Prime Highlight
- OpenAI has entered a major multi-year partnership with Cerebras to speed up AI responses and support real-time AI applications significantly.
- The deal reflects OpenAI’s strategy to build a flexible, low-latency compute infrastructure for its rapidly growing global user base.
Key Facts
- The agreement, valued at over $10 billion, will provide 750 megawatts of compute power to OpenAI between 2025 and 2028.
- Cerebras, known for its AI-focused chips, is reportedly seeking to raise $1 billion at a valuation of around $22 billion amid IPO delays.
Background
OpenAI has signed a major multi-year deal with AI chipmaker Cerebras to strengthen its computing capacity and improve the speed of its AI services. The agreement, valued at more than $10 billion, will see Cerebras deliver 750 megawatts of compute power to OpenAI from 2025 through 2028, according to people familiar with the matter.
Both companies said the partnership aims to deliver faster AI outputs for OpenAI’s growing global user base. OpenAI noted in a blog post that the new systems will reduce response times for tasks that currently take longer to process. The deal reflects OpenAI’s push to support real-time AI applications as demand for advanced AI tools continues to rise.
Andrew Feldman, co-founder and CEO of Cerebras, compared the shift to a major change in internet history. He said that just as broadband reshaped online services, real-time AI inference will transform how people interact with artificial intelligence.
Cerebras, which has been operating for more than a decade, has gained strong attention since the launch of ChatGPT in 2022 and the rapid growth of the AI sector. The company designs special chips for AI tasks and says its systems can run faster than traditional GPU-based solutions for some jobs.
The company filed confidentially for an initial public offering in 2024 but delayed the listing several times. Even so, Cerebras has continued to draw strong interest from investors. Reports this week said the firm is in talks to raise another $1 billion at a valuation of around $22 billion.
OpenAI executive Sachin Katti said the company is building a flexible compute strategy. He said Cerebras adds a low-latency inference option that will help deliver faster responses, smoother conversations, and better scaling for real-time AI use.
