Cerebras: Andrew Feldman Advocates Against AI Alarmism

⚡

Key Takeaways

1Andrew Feldman (Cerebras) dismisses alarmism about AI and advocates for a constructive approach

2He links some of the distrust to data center management and cites countries favorable to their establishment

3He believes that efficiency gains will not reduce the demand for chips and highlights the potential of AI

4Cerebras supplies major players like OpenAI and GSK

💡Why it matters — The perspective of a key player in AI hardware adds nuance to the debates on risks and emphasizes the importance of appropriate regulation and better communication about potential benefits.

At TechCrunch Disrupt 2026, Andrew Feldman advocated for a more constructive vision of AI, which he wants to be powerful, regulated, and better explained to the public. He links part of the current skepticism to local tensions surrounding data centers and believes that efficiency gains will not dampen the demand for chips. The CEO also acknowledges the interest of Cerebras, a supplier to players like OpenAI and GSK.

Efficiency and infrastructure will not dampen demand, according to Feldman

Andrew Feldman claims he is not concerned that efficiency advancements will reduce demand for Cerebras chips or those of other Nvidia competitors. According to him, hardware providers can enhance computing efficiency, while model creators can optimize their architectures to achieve the same goals. He believes that these joint efforts are necessary to enable significant breakthroughs. Feldman also attributes part of the negative perception of AI to how the data center industry has managed new AI-related projects. He considers that the sector has not sufficiently collaborated with local communities in recent years, particularly in the United States. He cites the energy, oil, and gas sectors as examples of industries that have not adequately involved local stakeholders, which has caused issues. Feldman compares this situation to American NIMBYism and mentions Brazil, Canada, Finland, Iceland, Norway, Mexico, and Sweden as countries where data center construction is encouraged.

⚡Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisis ton rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

A "reasonable" framework for a technology deemed powerful

Andrew Feldman believes that any powerful technology comes with the risk of misuse and that it is impossible to separate power from risk. He asserts that we must explicitly acknowledge the power of AI, seriously consider its implications, and establish reasonable regulation.

Against catastrophism: jokes, debates, and television parody

Andrew Feldman considers that alarmism surrounding AI does not benefit the industry. At TechCrunch Disrupt 2026, he joked that some AI leaders perhaps should not be allowed to speak in public. He referred to the wave of negative discourse, ranging from scenarios of out-of-control models to expert debates on the possibility that AI could annihilate humanity by the end of the decade. This discourse has become so common that it was recently parodied in a Saturday Night Live sketch, where a character declared, "We do not endorse what we are doing." Feldman believes that, conversely, the promise and potential of AI are not highlighted enough.

Acknowledged commercial interest and emphasized promise

Cerebras counts among its clients OpenAI, Block, Figma, GSK, and Lovable, placing Andrew Feldman at the heart of the AI market. He believes that industry leaders should better explain to the public that AI could enable their grandchildren to never know someone who dies from cancer. Andrew Feldman is the CEO and co-founder of Cerebras Systems.