Cerebras Systems just drew a line in the sand. The wafer-scale chip maker announced it will partner with essentially every major AI data center component supplier, with one very conspicuous exception: Nvidia.

CEO Andrew Feldman made the declaration on June 3 during the Bloomberg Tech conference in San Francisco, framing the initiative as the formation of a “non-Nvidia” alliance.

From underdog to OpenAI’s partner

The company secured a landmark multi-year deal with OpenAI in January 2026, committing to provide 750 megawatts of AI inference capacity. To put that number in perspective, a single megawatt can power roughly 750 average US homes.

Cerebras completed a successful IPO earlier in 2026, a milestone that looked far from guaranteed after the company withdrew an earlier filing in 2024. The reason for that withdrawal was telling: concerns over reliance on a single customer. The fact that Cerebras came back to public markets with a diversified client base, including a collaboration with Amazon Web Services for AI inference workloads, suggests the company has addressed that vulnerability head-on.

Founded in 2016, Cerebras built its reputation on a genuinely different approach to chip design. Its core product, the Wafer-Scale Engine, is substantially larger than standard GPUs. Rather than cutting a silicon wafer into hundreds of individual chips, Cerebras uses the entire wafer as one massive processor.

What this means for investors

For those watching Cerebras specifically, the newly public company presents a different risk profile. Its IPO was initially delayed over customer concentration concerns, and while the OpenAI and AWS partnerships diversify its revenue base, the company is still early in proving it can manufacture and deliver wafer-scale chips at the volumes required by hyperscale data centers. Semiconductor manufacturing at this scale is notoriously unforgiving. A single yield issue could derail delivery timelines.