SemiAnalysis: renting US compute to China yields Nvidia ~$7M profit per megawatt
xiaosun86 · x · 2026-11-16
SemiAnalysis makes a devil's-advocate case for letting China rent American compute, framed as an economics argument:
- Nvidia makes about $7M profit per deployed megawatt, and cloud providers make another $7–10M at today's pricing — meaning China renting compute pays huge margin dollars into the US ecosystem
- If ByteDance wants to compete with Meta on global video recommendation, being allowed to rent (but not buy) chips could leave Meta with a fourth of the compute cost
- Application-layer companies built on open models add up to well over $10B in ARR; DeepSeek makes less than Baseten and Fireworks do serving its models — "America is making so much money off Chinese models"
The thread is a contrarian take on US compute export policy: rather than blocking China, lock it into paying for US compute and models.
More from Infra
- WSJ ran 66 Nvidia/Jensen headlines in 4 days, joke 'benchmark' shows — xiaosun86 · 2026-11-16
- Inkling too small to saturate Tensor Cores; BBBK speedups scale with model size — andrew_n_carr · 2026-11-16
- Reuters: Nvidia chips pile up in Hong Kong as China withholds import approval — _weiping · 2026-11-16
- Tongxin Intelligent raises hundreds of millions RMB to build AGI-native inference chips — 量子位 · 2026-11-16
- CMU's Singularity-Bench: Opus 5.5 builds AI infra beating SGLang, scores 1.25 vs human baseline 1.0 — 新智元 · 2026-11-16
- GLM-5.3-Flash on two Mac Studios hits ~140 tokens/s via tensor parallelism — cryotic · 2026-11-16