Earlier this year WIRED said AMD CEO Lisa Su was “out for Nvidia’s blood.” The US chipmaker is still petite compared to the behemoth that is Nvidia – its market capitalization is $353 billion and $4.4 trillion, respectively – but Su’s company is gaining momentum. Today, when Su took the stage at the WIRED Gigantic Interview conference in San Francisco, she had something else in her sights: the artificial intelligence bubble.
When asked by WIRED senior writer Lauren Goode whether the tech industry is in an artificial intelligence bubble, her answer was “definitely, from my perspective, no.” The artificial intelligence industry will need dozens of chips from companies like AMD, and fears about such a bubble, Su said, are “a little overblown.”
It may sound brave, but bravery is Su’s art. Since becoming AMD’s CEO in 2014, she has increased the company’s market capitalization from $2 billion to $300 billion. Su is now betting large on demand for much more computing power for artificial intelligence and the data centers needed to provide it.
However, AMD faces many obstacles. One is to build all these data centers, and the second is to make its chips available to as many customers as possible. During the discussion, Goode asked AMD’s CEO about chip sales to China. It confirmed that AMD will pay the 15 percent tax imposed by the Trump administration on MI308 chips, which it plans to resume shipping to China. The U.S. government previously halted chip sales to China but began reviewing applications again in the summer. AMD said earlier this year that U.S. export restrictions on MI308 chips would cost the company about $800 million.
Earlier this year AMD did a huge business with OpenAI, under which the AI company will deploy 6 gigawatts of AMD Instinct GPUs over the course of several years. As part of the deal, AMD agreed to allow OpenAI to purchase 160 million shares of the company at a penny per share. effectively giving him 10 percent of the company. The first gigawatt deployment is expected to take place in the second half of next year.
This is one of several large bets AMD is making on AI data centers to power artificial intelligence. Su said she wasn’t worried about competition from Nvidia or even Google or Amazon, which have their own chipmaking plans. “When I look at the landscape, I can’t sleep at night: ‘How can we move faster when it comes to innovation?’” Su said.
Su believes that artificial intelligence is still in its infancy and her company needs to be ready to supply chips for the future. “No matter how good today’s models are,” he says, “the next one will be better.” Artificial intelligence has enormous potential and “there is no reason not to promote this technology” in the future.