SpaceXAI plans massive GPU buildout to 1.44 million chips this year
The expansion accelerates SpaceXAI’s compute scale while it reshapes power use and lays groundwork for orbital data centers.
Elon Musk says SpaceXAI is preparing a major expansion at its Colossus sites, with 660,000 more Nvidia GPUs planned for this year. The deployment schedule calls for 220,000 GB300s by next week, another 220,000 in November and a final batch of 220,000 in late December if the company is fortunate. The additions would bring the operating fleet to 1.44 million GPUs, including 1.1 million GB300s; SpaceXAI says it will remove its gas turbines over a year as a 1.2-GW power plant comes online. Musk also says the company aims to increase data-center capacity sevenfold by 2027 and reach 50 million H100-equivalent GPUs by 2030. SpaceXAI is also planning an Orbital Data Center System involving one million satellites.
Why it matters
SpaceXAI’s plan to lift its GPU fleet to 1,440,000 units, including 1.1 million GB300s, signals an aggressive push to secure AI compute at scale, far beyond typical data-center growth. The shift from gas turbines to a 1.2‑GW power plant changes how that capacity is powered, while the aim to grow data-center capacity sevenfold by 2027 points to much larger future workloads. Longer term, a proposed orbital system of one million satellites suggests a move to distribute compute infrastructure well beyond terrestrial facilities.
Signal or noise?
Does this story matter, or is it hype? Decide before you see what everyone else thinks.
Sources
- Tom's Hardware