NVIDIA's DSX POD: a 1GW supercomputer costs $50B but yields $40B a year
LinusEkenstam · x · 2026-10-21
Linus Ekenstam clarifies that one shipping-container build doesn't produce 1GW of compute — his point is the economics driving these deployments. A 1GW supercomputer costs roughly $50B installed, generates about $40B in revenue within 12 months, and lasts 10+ years. Containerized delivery means being online in days instead of months: at these rates, 3 months without token generation is a $10B opportunity cost. The original post noted NVIDIA's DSX POD packs 17km of wiring into a single container.
Related event: NVIDIA Unveils Vera Rubin DSX POD: AI Factory in a Shipping Container(8 posts)→
More from Infra
- Mistral Partners with NVIDIA to Build AI-Native Engineering Workflows — Mistral AI · 2026-10-21
- Firmus, a neocloud set to list at $44B, started as a Bitcoin miner — victor_explore · 2026-10-21
- TokenGate: learnable token pruning cuts LLM prefill TTFT by up to 11.3x — pmttyji · 2026-10-21
- How OpenAI ingests petabytes of logs every day: a look at AI observability — blaizedsouza · 2026-10-21
- Lambda Featured in NVIDIA GTC Berlin Keynote on DSX MaxLPS and Confidential Bare Metal — TheZachMueller · 2026-10-21
- B300 cluster earns $38/GPU-hour; DeepSeek open-sources code but keeps the crown jewels — vincent_koc · 2026-10-21