SpaceX folds xAI into its business to pursue space-based AI data centers

SpaceX says it has acquired Elon Musk’s AI startup xAI, positioning the combined company around a single big bet: moving compute off Earth. For B2B teams building on LLMs, the deal is a reminder that infrastructure constraints—power, cooling, and capacity—are becoming strategic, not just technical.

Space-based data centers become the merger’s headline goal

In a memo published by SpaceX, Musk argued that progress in AI is increasingly gated by terrestrial data centers and their energy and cooling requirements, and that “global electricity demand for AI simply cannot be met with terrestrial solutions” (SpaceX memo). The proposed alternative is a network of space-based data centers that would require a “constant stream” of satellites, effectively tying xAI’s compute roadmap to SpaceX’s launch cadence.

That satellite dependency matters commercially: satellites must typically be de-orbited after about five years under US Federal Communications Commission rules, creating a built-in replacement cycle and recurring launch demand.

Financial pressure, IPO timing, and near-term risks

The transaction values the combined entity at 1.25 trillion dollars, according to Bloomberg. Bloomberg also reported xAI is burning around 1 billion dollars per month, putting pressure on monetization and funding.

SpaceX’s own cash engine is increasingly internal: as much as 80 percent of revenue is tied to launching Starlink satellites, per Reuters. Meanwhile, SpaceX has been reported as preparing an IPO as early as June; Musk did not address timing in the memo.

Operationally, the two companies face very different immediate goals: SpaceX is focused on proving Starship for lunar and Mars missions, while xAI is in a fast-moving model race. The Washington Post reported that pressure contributed to loosened safeguards on Grok, which enabled nonconsensual sexual imagery generation.

For marketers and commerce operators, the near-term takeaway is not “space compute tomorrow,” but that compute scarcity, governance, and safety issues are converging with corporate structure—and will shape AI pricing and reliability long before any orbital data center is online.