On June 12, 2026, SpaceX is set to debut on the Nasdaq at a valuation of roughly $1.77 trillion — the single largest IPO in the history of the world, by a wide margin. The roadshow kicked off June 4, pricing lands after the close on June 11, and at $135 a share Elon Musk's company will instantly become the seventh-most-valuable company in America, worth more than Tesla.
Here's the thing almost every headline gets wrong, though, and it's the reason I wanted to write this for you specifically. Everyone is calling this a "rocket IPO." It isn't. Underneath the Starship launches and the Mars talk, SpaceX has quietly become one of the largest artificial-intelligence companies on Earth — and if you don't see that, you'll completely misjudge what you're actually buying. Let me show you the AI story hiding inside the hype, because it changes everything about how you should think about this deal.
The numbers behind the biggest IPO ever
Let's start with the raw scale, because it's genuinely historic and worth sitting with for a second.
| Metric | SpaceX IPO |
|---|---|
| IPO date | June 12, 2026 (Nasdaq) |
| Share price | $135 |
| Valuation | ~$1.77 trillion |
| Amount raised | ~$75 billion (largest ever) |
| Shares offered | ~556.6 million |
| Rank in US by market cap | ~7th, above Tesla (~$1.6T) |
| Starlink 2025 revenue | $11.4 billion (61% of total) |
| Starlink share of revenue, Q1 2026 | 69% |
A $75 billion raise is roughly three times the size of the previous record IPO. To put $1.77 trillion in perspective: that's larger than the entire annual GDP of countries like Spain or Australia, attached to a single company that didn't even exist 25 years ago. This is the kind of event that defines a market cycle — and I covered how it fits into the broader wave in the $3 trillion AI IPO race, where SpaceX, Anthropic, and OpenAI all rush public at once.
The part the headlines miss: this is an AI company wearing a rocket costume
Now the piece I really want you to understand, because it's where the value — and the risk — actually lives.
In February 2026, Musk merged SpaceX with xAI, his artificial-intelligence company, in an all-stock deal that valued the combined entity at $1.25 trillion. By May, xAI was dissolved as a standalone company entirely — Grok (the AI model) and X (the social platform) now operate as SpaceX's internal AI division. Read that again. The company going public next week isn't just rockets and satellites. It owns a frontier AI lab, a major social network, and one of the most-used chatbots in the world.
And this isn't a side project. Look at where the money actually went last year: of SpaceX's $20.7 billion in total capital expenditure, 61% — about $12.7 billion — went to AI infrastructure. More than half of everything SpaceX spent building for the future went toward AI, not rockets. The AI segment (X advertising, subscriptions, and xAI compute) already adds roughly 17% to revenue. When a "rocket company" spends the majority of its capital budget on AI, the label on the box is wrong. You're buying an AI conglomerate that happens to own the world's best launch business.
The genuinely wild part: data centers in space
Here's where it stops being a clever financial maneuver and becomes a real strategic bet — the kind only Musk would attempt.
The single biggest constraint on AI right now isn't ideas or talent. It's power and compute capacity. I've written about how desperate that crunch has become — it's the entire reason behind the $67 billion NextEra-Dominion utility merger, where the bottleneck on AI turned out to be electricity itself. Every hyperscaler on Earth is fighting over gigawatts.
SpaceX's answer is audacious: build the data centers in orbit. The plan is constellations of satellites packed with AI compute — targeting up to 100 gigawatts of AI compute capacity annually — powered by unlimited solar energy in space, cooled by the vacuum, and launched cheaply by Starship. Starlink provides the low-latency global connectivity to reach them. Deployment could begin as early as 2028.
Think about why this is so clever. On Earth, AI data centers are choking on power limits, cooling costs, water use, and angry neighbors. In orbit, the sun never sets, cooling is free, and there are no ratepayers to revolt — the exact problems I described in how ordinary people are footing the AI power bill. SpaceX is the only company on the planet that owns all three pieces needed to pull this off: the rockets to get there cheaply, the satellite network to connect it, and now the AI lab to fill it. That vertical integration is the real thesis of the IPO.
Follow the money: how Starlink secretly funds the AI dream
Here's the financial engine that makes the whole thing work, and it's beautiful in its simplicity once you see it.
