We believe we have identified the largest actionable total addressable market in human history.

— SpaceX prospectus

SpaceX, hence the name, was a space company. “Was,” past tense.

They still launch rockets and operate satellites, of course, and they still talk a lot about things like lunar infrastructure, satellite-based data centers and colonies on Mars. But when the company absorbed xAI in February (nearly a year after xAI absorbed X), valuing the combined enterprise at $1.25 trillion, the message was clear enough: Elon Musk knows a speculative mania when he sees it, and he wasn’t about to let the opportunity go to waste.

Now, SpaceX looks more like an AI company than a starship enterprise, just like Tesla looks, on some days, more like a(nother) holding company for Musk’s sundry hobbies and side projects than it does an EV manufacturer.

The figure below gives you a sense both of how astronomical SpaceX’s “projections” for the AI market are, and also how relatively negligible space itself is in the grand cosmic scheme of SpaceX’s investor pitch.

So, the total addressable market for AI, inclusive of enterprise, infrastructure, subscriptions and ads, is damn near $27 trillion. Allegedly. The cosmos, meanwhile, is a $400 billion market. I dare say the universe never looked so small.

In the preliminary prospectus quoted here at the outset, SpaceX focused on plans to tap into “consumer and enterprise applications” by way of what the company described as “leading frontier models.” I imagine OpenAI and Anthropic might quibble with the term “leading,” and SpaceX will have to compete with both for investors’ affection later this year, when the two go public.

Just to be clear: We’re to believe (and, if we buy the stock, we’re to trust) that last year’s rocket and satellite company is now, or will be soon, the leading name in AI.

It’s not as far-fetched as it seems. Musk was, let’s not forget, one of OpenAI’s co-founders and xAI’s a real (loss-making) business with real models and real contracts to provide access to GPU capacity at real data centers. Moreover, Musk is Musk. Betting against him’s a fool’s errand.

That said, there’s something inescapably — and I’ll choose my words carefully here, lest I inadvertently break my own rule by making a rhetorical bet against Musk — unconventional about the idea that history’s largest IPO is a company whose business model underwent an overnight overhaul a mere four months ago.

Musk would say that’s a ridiculous characterization, and he wouldn’t be wrong. At least not in spirit. Conceptually, there’s not a lot of utility in drawing a distinction between his businesses. There’s a very real sense in which xAI was always SpaceX and vice versa, just like Tesla’s Neuralink and Neuralink’s The Boring Company, and so on.

For years, Tesla’s traded like “liquid Elon,” which is to say shares of the EV maker were just a way to bet on the fortunes of Musk — the spaceman, the social media tycoon, the brain chip guy, the tunnel digger, the AI king and, briefly, the US shadow president.

But on paper, SpaceX was a separate entity. That entity launched rockets, sold satellite broadband and was worth less than $50 billion in the summer of 2020. Midway through last year, SpaceX was worth around 10 times that, at $450 billion. By December, the figure was $800 billion. Then, following the internal Musk merger which transformed the company into an AI power player (and a de facto data center REIT), $1.25 trillion.

Fast forward another four months and SpaceX was set to debut as the seventh- or eighth-largest company by market value on the planet, worth some $1.8 trillion after pricing 555.6 million shares at $135 each.

As the figure shows, SpaceX’s debut — which was four times oversubscribed — is three times larger than Aramco’s 2019 offering, and more than double the size of what Ant Group would’ve likely raised in November of 2020 had Jack Ma not pissed off Xi Jinping the previous month.

Skeptics abound. “The largest IPO in history is a perfect storm of everything that’s wrong with Wall Street and the clearest signal that we are in the midst of an equity market bubble,” JonesTrading’s Mike O’Rourke said. “Investors will be paying 95x sales for a cash-burning enterprise.”

SpaceX, O’Rourke went on, has been around for nearly a quarter of a century. “Apparently 2026 is the year it all comes together,” he quipped. “That’s more indicative of the times than SpaceX’s fundamental prospects.”

In the same note, O’Rourke called SpaceX (and this is hilarious if you get the joke), a “supersized version of CoreWeave.” Jim Chanos echoed that sentiment. “The real excitement should be about developing some new, powerful agentic model,” he said. “But xAI seems to be suddenly changing its business from developing models to basically becoming a neocloud.”

As for Musk’s personal fortune, he’ll be worth nearly a trillion on Friday, bringing his on-paper net worth gain since the onset of the pandemic to $935 billion, give or take.

As the figure shows, nobody’s close. And no one’s likely to be close ever again. Musk is a god among gods.

I, for one, won’t be buying SpaceX. It’ll be in the benchmarks soon enough (and, depending on the benchmark, sooner than critics believe is safe, fair and warranted), which means I’ll own it passively anyway.

Note that not betting on Musk with a single-stock purchase isn’t the same as betting against him. Personally, I wish SpaceX all the best in its overnight conversion to an AI powerhouse. Presumably, the more successful it is in that arena, the more it’ll be able to plow into actual space exploration, which we’re gonna need once we “succeed” in turning Earth into the dying terrarium depicted over the first hour of Christopher Nolan’s Interstellar.

I’ll give the last word to Chanos who, while speaking at a conference this week, said, “The total addressable market for space is infinite. You can build whatever stories you want to justify the valuation.”