom may be akin to 1997, with years of expansion ahead, or to late 1999, when the music was about to stop. The only safe approach, Blodget writes, is to invest, learn, and take advantage of the new technology without betting more than you can afford to lose.

The piece closes with Blodget's personal stake: he owns stocks, does not run an AI company, and hopes there is never a bust. But he has lived through two of them, and he sees the same shape forming again.