• Elon Musk declares electrical power, not chips, as the primary bottleneck for AI scaling, proposing solar-powered orbital data centers as the solution.
  • Through a merger with xAI, SpaceX aims to deploy up to one million solar-powered satellites, targeting cost parity with Earth-based computing within 2-3 years.
  • The move addresses a looming energy crisis, with AI data centers projected to double U.S. electricity consumption, potentially reshaping cloud infrastructure and competitive dynamics.

In a bold strategic pivot, Elon Musk has positioned SpaceX at the forefront of solving what he calls the "fundamental limiting factor" for artificial intelligence: energy. Speaking at a recent industry event, Musk emphasized that while chip production is ramping up, the real constraint is electricity, warning that "we're very soon—maybe even later this year—we'll be producing more chips than we can turn on." This insight comes as data centers operated by tech giants like Google (GOOGL) and Amazon (AMZN) consume power equivalent to small cities, with projections suggesting AI could drive U.S. electricity demand to double current levels.

The solution, according to Musk, lies beyond Earth's atmosphere. In early February 2026, he announced a massive merger of SpaceX with his AI startup xAI, creating a combined entity valued at $1.25 trillion. The goal is ambitious: deploy up to one million solar-powered satellites that function as orbital data centers, leveraging constant sunlight and passive cooling in space's vacuum. "Space-based AI is obviously the only way to scale," Musk declared, citing advantages like uninterrupted solar energy and reduced cooling costs. People familiar with the matter say internal timelines aim for these platforms to become the lowest-cost computing location within two years, though regulatory approvals and technical hurdles remain.

On the ground, the energy crunch is intensifying. Data center electricity consumption is expected to surge by 15% over the next four years, potentially accounting for 12% of total U.S. power generation by 2030. Musk noted that the United States currently uses about 0.5 terawatts, but AI needs could push that to 1 terawatt, straining grid infrastructure and slowing expansion. In response, SpaceX and Tesla (TSLA) are independently working to build up to 100 gigawatts per year of solar manufacturing capacity domestically, a process estimated to take three years. Meanwhile, efforts to secure partnerships for terrestrial power have hit snags, with some utilities struggling to meet demand.

Industry analysts are cautiously optimistic but highlight challenges. One source close to the discussions pointed out that developing reliable orbital infrastructure at scale requires overcoming heat dissipation and data transmission issues, while regulatory bodies have yet to greenlight such massive satellite constellations. Musk, however, remains bullish, predicting that "we might have AI that is smarter than any human by the end of this year" and claiming AGI will arrive in 2026. As negotiations for initial deployments continue, the focus is on whether SpaceX can achieve its aggressive cost reduction timeline, potentially reshaping how the tech sector operates at scale. Attempts to reach xAI for further comment were unsuccessful at press time.

Correction: An earlier version misstated the projected timeline for cost parity; it is within 2-3 years, not immediately.