A clash between AMD and NVIDIA could be in store for CES. AMD CEO Lisa Su has confirmed for the first time that the company is set to release its next-gen PC GPUs early next year.
“In gaming graphics, revenue declined year-over-year as we prepare for the transition to our next-gen Radeon GPUs based on our RDNA 4 architecture,” Su told investors on AMD’s third-quarter earnings call. “In addition to strong increases in gaming performance, RDNA 4 significantly improves ray-tracing performance and adds new AI capabilities. We are on track to launch the first RDNA 4 GPUs in early 2025.”
The timing pretty much suggests that AMD will introduce those RDNA 4-based graphics cards at CES in early January. It’s rare for the company to unveil desktop GPUs at the trade show (laptop cards are typically the order of the day for AMD at that event). However, it is widely expected that NVIDIA will use its CES keynote address to introduce its next-generation 50-series GeForce RTX GPUs. We may get a little more clarity on this front when NVIDIA announces its own Q3 earnings results on November 19.
As PCWorld notes, AMD’s first RDNA 4 GPUs are expected to deliver mid-range performance at a similar price point to increase its market share. AMD’s gaming business (which includes the company’s GPU division) saw a 69 percent year-over-year decline in revenue in Q3, to $462 million.
The Biden administration has unveiled its “AI proliferation rule,” which aims to restrict exports of GPUs that are most coveted for AI applications. Although it doesn’t name the country, it is largely seen as a means to prevent China from overtaking the US in AI development.
The rule proposes three licensing tiers. The first tier is unrestricted and includes the domestic market as well as 18 strategic allies. Most countries fall into the second tier, which will have limits on how much computing power they can import from the US, including top GPUs. The third tier includes China, Russia, Iran, and North Korea, and it effectively prevents US companies from selling their most powerful GPUs there.
US-based companies will also be prevented from sharing many details of their AI software models with countries outside the first tier, and will have to get permission from the federal government before building large data centers in any second-tier country.
Several parties, including the Semiconductor Industry Association (SIA), released statements condemning the decision, believing that the restrictions will push countries to work with China. “The new rule risks causing unintended and lasting harm to the US economy and global competitiveness in semiconductors and AI by ceding strategic markets to our competitors,” the SIA wrote.
NVIDIA has also objected to it. Ned Finkel, the company’s vice president of government affairs, said the Biden administration “seeks to undermine America’s leadership with a 200-plus page regulatory morass, crafted in secret and without proper legislative review.” Comments on the rule have been given for 120 days, so the question is still open as to whether it will remain in place in the incoming Trump administration.