Earlier this 12 months, WIRED stated that AMD CEO Lisa Su was “out for Nvidia’s blood.” The American chipmaker remains to be small in comparison with the juggernaut that’s Nvidia—their market caps are $353 billion and $4.4 trillion, respectively—however Su’s firm is gaining steam. At this time, when Su took the stage at WIRED’s Large Interview convention in San Francisco, she had one thing else in her sights: the AI bubble.
When requested by WIRED senior author Lauren Goode if the tech trade is in an AI bubble, her response was “emphatically, from my perspective, no.” The AI trade goes to wish scores of chips from firms like AMD, and fears of such a bubble, Su stated, are “considerably overstated.”
Which may sound daring, however boldness is Su’s entire deal. Since she turned AMD’s CEO in 2014, she has elevated the corporate’s market cap from $2 billion to $300 billion. Now, Su is betting huge on the necessity for far more computing energy for AI, and the information facilities wanted to offer that.
Nonetheless, there are many hurdles forward for AMD. One is all of these knowledge facilities being constructed, and one other is getting its chips out into the arms of as many purchasers as doable. In the course of the dialogue, Goode requested the AMD CEO about promoting chips to China. She confirmed that AMD can pay a 15 % tax instituted by the Trump administration on MI308 chips it plans to renew delivery to China. The US authorities beforehand halted gross sales of the chips to China, however then started reviewing purposes once more over the summer season. AMD stated earlier this 12 months that US export restrictions on the MI308 chips would value the corporate roughly $800 million.
Earlier this 12 months, AMD made an enormous deal with OpenAI, below which the AI firm will deploy 6 gigawatts of AMD’s Intuition GPUs over the course of a number of years. As a part of the deal, AMD agreed to permit OpenAI to purchase 160 million shares of the corporate’s inventory for a penny per share. successfully giving it a ten % stake within the firm. The primary gigawatt deployment is about to rollout within the second half of subsequent 12 months.
It’s one in every of a number of huge bets AMD is making on AI knowledge facilities to energy synthetic intelligence. What Su stated she’s not frightened about is competitors from Nvidia, and even Google or Amazon, each of which have their very own chip-making plans. “After I have a look at the panorama, what retains me up at night time is ‘How will we transfer sooner in terms of innovation?’” Su stated.
Su believes that AI remains to be in its infancy and her firm must be prepared to offer chips for the longer term. “Nearly as good because the fashions are at the moment,” she says, “the following one will probably be higher.” There’s enormous potential in AI, and “there’s not a purpose to not preserve pushing that expertise” into the longer term.
