SpaceX and Nvidia plan to launch Starmind AI satellite in Q4 2027
On October 16 Elon Musk posted on X praising Jensen Huang and announced progress on “Starmind,” a space‑based AI data center being built by SpaceX and Nvidia. He said the first‑generation satellite, called AI1, will carry a space‑adapted version of Nvidia’s next‑generation system Vera Rubin NVL72, and that mass production would start in 2027, with a launch targeted for the fourth quarter of 2027…
Key points
- SpaceX and Nvidia target a Q4 2027 launch of the AI1 satellite carrying a Vera Rubin NVL72 system.
- FCC filing requests up to 1 million units; review accepted Feb 2026, not yet approved.
The FCC filing lists a maximum of 1 million units, accepted for review in February 2026 but not yet approved. The ground version bundles 72 Rubin GPUs and 36 Vera CPUs. IEEE Spectrum estimates about 80 square meters of radiator space needed per rack. Google’s Project Suncatcher test satellite, about the size of a refrigerator, carries four TPUs and can run AI models for roughly 15 minutes before thermal limits stop it.
SpaceX shares closed at $148.62 on the 15th and rose to $151.99 in pre‑market trading. Analysts note that heat‑dissipation hardware may be the first commercial opportunity, while actual orbital run time and first customers remain unknown.
The story so far
3 episodes →- SpaceX and Nvidia plan to launch Starmind AI satellite in Q4 2027this story
SpaceX and Nvidia's Starmind, Launch in Q4 2027: Reading the Numbers on the 'Space AI Data Center'
note.com · 19 October 2026
Loading the full article…
This text was published by note.com and written by 宇宙ビジネスキュレーター. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Hardware & Compute
All →- NVIDIA’s Vera CPU delivers up to twice the throughput of x86 servers for Modal’s AI sandboxes · 7 src
- Axera's AX8860P claims to outpace Nvidia DGX Spark in select LLM benchmarks · 1 src
- Opinion: Google’s AI focus leaves Pixel missing Samsung’s Good Lock and widget stacking · 1 src
- Nvidia projected to sell about $430 billion in chips in 2026 as data-center revenue hits $89 billion · 6 src
- Microsoft could make Maia 300 its primary inference engine with 2.5x memory bandwidth · 2 src
Comments
via GitHub Discussions