Victoria Bazir
That misses the more important question. What happens to the rest of the market when a handful of AI companies become vacuum cleaners for global liquidity at the same time?
Michael Burry’s warning about future listings from OpenAI, SpaceX, and Anthropic matters less as a prediction and more as a reminder of how market structure works at extremes. Bull markets usually look strongest right before capital starts getting redistributed underneath them. And historically, that transition can happen fast.
The Real Parallel With 2000 Isn’t Valuation
In 2000, the Nasdaq did not collapse because investors suddenly discovered the internet was useless. The crash came after an unprecedented flood of equity issuance exhausted available capital.
That year, 446 IPOs raised more than $108 billion - a record at the time. The Nasdaq peaked in March and eventually lost roughly 80% over the following years. The mechanics were simple. Funds needed cash for new offerings. To get cash, they sold existing winners. The selling pressure first appeared in the exact tech names investors assumed were untouchable.
The parallel with today breaks in one important way. Back then, capital dispersed across hundreds of companies. This cycle is compressing into a few gravitational centers. That concentration changes the risk profile entirely.
Dot-Com Peak vs AI IPO Cycle
| 2000 Dot-Com Era | Current AI Cycle |
| 446 IPOs flooded the market | A few mega-IPOs may dominate capital flows |
| Capital dispersed across hundreds of firms | Capital concentrated into several AI giants |
| Retail speculation drove momentum | Institutional and hyperscaler capital drive momentum |
| Many companies lacked real revenue | AI leaders already generate enterprise demand |
| Liquidity dried up gradually | Passive ownership could accelerate rotations |
| Nasdaq leadership was relatively broad | Market gains are concentrated in mega-cap tech |
AI’s Ownership Structure Is More Fragile Than It Looks
The modern AI market behaves less like a diversified ecosystem and more like a tightly connected dependency chain.
Microsoft’s cloud backlog is deeply tied to OpenAI demand. Oracle has aggressively positioned itself around AI infrastructure exposure. Nvidia remains the clearest public-market proxy for model training economics. Amazon and Google are simultaneously financing and competing against the same model ecosystem.
The same institutions overweight these companies would almost certainly need exposure to any OpenAI or Anthropic IPO on day one.
So where does that money come from? Usually from selling something else. Potentially the same mega-cap tech positions currently carrying the index. This is the structural tension underneath the AI trade that the market still seems reluctant to price in. Investors talk about AI as a growth theme, but the ownership structure is becoming increasingly circular.
Unlike the dot-com era, today’s AI leaders are already embedded inside the balance sheets of the world’s largest public companies before going public themselves.
That matters because modern markets are far more index-driven than they were in 2000. Passive funds, sovereign wealth allocations, pension exposure, and mega-cap concentration mean the same pools of institutional capital effectively own everything simultaneously.
A massive IPO wave no longer pulls liquidity from the edges of the market. It pulls directly from the core.
Why This IPO Wave Could Matter More Than Investors Expect
- The same funds buying AI IPOs already own mega-cap technology stocks
- AI exposure is increasingly concentrated across public markets
- Passive investing may amplify liquidity rotations instead of softening them
- Public markets could become exit liquidity for mature private AI ecosystems
- Massive IPO demand can pressure existing market leaders before sentiment changes
Why “This Time Is Different” May Only Be Partly True
Bulls would argue that comparisons to 2000 are overstated. Today’s AI companies generate real enterprise demand, hyperscalers have enormous balance sheets, and institutional capital pools are vastly deeper than they were during the dot-com era. OpenAI is not Pets.com. SpaceX is not a speculative fiber-optic startup burning cash without revenue visibility.
But deeper liquidity does not eliminate liquidity constraints. In some ways, passive ownership and concentrated positioning may amplify them.
The AI boom may become the first major tech cycle where private capital formation grows so large that public markets themselves become the exit liquidity provider for an already mature ecosystem.
The Risk Is Liquidity, Not AI Itself
The warning sign from 2000 was never valuation alone. It was deterioration in funding conditions after peak issuance.
In the first quarter of 2000, IPOs were still doubling on their opening day. By the fourth quarter, gross proceeds had collapsed, first-day gains deteriorated sharply, and capital markets tightened almost overnight.
Speculative systems tend to fail through funding pressure before sentiment fully adjusts. That is why the coming AI IPO cycle matters even if demand for AI remains fundamentally strong.
If OpenAI, SpaceX, and Anthropic eventually pursue public listings within a compressed window, the broader market may discover that enthusiasm and liquidity are not the same thing.
None of this guarantees a repeat of 2000. AI infrastructure demand is real, revenues are real, and institutional balance sheets are stronger than they were two decades ago. But markets rarely break because investors see the risk clearly in advance. More often, they break because too many participants quietly assume liquidity will always be there when they need it.
Victoria Bazir
