DeepSeek Might Need 100k GPUs to Hit ARR Target

teortaxesTex · x · 2026-07-15

Based on the current pricing and speed of V4, the author estimates that for DeepSeek to reach $8B ARR, it would need roughly 100,000 inference GPUs at peak capacity.

The post also provides rough calculations:

This post primarily discusses the relationship between inference compute demand and commercial scale, rather than just product opinions.

Related event: DeepSeek inference economics: sizing the GPU fleet behind its ARR(8 posts)→

Original post →

More from Infra

Infra channel →