AMD’s CEO Lisa Su: Fear of an AI Bubble Is Exaggerated

Earlier this year, WIRED reported that AMD CEO Lisa Su was “out for Nvidia’s blood.” While AMD remains significantly smaller than the behemoth that is Nvidia—boasting market caps of $353 billion and $4.4 trillion, respectively—Su’s company is picking up momentum. During her appearance at WIRED’s Big Interview conference in San Francisco today, she focused on another pressing issue: the AI bubble.

When WIRED senior writer Lauren Goode inquired whether the tech sector is experiencing an AI bubble, Su firmly replied, “emphatically, from my perspective, no.” According to her, the AI sector will require vast numbers of chips from companies like AMD, and the apprehensions about a bubble, she noted, are “somewhat overstated.”

Such a stance may seem audacious, but audacity defines Su’s approach. Since assuming the role of CEO in 2014, she has propelled AMD’s market capitalization from $2 billion to $300 billion. Presently, Su is making substantial bets on the escalating demand for computing power for AI and the necessary data centers to support that demand.

Nevertheless, AMD faces significant challenges ahead. One involves the construction of data centers, and another is ensuring its chips reach as many customers as possible. During the interview, Goode raised the issue of selling chips to China. Su confirmed that AMD will incur a 15 percent tax, introduced by the Trump administration, on MI308 chips it plans to resume sending to China. Although the US government previously halted sales of these chips to China, it began reassessing applications again over the summer. Earlier this year, AMD indicated that US export limitations on MI308 chips would cost the company approximately $800 million.

In a major move earlier this year, AMD struck a deal with OpenAI, which entails deploying 6 gigawatts of AMD’s Instinct GPUs over several years. As part of this agreement, AMD also allowed OpenAI to purchase 160 million shares of the company’s stock for a penny each, effectively giving OpenAI a 10 percent stake. The initial gigawatt deployment is scheduled for the latter half of next year.

This is just one of multiple significant investments AMD is making in AI data centers designed to drive artificial intelligence forward. When it comes to competition from Nvidia or even giants like Google and Amazon, who have their own chip initiatives, Su expressed that her primary concern is not competitors. “When I look at the landscape, what keeps me up at night is ‘How do we move faster when it comes to innovation?’” Su stated.

Su believes that AI is still in its nascent stages, and her company must be prepared to provide chips for the future. “As good as the models are today,” she asserts, “the next one will be better.” There’s immense potential in AI, and “there’s no reason not to continue advancing that technology” into the future.