The queue is longer than the runway

Something is happening in the IPO pipeline that deserves more scrutiny than it's getting. A legitimate cohort of AI infrastructure companies — businesses with real revenue, real customers, and real compute costs — is preparing for public markets. That part of the story makes sense. What's less examined is who's lining up behind them.

According to reporting from TechCrunch, startups are explicitly trying to "ride that SpaceX IPO wave" — timing their own listings to benefit from the market enthusiasm generated by high-profile AI and tech debuts. That's a strategy, not a business plan. And it's worth separating the two.

What a real AI IPO looks like

The companies with the strongest cases for going public right now share a few traits: they sell infrastructure or tooling that other AI builders depend on, they have recurring revenue that doesn't evaporate when a model generation turns over, and their growth isn't purely a function of the hype cycle.

CorWeave, which rents GPU (graphics processing unit) clusters to AI developers, fits that profile. So does any company sitting on contracted capacity from hyperscalers or enterprise customers with multi-year commitments. These businesses have something to show a public market investor beyond a pitch deck and a ChatGPT integration.

The second tier is the interesting problem

The more complicated question is what happens to the companies that don't fit that description but are going public anyway — or trying to.

Wave-riding is a time-honored Silicon Valley tradition. It worked for a lot of SaaS companies that went public during the 2020-2021 window. It also produced a graveyard of SPACs (special purpose acquisition companies — blank-check vehicles used to take companies public without a traditional IPO process) that are now trading at fractions of their debut prices.

The difference between those outcomes usually came down to one thing: whether the underlying unit economics — the per-customer revenue and cost structure — could survive a sentiment shift. Many couldn't.

Fundraising theater, now with a ticker symbol

There's a version of the current AI IPO moment that is essentially fundraising theater at scale. A company raises a large private round at a valuation that implies future dominance, uses that round to generate press, uses the press to generate customer interest, and then files for an IPO before the market has time to ask hard questions about churn or margins.

That cycle isn't new. But the AI context gives it extra cover, because the sector's genuine importance makes skepticism feel contrarian. It isn't. Asking whether a company's revenue is durable is not the same as doubting that AI matters.

What to watch

The IPO filings themselves will be the tell. S-1 documents (the registration statements companies file with the SEC before going public) are required to disclose revenue, customer concentration, and risk factors in detail that press releases are not. The gap between a company's fundraising narrative and its S-1 disclosures is often where the actual story lives.

For now, the wave is real. Whether everyone riding it belongs on the board is a separate question — and the more important one.