Particle.news
Download on the App Store

OpenAI Taps Cerebras for 750MW of Compute Through 2028 in Multi-Year Pact

The pact adds dedicated low-latency inference capacity that fits OpenAI’s push to diversify compute suppliers.

Overview

  • OpenAI and Cerebras say the capacity will come online in multiple tranches through 2028.
  • Multiple outlets report the contract is worth more than $10 billion over its life, though the companies have not disclosed financial terms.
  • OpenAI frames the integration as a dedicated low-latency inference solution to deliver faster, more natural real-time responses.
  • Cerebras will provide cloud services powered by its wafer-scale chips, with facilities it builds or leases, focusing on inference and reasoning workloads.
  • The agreement reduces Cerebras’ reliance on UAE firm G42 and is cited in reports as strengthening its fundraising efforts and plans to refile for an IPO.