Starlink is a cash machine. It threw off $11.4 billion in revenue in 2025, and its share of SpaceX's total revenue climbed from 61% to 69% in a single quarter. That profitable, fast-growing satellite-internet business generates the cash that funds the wildly expensive, not-yet-profitable AI ambitions — the orbital compute, the Grok training runs, the xAI talent.
It's the same playbook the AI giants use, just with rockets attached. Google funds DeepMind with search-ad money. Microsoft funds its AI push with Office and Azure. SpaceX funds its AI conglomerate with Starlink subscriptions. The pattern across the whole industry is identical: a boring, profitable core bankrolls an expensive, world-changing bet. Understanding that structure is exactly how you judge whether a sky-high valuation is sane, which is the lens I laid out in how VCs actually price AI companies in 2026.
What you should actually watch for (the risks nobody's shouting about)
I'd be doing you a disservice if I only sold you the dream, so here's the honest other side. A $1.77 trillion valuation prices in a lot of future that hasn't happened yet.
It loses money where it counts. SpaceX's IPO filing shows billions in losses, and the AI division in particular is a cash furnace. Starlink is profitable; the rest is a bet. You're paying a trillion-plus dollars partly for things — orbital data centers, Mars, mass-market Grok revenue — that don't exist yet.
Musk owns and controls it. Through super-voting shares, Musk keeps tight control. That's a feature if you trust his vision and a serious risk if you don't — public shareholders get the ride, not the steering wheel.
The AI bet might not pay off on schedule. Orbital data centers are a genuinely unproven idea targeted for 2028+. If the AI compute crunch eases on the ground before then — or if the economics of launching compute to orbit don't close — the most exciting part of the story slips. And the whole thing rides on AI demand staying red-hot, the same demand propping up Anthropic's $965B valuation and NVIDIA's $78 billion quarters. If that cools, every trillion-dollar AI bet reprices at once.
You may be late. This is the uncomfortable one. As I explained in how the AI wealth boom locked ordinary people out, most of the gains in companies like this were captured privately, years ago, by insiders. By the time a $1.77 trillion company IPOs, a lot of the easy upside has already happened. The IPO is, in part, early investors finding new buyers — and that new buyer might be you.
What this means for you
So how should you actually think about it, depending on who you are?
If you're tempted to buy on day one, slow down. Hot mega-IPOs frequently pop on opening day and then sag once the excitement fades and lockups expire. Decide whether you're buying the business for the next decade or chasing the headline for the next week — they lead to very different choices. If it's the long-term AI-plus-space thesis you believe in, there's rarely a need to win the first hour of trading.
If you want the AI exposure without the single-stock risk, remember you don't have to bet everything on Musk's execution. A diversified basket gets you exposure to the whole AI-and-robotics wave — SpaceX-adjacent names included — without your outcome hinging on one founder and one unproven orbital-compute bet. I walked through that approach in how to invest in the AI boom without picking stocks.
If you just want to understand the moment, here's the takeaway to hold onto: the biggest IPO in human history is, at its core, a bet that AI demand is so insatiable we'll literally launch the data centers into space to feed it. Whether or not you buy a single share, that tells you something profound about where this whole era is heading — the same trillion-dollar trajectory I traced in the AI race that's reshaping everything.
The honest take
Calling SpaceX a "rocket IPO" is like calling Amazon a "bookstore IPO" in 1997 — technically true, and completely missing the point. The rockets are real and the Starlink cash is real, but the reason this thing is valued at $1.77 trillion is the bet stacked on top: that SpaceX can use its launch monopoly to win the AI-infrastructure race by going somewhere no competitor can follow — orbit.
It might work spectacularly. It might be a decade early. What I'm confident about is that you should walk into this with your eyes open about what it actually is: not a bet on rockets, but the most ambitious, most vertically integrated AI bet anyone has ever taken public. Price it as the AI conglomerate it is, not the rocket company the headlines call it, and you'll make a far smarter decision than the crowd chasing the launch-day pop.
So before June 12, ask yourself the real question — not "do I want to own a piece of SpaceX?" but "do I believe AI demand gets so big that data centers in space become inevitable?" Answer that honestly, and you'll know exactly what to do.
