Particle.news
Download on the App Store

OpenAI Strikes Multi-Year Cerebras Deal for 750MW of Compute Through 2028

The multi-year commitment adds a dedicated low-latency inference platform to OpenAI’s multi-vendor compute strategy.

Overview

  • Cerebras will deliver 750 megawatts of compute to OpenAI starting this year, with capacity coming online in tranches through 2028.
  • OpenAI says the integration provides a dedicated low-latency inference solution to speed responses and make interactions feel more natural.
  • The agreement is reported to be worth more than $10 billion over the life of the contract, according to people familiar with the matter.
  • Cerebras will supply cloud services powered by its wafer-scale chips, building or leasing data centers to run inference and reasoning workloads for OpenAI.
  • The deal broadens OpenAI’s hardware base and helps diversify Cerebras beyond UAE-based G42, and reports say Cerebras is in talks to raise nearly $10 billion at about a $22 billion valuation.