The AI unit of SpaceX is rushing to fulfill extensive compute obligations by activating a massive fleet of hardware in the coming weeks. According to The Information, the company plans to bring roughly 420,000 Nvidia GPUs online in November to satisfy agreements worth billions of dollars, a target that exceeds previous estimates of 220,000 units for the month.
This acceleration is part of a broader rollout of the Colossus 1 and 2 clusters, which could reach a total of 1.44 million GPUs by year-end. Elon Musk's previously disclosed schedule includes 220,000 GB300 chips coming operational this week and another batch in December. While analysts project total server power will reach 2.7 GW by the end of the year, some critics argue currently operational capacity is only 1 GW. The investment for GPUs in the fourth quarter alone is estimated at over $50 billion.
Key sources
- SOURCE@theinformation“roughly 420,000 Nvidia GPUs online in November as it races to meet billions of dollars in new compute commitments”x.com
- SOURCE@elonmusk“Colossus 1 is 150k H100, 50k H200 and 30k GB200. Colossus 2 is 110k GB200 and 440k GB300”x.com
- SUPPORT@kakashiii111“while SpaceX claimed in its latest 10-Q to have deployed an additional 0.4GW, that capacity was not operational or online”x.com
- SUPPORT@sawyermerritt“possible total of 660,000 GB300s coming online in Q4 alone. Upwards of $50 billion of GPUs”x.com
- SUPPORT@shanumathew93“implies ~2.7 GW of server power (~3+ GW facility) by year-end”x.com
- SUPPORT@wintonark“If they deliver on these installs it would put them 20% ahead of our 2026 expectations”x.com
- SOURCEhuggingnewshuggingnews.com