AMD will decrease AI GPUs to concede with US laws, trade in China

August 3, 2023: AMD will decrease AI GPUs to concede with US laws, trade in China.

Lisa Su, CEO of AMD, stated on an earnings call that China is an “important” need and that the semiconductor giant wants to fully comply with U.S. export controls.

“As we think about definitely the accelerator market, we plan to, of course, be fully compliant with U.S. export controls, but we do believe there’s an option to develop products for our customer set in China that is looking for A.I. solutions, and we’ll resume working in that direction,” Su said.

Accelerator chips are the semiconductors required to train vast data for artificial intelligence applications.

AMD is gearing up to increase production of its MI300 chip which it is positioning as a rival to Nvidia’s pictures processing units used for A.I. training. Nvidia dominates the need, but AMD hopes to challenge it with its latest chip.

Earlier this year, the U.S. government restricted Nvidia from selling its A100 and H100 chips to China. The H100 is one of Nvidia’s key A.I. chips. Nvidia decided to create a chip with tweaks to the H100′s specifications that complied with the export curbs.

Intel also made a modified version of its Gaudi 2 A.I. chips for the Chinese market.

China stays a lucrative market for U.S. chipmakers, particularly in A.I., where there are few homegrown alternatives to Nvidia.

A lot is riding on its MI300 AI chip for AMD as it looks to take on Nvidia. The company expects the chip to help it rapidly grow its data center industry for the rest of the year.

Su said AMD is looking at around 50% growth in the year’s second half versus the first half in its data innards industry, partly thanks to the new A.I. chip.

About Us

We provide the insights on leaders who are responsible for taking their organization to new heights, all the while bringing together a group of talented individuals.

Recent Posts | Make Profit Not Waste | Stephen Pratt

Supply chain leaders, including VPs, often find themselves making bets with unknown financial consequences. Having started with largely infeasible plans generated by deterministic “one-number” planning systems, compounded by unexpected supply or demand disruptions, they are forced into last-minute adjustments to meet operational and financial goals.

Intelichain | Revolutionary Supply Chain Planning Solutions | Roei Aviram

Unsurprisingly, today’s supply chain faces numerous issues, including sustainability, technology, global economic and political instability, talent management, and supply chain resilience. Advanced forecasting algorithms and predictive analytics are used in supply chain management to help organizations of all sizes make better decisions by providing insights into what’s going on in their business at any given moment and predicting future trends.

AMD will decrease AI GPUs to concede with US laws, trade in China