The $1.25 Trillion Bet: SpaceX Swallows xAI and Plans Data Centers in Space

The Biggest Deal in History

When Elon Musk announced that SpaceX would acquire xAI in early February, the numbers alone were staggering. The combined entity is valued at $1.25 trillion, with SpaceX contributing $1 trillion and xAI clocking in at $250 billion. That makes it the largest merger of all time, period. Not the largest tech merger. Not the largest in the AI space. The largest, full stop.

The deal brings together SpaceX's rocket fleet, the Starlink satellite network, xAI's Grok chatbot, and the X social media platform under one corporate roof. Musk described it as building "the most ambitious, vertically integrated innovation engine on (and off) Earth." That's the kind of sentence that either ages brilliantly or becomes a punchline; there's rarely a middle ground with Musk.

Why Space Data Centers?

The strategic pitch is straightforward, if audacious. AI model training demands enormous computational power, which demands enormous amounts of electricity and cooling. Musk's argument is that this demand will eventually become unsustainable on Earth without "imposing hardship on communities and the environment." His proposed solution: put the data centers in orbit and power them with solar energy.

SpaceX has already requested regulatory approval in the United States to launch one million orbital data centers. Yes, one million. The idea is that space offers essentially unlimited solar power, natural cooling (space is rather cold), and no NIMBY neighbors to worry about. Whether the economics and engineering actually work out is a separate question entirely, but the ambition is unmistakable.

xAI is currently burning through roughly $1 billion per month, according to Bloomberg. That cash burn rate becomes a lot more palatable when SpaceX generates massive revenue from Starlink satellite launches. The merger essentially gives xAI a sugar daddy with actual cash flow.

The Talent Exodus Nobody's Ignoring

Here's where the story gets complicated. Six of xAI's original twelve co-founders have left the company. That's half the founding team, gone. Jimmy Ba and Tony Wu departed in early February, following earlier exits by Kyle Kosic, Igor Babuschkin, and Christian Szegedy. Fortune called it an "X-odus," and the headline stuck.

Musk responded by announcing a reorganization, structuring xAI into four core divisions: Grok (the chatbot and voice product), Coding, Imagine (video generation), and something called Macrohard, described as "an AI software company run by digital agents." Musk characterized the departures as the company "parting ways with some people" to improve speed of execution, suggesting at least some of the exits were not entirely voluntary.

The timing is awkward, to put it mildly. These departures come as xAI faces regulatory probes in multiple jurisdictions across Europe, Asia, and the United States over whether Grok enabled mass creation of non-consensual deepfake images. Losing half your founding brain trust while navigating regulatory minefields is not ideal.

The IPO Factor

None of this is happening in a vacuum. SpaceX is planning a blockbuster IPO, reportedly targeting late June 2026 with a valuation that could reach $1.5 trillion. That would make it one of the largest public offerings in history.

The xAI merger positions SpaceX as more than a rocket company heading into that IPO. It's now an AI company, a satellite internet company, a social media company, and a space infrastructure company all wrapped in one. The bull case writes itself for investment bankers.

But the co-founder departures could complicate the narrative. Public market investors tend to get nervous when key technical talent walks out the door right before an IPO. Musk will need to convince Wall Street that the reorganization strengthened xAI rather than hollowed it out.

What the Merger Means for AI

The bigger picture here is about compute infrastructure. The AI industry's appetite for computing power shows no signs of slowing down, and the first gigawatt-scale data center clusters are already coming online in 2026. Every major AI company is scrambling for power, cooling, and physical space.

Musk's orbital data center concept is the most extreme answer to this problem anyone has proposed. If it works, it could give the merged SpaceX-xAI entity a compute advantage that no Earth-bound competitor could match. If it doesn't work, well, that's a lot of expensive satellites floating around doing nothing.

The merger also raises questions about vertical integration in AI. When one company controls the rockets, the satellites, the AI models, the social media distribution platform, and potentially the orbital compute infrastructure, that's a level of integration that makes even the most ambitious tech conglomerates look modest.

What to Watch

The SpaceX IPO in late June will be the first real market test of whether investors buy the combined vision. Before that, watch for any additional co-founder departures, progress (or lack thereof) on the orbital data center regulatory approvals, and how xAI's Grok performs against competitors like Claude, GPT, and Gemini. The deepfake regulatory probes could also become a significant headwind if they result in fines or operational restrictions in major markets.

The $1.25 trillion bet is placed. Now comes the hard part: actually building it.