10 hours ago
AMD CEO on taking on Nvidia: People used to think that $$500 billion was a very large number, and now ...
AMD appears to be aiming to do what Intel has not been able to: take on AI chip giant Nvidia. Advanced Micro Devices (AMD) CEO Lisa Su announced the company's latest MI350 series AI processors, positioning them as formidable competitors to Nvidia's dominant chips in the rapidly growing AI accelerator market. Speaking at a company event in San Jose, California, Su projected the AI market to surpass $500 billion by 2028, revising her earlier estimate as demand continues to surge.
Though AMD remains a distant second to Nvidia in AI accelerators, it aims to start catching up with these new products. 'People used to think that $500 billion was a very large number,' she said in a briefing following her presentation. 'Now it seems well within grasp.'
AMD claims its MI355 chips outperform Nvidia's
The MI355 chips, which began shipping this month, boast a 35-fold performance increase over their predecessors and outperform Nvidia's offerings in speed, Su claimed. She highlighted that the chips deliver up to 40% more tokens per dollar, offering cost efficiency for customers. AMD also introduced the MI350 Series GPUs and the Helios AI Rack Scale solution, set for release next year, alongside a developer cloud access program to spur innovation through open architecture -- a contrast to Nvidia's approach.
Despite Nvidia's entrenched position and high demand for its GPUs, which often face long wait times, AMD is gaining traction. Su noted that seven of the top 10 AI customers, including India's Reliance Jio, are deploying AMD Instinct Accelerators. The MI350 series, now in production with a fourfold compute increase over the prior generation, will be commercially available from Q3.
Su's presentation featured discussions with industry leaders, including OpenAI's Sam Altman and executives from Meta, Microsoft, and Oracle, underscoring AMD's growing partnerships. While AMD trails Nvidia in the AI accelerator space, Su emphasized that the new chips signal a push to close the gap, restoring momentum after a February forecast of slower data center growth disappointed some analysts.
'People used to think $500 billion was a very large number,' Su said in a briefing. 'Now it seems well within grasp.' With its latest innovations, AMD is betting on open architecture and performance gains to challenge Nvidia's dominance in the AI revolution.
AI Masterclass for Students. Upskill Young Ones Today!– Join Now
