Scale

Meta has detailed Prometheus, its gigawatt-scale supercluster in Ohio, which will house approximately 1 million GPUs at full scale. The supercluster was used to train Meta's Muse foundational models.

Infrastructure • Techmeme

CERN and Mistral have signed a statement of intent to establish a long-term strategic partnership focused on developing foundation models for scientific discovery. The collaboration aims to combine CERN's expertise in fundamental physics with Mistral's large-scale AI models.

Research • Hacker News

A new report from MIT Technology Review examines how enterprise AI economics are shifting as usage scales. The analysis suggests the question for companies is moving from 'can we use it?' to 'how efficiently are we using it?' as AI becomes core infrastructure.

Industry • MIT Technology Review

Tenstorrent • Unknown

Tenstorrent, a Nvidia challenger, raised $1.5 billion, reaching a $6.2 billion valuation to scale open-weight AI chips and remain independent after drawing acquisition interest earlier this year.

Nscale • Unknown

Nscale and Anthropic are reportedly in talks for a $45 billion compute deal, which would be the latest in a long string of giant compute deals Anthropic has struck over the past year.

DeepSeek released DeepInfer, the inference framework behind its large-scale model services, as open source. The framework supports Nvidia GPUs and Huawei Ascend processors and includes capabilities like distributed inference and speculative decoding.

Infrastructure • TechNode