
The chip war escalates! AMD MI355 chip challenges NVIDIA B200. Lisa Su boldly claims a $500 billion market is "within reach."

AMD CEO Lisa Su stated that the latest AI processor MI355 is expected to challenge NVIDIA, with the artificial intelligence market projected to exceed USD 500 billion in the next three years. Although AMD still lags behind NVIDIA in the AI accelerator market, it plans to narrow the gap with new products. The MI355 chip offers a 35-fold performance improvement at a price lower than competitors. Investor reaction has been lukewarm, with the stock price dropping by 2.2%
According to Zhitong Finance APP, AMD (AMD.US) CEO Lisa Su stated that the company's latest AI processors are expected to challenge NVIDIA chips, and she anticipates that the market size will exceed $500 billion in the next three years.
Lisa Su mentioned at a company event in San Jose, California, that the new MI350 series chips outperform NVIDIA's competing products, achieving significant breakthroughs compared to previous generations. The MI355 chip, which began shipping this month, has a performance improvement of 35 times.
Although AMD still lags far behind NVIDIA in the AI accelerator market (chips used for developing and running artificial intelligence tools), the company plans to narrow the gap with its new products. The stakes in this competition are unprecedented: Lisa Su previously predicted that the market size would reach $500 billion by 2028, and now she believes this figure will be surpassed.
"People once thought $500 billion was an unattainable number," she said at a briefing after her speech, "but now it seems within reach."
In February of this year, AMD's forecast for data center business growth was below some analysts' expectations. The company believes that the new MI series products will revitalize growth momentum and prove its ability to compete directly with larger rivals.
AMD claims that the MI355 outperforms NVIDIA's B200 and GB200 products when running AI software, matches or exceeds performance in code generation, and is priced significantly lower than competing products.
Investors reacted lukewarmly to AMD's new product launch, with the stock price falling 2.2% to $118.50 at Thursday's close in New York, erasing this year's modest gains.
As two major suppliers of advanced computer graphics chips, NVIDIA and AMD's products have become core components for AI development. The global giants' hundreds of billions of dollars in infrastructure investments have led to a continuous shortage of chips, with prices for individual chips soaring to tens of thousands of dollars.
The AI accelerator business has helped AMD break free from the pressure of its old rival Intel in the PC processor field, but NVIDIA has surpassed both. While AMD generates billions in revenue from AI accelerators, NVIDIA's annual revenue has exceeded $100 billion.
Lisa Su also previewed the MI400 series, set to launch next year, claiming it will establish a clear advantage over NVIDIA's chips at the same time. She stated that the new products will increase memory and high-speed information reading components, which are crucial for running AI software.
OpenAI CEO Sam Altman appeared on stage, revealing that the development of the MI400 absorbed suggestions from the company's engineers. Altman expressed initial skepticism when he first heard about the MI400's specifications, thinking it was "impossible," but noted that AI services have transformed from a novelty into a vehicle for commercial value, leading to an explosive demand for computing infrastructure.
The key question is whether AMD can expand its AI chip sales. Due to U.S. trade restrictions, both AMD and NVIDIA are prohibited from selling their highest-performance components to China. Lisa Su stated that she is actively lobbying the Trump administration to ease export restrictions on AI components to other countries She revealed that she had consulted Washington prior to signing recent orders from Saudi Arabia to ensure compliance, and emphasized the need for more trade freedom to secure the core position of American technology in the global AI development