The AI IPO Gold Rush Is Here, and I've Seen This Movie Before

What we do know is that the AI model companies are in an arms race that requires staggering amounts of capital. Training runs cost hundreds of millions of dollars. Inference at scale isn't cheap either. Going public is one way to fund that race, though it comes with its own set of headaches (quarterly earnings calls, activist investors, the whole circus).

Alphabet's $80 billion ask

Speaking of staggering amounts of capital, Alphabet wants to raise $80 billion in equity to fund AI infrastructure expansion. Eighty billion! That's not a typo.

For context, that's roughly what Apple spends on R&D over four years. It's more than the entire market cap of most companies in the S&P 500. And Alphabet is raising it specifically for AI infrastructure, which means data centers, chips, cooling systems, and all the physical stuff that makes the AI magic happen.

This is where I start to get, let's call it, concerned. Not because Alphabet is wrong to invest in AI infrastructure. They're probably right that this is a generational shift in computing. But because the scale of capital being deployed is creating a situation where only a handful of companies can even play the game.

If you're a startup trying to compete in foundation models, you're looking at incumbents who can casually raise $80 billion for infrastructure. That's not a level playing field. That's not even the same sport. The young founders I talk to are increasingly pivoting to application-layer companies because the infrastructure layer is just too expensive to compete in.

The SpaceX situation

And then there's SpaceX, which is apparently pitching investors on a $28.5 trillion market opportunity. According to Bloomberg, the company is combining rockets and AI in its pitch, which is an interesting way to frame things.

Look, SpaceX has done genuinely impressive work. They've made reusable rockets a reality. Starlink is a real business with real revenue. But $28.5 trillion? That number is so large it's basically meaningless. It's the kind of figure you throw out when you want to sound visionary but don't want to be held accountable to anything specific.

The other interesting tidbit is that SpaceX is negotiating to pay razor-thin fees to the Wall Street firms handling its IPO. Which, good for them I guess? When you're Elon Musk and you're pitching a $28.5 trillion opportunity, you have leverage. The banks want the prestige of the deal even if the fees are minimal.

But what do I know. Maybe SpaceX really will capture a meaningful slice of a $28.5 trillion market. Maybe Starlink becomes the backbone of global internet infrastructure. Maybe the AI angle is more than marketing. I'm skeptical, but I've been wrong before.

The pattern I'm worried about

Here's what bothers me about all of this. We're in a moment where the AI hype is so intense that companies can raise essentially unlimited capital on the promise of future returns. Anthropic can go public without anyone knowing if they're profitable. Alphabet can ask for $80 billion and investors will probably give it to them. SpaceX can pitch a market opportunity larger than most national economies.

This is the self-driving car hype cycle all over again, except the numbers are bigger and the timeline is compressed. Back in 2016, every major automaker and tech company was convinced they'd have fully autonomous vehicles by 2020. They raised billions. They made promises. And then reality set in, the timeline stretched to "eventually," and a lot of that capital got written off.

I'm not saying AI is a bubble that's going to pop. The technology is real, the applications are real, and some of these companies will probably generate enormous returns. But the current moment, where any AI-adjacent company can raise unlimited capital at sky-high valuations, that moment won't last forever. It never does.

So what

If you're an investor, I'd be asking hard questions about unit economics and path to profitability. If you're a founder, I'd be thinking about how to build a sustainable business rather than chasing the biggest possible fundraise. And if you're just someone trying to make sense of the AI landscape, I'd suggest taking the big numbers with a grain of salt.

Anthropic filing before OpenAI is genuinely interesting. Alphabet's infrastructure bet is probably smart, even if the scale is wild. SpaceX is SpaceX, which means equal parts impressive engineering and Musk-style hyperbole.

But $28.5 trillion markets and $80 billion raises? Call me old-fashioned, but I'll believe it when I see the actual returns. My email's on the about page if you want to argue.

James Chen · 4 hours ago · 4 min