Jensen Huang: doubling model speed doubles AI lab revenue while demand outruns capacity
rohanpaul_ai · x · 2026-10-29
Jensen Huang claims doubling model speed (e.g. via software or kernel optimization) doubles an AI lab's revenue — but only while demand outruns capacity.
- Inference revenue = tokens served × price per token
- Faster kernels raise tokens per second on the same GPUs, power, and building
- If every extra token finds a buyer at today's price, a 2x speedup doubles revenue with zero new capex
More from Venture
- Celesto AI lets you run GitHub Actions on its cloud computers with per-job pricing — aniketmaurya · 2026-10-29
- In SF, if your project isn't a $10B pitch, it's dismissed as a 'lifestyle business' — carlosdponx · 2026-10-29
- Google and Microsoft market caps momentarily identical at $3.94T each — JOBhakdi · 2026-10-29
- Databricks and Snowflake CEOs trade barbs on X over AI tokens and growth numbers — Yuchenj_UW · 2026-10-29
- Oracle's $664B backlog is real, but free cash flow is negative $23.7B — pdamodaran · 2026-10-29
- Bun creator defends VC-funded open source: still MIT-licensed and shipping — ycombinator · 2026-10-29