Cerebras CEO Feldman Urges Balanced View of AI Risks and Benefits

Cerebras Systems CEO and co-founder Andrew Feldman urged the artificial intelligence industry to balance warnings about catastrophic risks with discussion of AI's potential benefits, including advances in health care.

Speaking Oct. 13 at TechCrunch Disrupt 2026 in San Francisco, Feldman said some AI leaders “maybe shouldn’t be allowed to talk in public” because their focus on worst-case outcomes is overshadowing the technology’s promise.

He said AI leaders should also explain how the technology could improve medicine, including a future in which people’s grandchildren “never meet anybody who dies of cancer.”

Feldman acknowledged that powerful technologies can be misused. “I don’t think there are powerful technologies that don’t come with the challenge of ‘what happens when powerful technologies are in the hands of bad people?’” he said. “I think you can’t make one and not have the other.”

He said the industry should recognize AI’s power, approach its development carefully and regulate it reasonably.

Feldman also attributed some of the negativity around AI and new data-center construction to the data-center industry’s failure to work closely with local communities, particularly in the United States. He said Brazil, Canada, Finland, Iceland, Norway, Mexico and Sweden are encouraging companies to build data centers.

By the end of 2027, Cerebras expects its operational and contracted data-center capacity to exceed 600 megawatts. The company also projects that its manufacturing capacity will grow more than tenfold in 2026.

Feldman said efficiency improvements must advance across both hardware and software. Hardware vendors can make computing more efficient while model makers improve their models, he said, with both efforts needed to achieve meaningful progress.

The appearance followed Feldman’s earlier discussion of AI scaling limits as Cerebras expands its computing infrastructure.