This isn’t a new story. It’s an ongoing one that hasn’t received the analytical treatment it warrants.
SpaceX and xAI agreed to merge on January 31, 2026, in an all-stock transaction. The Wall Street Journal confirmed the combined entity’s valuation at $1.25 trillion, describing it as the largest U.S. corporate tie-up by value. An internal employee memo followed on February 2. CNBC independently confirmed both the valuation and the orbital data center rationale.
The stated logic of the deal is worth taking seriously. Musk’s public framing centered on combining SpaceX’s Starlink global network and launch infrastructure with xAI’s large language model capabilities. The deal announcement referenced plans to develop what Musk described as orbital data centers, reportedly referred to in one source as “Grok-Sats”, though that specific term comes from a single report and shouldn’t be treated as an established product name. What WSJ and CNBC confirm is the orbital data center concept and the Starlink integration rationale. The technical architecture remains to be established.
Why does this matter for the AI infrastructure story? The hyperscaler model for AI compute is terrestrial: massive data centers, power-constrained locations, fiber networking. An orbital compute layer, if technically and operationally viable, would sit outside those constraints. Latency, power sourcing, and physical geography become different problems. Whether Musk’s orbital data center concept is technically achievable on a useful timeline is genuinely uncertain; the merger creates the organizational conditions to attempt it, not a demonstrated capability.
The deal’s scale deserves context without inflation. At $1.25 trillion, the combined entity is valued higher than any prior U.S. corporate merger, the WSJ’s characterization. For comparison, the merger alone is valued larger than the entire reported Q1 2026 global M&A figures circulating in some sources, which is part of why those broader figures couldn’t be independently verified and aren’t cited here. The SpaceX-xAI deal is an outlier by any historical measure. It doesn’t need Q1 context to be significant.
What to watch: the all-stock structure means no immediate cash transaction, but it creates a combined entity with the capital markets access to fund orbital infrastructure at scale. Watch for any SEC or FTC review disclosures, for xAI’s model development roadmap as a combined entity, and for Starlink subscriber data as a proxy for the network capacity that would underpin any orbital compute strategy. The February employee memo was the last official internal communication confirmed by cross-reference sources, any further internal announcements would be the next meaningful data point.
The infrastructure theme running through this hub’s recent coverage, CoreWeave, Oracle, Alaska data centers, today’s hyperscaler performance data, has a logical extension in the SpaceX-xAI deal. All of it points toward the same question: who controls the compute substrate that AI runs on? Musk’s answer, on January 31, was to build a new layer of it from orbit.