FULL STORY
Sequoia's Maguire Runs the Numbers on SpaceX Compute
Sequoia partner Shaun Maguire built a financial model of SpaceX's compute ambitions, calling the numbers "crazy" while arguing SpaceX can build data centers 2-3 years faster than rivals, sparking community debate over feasibility.
2026-10-31 ~ 2026-11-02 · 2 episodes · 4 posts
Episode 1 · Sequoia's SpaceX compute model questioned (2026-10-31, 2 posts)
Sequoia's Shaun Maguire built a bottom-up SpaceX financial model to 2028, but skeptics note the AI compute market is only $224B/year, and SPCX would need to capture 85% of market growth just to break even.
- Sequoia's SpaceX model implies massive 2028 compute demand—but does that much AI demand even exist? — aronchick · 2026-10-31
- AI compute market is ~$224B/year — SPCX would need to capture 85% of all growth to pencil out — aronchick · 2026-10-31
Episode 2 · SpaceX Could Build Data Centers 2-3 Years Faster Than Rivals (2026-11-02, 2 posts)
Sequoia partner Shaun Maguire argues SpaceX is a low-cost compute producer that can build data centers 2-3 years faster than competitors; analysts note that with $50B per GW, build speed—not capex—is the key variable for capital efficiency.
- AI datacenters priced at ~$50B per gigawatt, but build speed is the metric that matters — JOBhakdi · 2026-11-02
- Shaun Maguire: SpaceX can spin up datacenters 2-3 years faster than rivals, rev/GW to fall to ~$10B — shaunmmaguire · 2026-11-02