Cerebras CEO says AI optimism is being drowned out by doomsday warnings

Cerebras CEO and co-founder Andrew Feldman has criticised the increasingly gloomy tone of AI debate, arguing that warnings about catastrophic risks are overshadowing the technology’s possible benefits. He made the comments during an interview at TechCrunch Disrupt 2026. ¹

A counterpoint to AI doomerism

Feldman said some AI leaders may be talking about the risks so loudly that they are failing to explain what useful progress could look like. He pointed to potential advances in areas such as cancer treatment, while acknowledging that powerful technology can also end up in the hands of bad actors. ¹

The Cerebras chief did not dismiss regulation. Instead, he argued that the industry needs to be thoughtful and that governments should regulate AI reasonably. He also said data-centre companies had not always worked well enough with local communities during the recent build-out, particularly in the United States. ¹

Efficiency does not mean less demand

Feldman also rejected the idea that more efficient AI hardware would automatically shrink demand for Cerebras or rival chipmakers. His argument is that hardware vendors and model developers can improve efficiency together while still delivering more capable systems. ¹

That position matters because Cerebras sells AI compute infrastructure to customers including OpenAI, Block, Figma, GSK and Lovable, giving the company a direct commercial interest in the technology’s continued expansion. ¹

Our opinion

Feldman’s optimism deserves to be read as an industry argument, not a neutral forecast: Cerebras benefits if AI investment keeps growing. Still, the useful part of his message is the attempt to hold two ideas together. AI risks need serious oversight, but treating every discussion as an apocalypse also makes it harder to talk clearly about engineering, regulation and the benefits people might actually experience.