AMD’s Lisa Su Meets China Vice Premier As AI Chip Politics Heat Up
AMD CEO Lisa Su made a high-profile China stop this week, meeting Chinese Vice Premier He Lifeng at Beijing’s Great Hall of the People and signalling that AMD wants to keep expanding its presence in the country.
According to Xinhua, He encouraged multinational firms, including AMD, to deepen cooperation with China. The meeting came shortly after the May 13-15 Xi Jinping-Trump summit, which Chinese officials framed as having produced “balanced and positive” results.
For AMD, this is more than just another executive handshake. China remains one of the biggest battlegrounds for AI hardware, and companies like AMD need to keep one foot in that market while also staying aligned with US export rules. That balancing act is getting harder as AI accelerators become a major geopolitical issue.
Su was not part of the business delegation that joined President Trump’s China visit. That group included Apple’s Tim Cook, Tesla’s Elon Musk, and Qualcomm’s Cristiano Amon. Nvidia boss Jensen Huang was also reportedly not on the original list before Trump personally invited him during a stop in Anchorage. Huang later had his own separate meeting with He Lifeng.
The bigger issue behind all this is AI chip access. AMD and Nvidia both received export approvals late last year for China-specific products, with AMD’s MI308 and Nvidia’s H20 reportedly carrying a 15% revenue fee. But approval does not automatically mean smooth sales. Nvidia has reportedly not completed any H200 sales to Chinese buyers, with US licensing delays and Beijing’s own restrictions both creating friction.
It is still not clear whether AMD’s MI308 is moving through the market more successfully, though AMD is believed to be progressing with shipments. AMD’s more powerful MI325X is also in a more complicated position. The chip, rated at 1,300 TFLOPS of FP16 performance and equipped with 256GB of HBM3E plus 6TB/s of bandwidth, is now handled under case-by-case review rules from the US Bureau of Industry and Security. Those rules, which took effect in January, also cover Nvidia’s H200 and include a separate 25% tariff for chips routed through the US before export.
After the Beijing meeting, Su went to Shanghai for AMD’s AI Developer Day keynote. There, she said around five billion people could be using AI every day by 2030, and described China as a key part of AMD’s global operation. AMD currently has more than 4,000 engineers across R&D centres in Beijing, Shanghai, Shenzhen, and Taipei.
Why should Malaysian and SEA tech fans care? Because this chip war affects more than just hyperscale data centres in America and China. AI GPU availability shapes cloud pricing, AI startup costs, gaming hardware roadmaps, and even the future of gaming laptops and creator PCs. If AMD can keep selling into China while scaling US AI infrastructure, it strengthens its position against Nvidia — and that competition is good for the whole hardware ecosystem.
At the same time, AMD is also going big in the US. The company is working on a gigawatt-scale data centre project for OpenAI, with the first phase expected in the second half of this year. So AMD is basically playing both lanes: supporting America’s AI infrastructure push while trying not to lose access to China’s massive AI developer market.
For SEA, including Malaysia, this is worth watching closely. We may not be the centre of the semiconductor power struggle, but we feel the effects through GPU prices, AI cloud access, enterprise hardware supply, and the devices that eventually land on Shopee, Lazada, and local retailers. If the AMD-Nvidia-China-US triangle shifts, the ripple can reach our gaming rigs and AI tools faster than people think.
Source: Tom's Hardware
