# Nvidia cites production AI-factory grid-response result

_Published Tuesday, September 15, 2026 at 2:07 PM EDT · AI, Infrastructure · Latest · Tier 2 — Notable_

![Nvidia cites production AI-factory grid-response result — Primary](https://blogs.nvidia.com/wp-content/uploads/3026/09/end-to-end-tech-blog-dsx-ee-kv-1920x1080-5601273.jpg)

Nvidia said its Eos AI factory has responded to more than 200 demand signals from Silicon Valley Power through Emerald AI's Conductor platform. The system reduces or reschedules lower-priority computing work while preserving higher-priority services. Nvidia also reported that Lambda ran 19 nodes within the power budget normally assigned to 16, increasing token throughput from about 4 million to 5 million per second.

## Sources

- [NVIDIA Blog](https://blogs.nvidia.com/blog/from-megawatts-to-tokens-how-nvidia-maximizes-ai-factory-production/)

---
Canonical: https://techandbusiness.org/newswire/uxjFMUYRgI3SrJJktE2vAN
Published: 2026-09-15T18:07:27.124Z
Story chronology: 2026-09-15T16:55:59.000Z
Retrieved: 2026-09-15T20:01:23.339Z
Publisher: Tech & Business (techandbusiness.org)