TL;DR
The FCC has formally accepted SpaceX's filing for 100,000 "Starmind" AI computing satellites, opening the plan to public review and marking the first regulatory milestone for what could become a 1-million-satellite orbital supercomputer.
What happened
- On September 18, 2026, the FCC's Licensing Division accepted SpaceX's Starmind application as an official filing, triggering a public comment period.
- The 100,000-satellite tranche is the opening phase of a plan that scales to 1 million satellites.
- Each Starmind satellite weighs up to 4,000 kg, spans 75 metres, draws 250 kW of power, and carries a SpaceX-designed Nvidia Vera Rubin NVL72 computer.
- Elon Musk stated on September 20 that connectivity per satellite will reach 10 Tb/s in both directions, with a path to 100-plus Tb/s.
- Planned orbits sit at two very low Earth bands: 323 to 327.5 km and 473 to 477.5 km altitude.
Why it matters
- Acceptance is not approval: the filing clears a paperwork threshold only, and the FCC explicitly notes each application's fate may depend on earlier queue items, meaning regulatory risk remains high.
- A 250 kW power draw per satellite at 100,000 units implies 25 gigawatts of orbital power demand, a scale that has no precedent and will draw intense scrutiny from astronomers, spectrum regulators, and rival operators.
- The Nvidia Vera Rubin NVL72 pairing signals that SpaceX is positioning Starmind as distributed orbital AI infrastructure, not just connectivity, competing with terrestrial hyperscale data centers on latency-sensitive compute.
- Starlink Version 3 satellites (also covered in the application) are rated at 1 Tb/s downlink and six 400 Gb/s space lasers, and Starship launches add roughly 20 times the capacity of Falcon 9 V2 missions, compressing the deployment timeline dramatically.
- If even a fraction of the constellation launches, it would reshape the orbital compute market and pressure cloud providers, satellite broadband rivals, and national space agencies simultaneously.
What to watch next
- Public comment responses from competitors (Amazon Kuiper, OneWeb, Telesat), astronomers, and spectrum bodies will signal how much regulatory friction SpaceX faces before any grant.
- The September 28 Starship launch is the near-term hardware proof point: payload capacity and cadence will determine whether the deployment math behind 100,000 satellites is credible on any near-term schedule.
- Watch for Nvidia's formal acknowledgment of the Vera Rubin NVL72 orbital design win, which would confirm supply chain commitment and give investors a clearer picture of the compute architecture.
Originally published on Present of AI, a daily source-linked AI news timeline. Read the full timeline or browse the open dataset.