Meta plans to expand Ohio’s Prometheus cluster to about 1 million GPUs
The first phase of the gigawatt-scale supercluster in New Albany is complete and was used to train Meta’s Muse AI models. The system already operates hundreds of thousands of Nvidia GPUs across multiple buildings, according to The Next Web. Meta published an engineering account explaining the infrastructure and its operation.
- RuntimeWire reports Meta achieved 85% goodput while training Muse Wonder.
- According to WinBuzzer, spare processors and saved training states help recover from hardware failures.
Covered by 9 publishers within 26 hours of the first report.
Reporting9
Unite.AI Meta Details Prometheus Supercluster Scaling Toward 1 Million GPUs · Theo Nash, AI Infrastructure & Compute, AI Research Agent
RuntimeWire Meta says Prometheus delivered 85% goodput while training Muse Wonder · Ryan Merket
WinBuzzer Meta’s Prometheus AI Supercluster is Already Humming · Markus Kasanmascheff
Ars Technica Meta Muse exposes users' personal and business secrets when they share conversations · Dan Goodin
Quartz Researchers found 6,800 shared Meta Muse conversation links through public DNS records · Cris Tolomia
The Verge Meta pauses Muse sharing after researchers find AI chats on the web · Emma Roth
Engadget Meta suspends Muse chats sharing feature after it left private details exposed · staff@engadget.com (Anna Washenko)
The Next Web Meta details Prometheus, the Ohio supercluster heading for 1 million GPUs · Ana Maria Constantin
HPCwire Inside Prometheus, the Gigawatt Supercluster That Trained Meta’s Muse · Alex Woodie
Discussion
Hacker News 9 points, 1 comments