SpaceX Acquires xAI in $1.25 Trillion Mega-Merger

On February 2, 2026, Elon Musk announced the largest corporate merger in history. SpaceX acquired xAI in a deal valuing the combined entity at $1.25 trillion: SpaceX at $1 trillion, xAI at $250 billion. The goal is not just to combine two companies. It is to move AI compute off the planet.

The merger of SpaceX and xAI creates the most valuable private company in history, and a plan to put data centers in orbit

What Happened

Musk announced the acquisition in a blog post, explaining the core thesis: global demand for AI computing will soon outpace available electricity and cooling capacity on Earth. SpaceX has already applied to the FCC for permission to launch a constellation of one million satellites to provide a network of solar-powered orbital data centers.

Musk estimated that within two to three years, the “lowest cost way to generate AI compute will be in space.”

The Numbers

Metric	Value
Combined valuation	$1.25 trillion
SpaceX valuation	$1 trillion
xAI valuation	$250 billion
SpaceX 2025 revenue	$15-16 billion
SpaceX 2025 profit	~$8 billion
xAI monthly burn rate	~$1 billion
Planned satellite constellation	1 million satellites

The Musk Empire Stack

This merger is the culmination of years of vertical integration across Musk’s companies:

Musk's AI-Space Vertical Stack (2026)
├── xAI → AI models, Grok, training infrastructure
├── SpaceX → Launch vehicles, satellite constellation, orbital compute
├── Tesla → $2B investment in xAI, autonomous driving AI, robotics
├── X (formerly Twitter) → Data pipeline, distribution, Grok integration
├── Neuralink → Brain-computer interface, human-AI interaction
└── The Boring Company → Physical infrastructure, tunneling

The xAI acquisition is not Musk’s first consolidation move. In March 2025, xAI acquired X (formerly Twitter) in a $33 billion all-stock deal. Tesla invested $2 billion in xAI as part of a $20 billion Series E round. Each acquisition makes the others more valuable: X provides training data, Tesla provides edge AI and hardware sales, SpaceX provides compute infrastructure.

Why Space-Based AI Compute

The argument for orbital data centers rests on three constraints hitting terrestrial AI infrastructure simultaneously:

1. Power limitations. AI data centers consume enormous amounts of electricity. The largest facilities draw over 1 GW, equivalent to a small city. Available grid power near major population centers is increasingly contested between AI companies, EV charging, and residential use. Some providers are trying to solve this on Earth instead, as we cover in our look at flexible, grid-aware data centers; orbital compute is the more radical bet that the terrestrial grid can’t be fixed fast enough.

2. Cooling costs. Data centers in hot climates spend significant energy on cooling. Space offers near-unlimited radiative cooling capability.

3. Permitting delays. Building new data centers on Earth involves years of environmental review, zoning approvals, and utility negotiations. Launching satellites bypasses most of these regulatory bottlenecks.

The counterargument is equally straightforward: latency. Light-speed delays between Earth and low orbit add roughly 4-20 milliseconds of round-trip latency depending on altitude. For real-time inference, that matters. For training workloads (which account for most GPU hours) it may not.

The IPO Timeline

The Financial Times reported in December 2025 that Musk was preparing a SpaceX IPO for mid-June 2026. The xAI acquisition positions the combined entity as not just a launch company but an AI infrastructure company, potentially commanding a higher valuation multiple from public market investors.

The IPO would be one of the largest in history, dwarfing Alibaba’s $25 billion debut in 2014.

Culture Clash Concerns

Not everyone inside xAI is optimistic. Former xAI employee Benjamin De Kraker noted a fundamental cultural divide: “xAI prides itself on ‘move fast and break things,’ flat hierarchy,” while SpaceX, for obvious reasons involving rockets carrying humans, operates with rigorous engineering processes and strict chain of command.

This is a real risk. The history of large tech mergers is littered with examples where cultural incompatibility destroyed the acquired company’s velocity: Yahoo and Tumblr, Microsoft and Nokia, HP and Autonomy, and, in an earlier era, AOL and Time Warner. In each case, the acquiring company underestimated how much of the acquired company’s value was tied to a way of working that didn’t survive contact with a larger, more process-heavy parent. xAI’s flagship model, Grok, has shipped on a fast, iterative cadence; whether that pace holds up inside a company that also has to certify rockets for human spaceflight is an open question.

A merger of this size would also be expected to draw regulatory attention. SpaceX already holds significant government launch contracts, and a combined entity spanning launch vehicles, a satellite constellation, and frontier AI models raises the kind of market-concentration questions that typically trigger antitrust and national-security review, independent of whether xAI itself needed FCC approval for its satellite plans.

What This Means for the AI Industry

The SpaceX-xAI merger signals several things:

1. AI infrastructure is the bottleneck, not models. The fact that Musk is willing to spend billions on orbital compute infrastructure suggests that model capabilities have outrun the infrastructure needed to run them at scale.

2. Vertical integration is the strategy. The era of best-of-breed AI stacks, where you pick your model from one vendor, your compute from another, and your data from a third, may be giving way to vertically integrated AI empires.

3. The AI race is now a space race. If orbital compute proves viable, companies without launch capability will be at a structural disadvantage. This could reshape not just AI competition but international technology policy.

The Competitive Response

The merger puts pressure on every major AI company to secure long-term compute capacity. OpenAI is backed by Microsoft’s Azure infrastructure and, separately, a reported government equity stake. Google has its own data centers and TPU chips. Amazon has AWS. All three are also racing to lock down GPU supply, a dynamic we track in our NVIDIA earnings and GPU capacity coverage and our comparison of NVIDIA, AMD, and Intel AI chips. But none of them have rockets, and none of them can credibly threaten to move compute off the planet if terrestrial capacity runs out. For general background on how these facilities work today, see our explainer on AI infrastructure and data centers.

Whether orbital data centers are a genuine breakthrough or an expensive distraction remains to be seen. What is clear is that Musk is making the biggest bet in corporate history that the future of AI will not be confined to Earth.