Inside Prometheus: Meta’s Journey to a Gigawatt-Scale Supercluster - Engineering at Meta

kept by eddie oct 20

Meta's Prometheus supercluster achieves over 85% goodput on training runs and is scaling to ~1 million GPUs using NVIDIA GB200, GB300, and upcoming Vera Rubin platforms.

When AWS Grew Wings: The “Boring” Genius of Amazon Leo – SatNews
Amazon Leo's consumer broadband terminals serve as a volume play to reduce costs of its custom Prometheus chip, enabling a profitable B2B backhaul infrastructure play rather than competing directly with Starlink's consumer model.
also on Prometheus, #infrastructure, #hardware
Jeff Bezos Wants to Build an ‘Artificial General Engineer’
Jeff Bezos's startup Prometheus is building an artificial general engineer using AI to accelerate the design and manufacturing of devices like computers and jet engines.
also on Prometheus, #ai
Higher usage limits for Claude and a compute deal with SpaceX
Anthropic doubles Claude Code rate limits and API capacity for Opus models, enabled by a new compute partnership with SpaceX's Colossus 1 data center.
also on #data-centers, Nvidia, #infrastructure, #ai
Nvidia Unveils Faster AI Chips Sooner Than Expected - WSJ
At CES, Nvidia CEO Jensen Huang unveiled the next-generation Vera Rubin AI server systems, signaling faster chip development cycles driven by the AI race.
also on #data-centers, Nvidia, #hardware, #ai