OpenAI Signs $10 Billion AI Compute Deal with Cerebras Systems

Jan 15, 2026
OpenAI has entered a multi-year agreement worth over $10 billion with AI chipmaker Cerebras Systems to secure 750 megawatts of computing capacity through 2028. The partnership aims to enhance the speed and efficiency of OpenAI’s inference workloads, including ChatGPT.

Bloomberg reports that OpenAI has signed a multi-year agreement valued at more than $10 billion with Cerebras Systems to expand its computing capacity. Under the deal, Cerebras will provide 750 megawatts of compute power through 2028, hosting the infrastructure across multiple data centers.

The agreement will supply OpenAI with dedicated hardware optimized for faster inference, enabling quicker responses in applications such as ChatGPT. Both companies said the rollout will occur in stages and focus on improving real-time AI performance.

Cerebras, known for its wafer-scale chips designed for large-scale AI processing, will operate the systems and sell cloud services powered by its technology. OpenAI will pay for access to these services to support its growing suite of AI products.

Cerebras CEO Andrew Feldman described the partnership as a major milestone for the startup, positioning its technology as a high-speed alternative to GPU-based systems. OpenAI executives said the collaboration strengthens the company’s compute portfolio and supports scaling AI applications to more users.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like Silicon Brief or Daily AI Brief.

Also, consider following us on social media:

AI Chips & Datacenters AI Brief AI Brief (X)
Free newsletter

Silicon Brief

Weekly coverage of AI hardware developments including chips, GPUs, cloud platforms, and data center technology.

AI Token Webinar

Whitepaper

Tensordyne Napier: What If One Rack Could Do the Work of Nine?

Tensordyne

This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.

Read more

You may also like

Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.