Nvidia sets out three factors behind AI factory returns
What happened
On 1 October 2026, Nvidia published a blog post arguing that the return on an AI factory depends on three factors: output per megawatt, hardware service life and token demand. Using Vera Rubin NVL72 as an example, the post says its throughput per megawatt is more than 30 times higher than GB300 NVL72, and its cost per million tokens on DeepSeek V4 Pro is up to 45 times lower. This is currently the only report on the matter, with no follow-up or corroboration from other sources.
Written by AI from the reports · updated 1 h ago
Timeline
Follow the reports to see the event from each side.
- NVIDIA BlogProductive, Durable, Fungible: How NVIDIA AI Factories Maximize Return on Investment
AI factories are built by the megawatt, even by the gigawatt. Each megawatt factory costs roughly $60 million, and AI factory operators will only commit capital on that scale with a clear view of the return on investment.
Heat of this event
Not enough continuous observations yet to draw a trend.