A senior AMD executive has signaled that the company is focusing future chip improvements on AI performance and stepping away from high-end gaming systems.
For chip buyers, the move means more options for AI efforts: “For large-volume chip purchasers, AMD’s shift towards AI is a pivotal move that signals a fundamental change in their market positioning,” said Scott Dylan, the managing partner at NexaTech Ventures, a London-based venture capital fund investing in AI and tech startups.
“AMD’s decision to deprioritize flagship gaming GPUs in favor of AI makes sense from a financial standpoint. The gaming GPU market is mature and Nvidia has a near-monopolistic hold on the high-end segment. AMD is essentially acknowledging that, to capture significant market share, chasing Nvidia’s top-tier products may not be worth the resources,” Dylan said. “Instead, by doubling down on AI chips, particularly with the success of their EPYC and MI300 series in data centers, AMD is positioning itself as a formidable player in an industry where demand for AI accelerators is only set to grow exponentially.”
AMD’s positional shift can be seen by piecing together remarks by Jack Huynh, general manager of the chip maker’s computing and graphics business group, in two separate articles by Tom’s Hardware.
In one, Huynh described how AMD split its GPU microarchitecture into two families back in 2020: RDNA for gaming and CDNA for high-performance computing and AI applications in data centers, but now plans to recombine them under a new name, UDNA.
“We forked it because then you get the sub-optimizations and the micro-optimizations, but then it’s very difficult for these developers, especially as we’re growing our data center business, so now we need to unify it,” he told Tom’s Hardware.
Although that implies that the designs will be optimized for performance in both applications, his other remarks suggest that the focus will be on delivering the highest AI performance at the lowest total cost of ownership (TCO), while in gaming the focus will be on winning market share.
“In the server space, when we have absolute leadership, we gain share because it is very TCO-based. In the client space, even when we have a better product, we may or may not gain share because there’s a go-to-market side, and a developer side; that’s the difference,” Huynh told Tom’s Hardware.
“My priority right now is to build scale for AMD. Because without scale right now, I can’t get the developers,” he said of his approach to the market for gaming consoles and PCs. “If I tell developers, ‘I’m just going for 10 percent of the market share,’ they just say, ‘Jack, I wish you well, but we have to go with Nvidia.’ So I have to show them a plan that says, ‘Hey, we can get to 40% market share with this strategy.’ Then they say, ‘I’m with you now, Jack. Now I’ll optimize on AMD.’ Once we get that, then we can go after the top,” he told the PC hardware publication.
The move is not surprising, both given earlier signals from AMD of its interest in AI hardware as well as difficulties facing Intel due to its lack of perceived AI focus. AMD has made some small gains against Intel, but most question whether it was sufficient.
NexaTech’s Dylan sees the potential focus on chip price-performance as good news for buyers.
Nvidia’s dominance of AI chip development to date “has led to supply shortages and inflated pricing during periods of high demand,” he said, but “AMD’s pivot means a more balanced market, offering an alternative that could potentially drive prices down and improve lead times for large-volume purchases.”
For anyone sourcing chips for AI workloads, this is good news, he said. “The AI market is obviously still in its growth phase, and by focusing its resources here, AMD can innovate faster and provide more competitive pricing and performance-per-dollar ratios. This benefits buyers looking for scalable AI (offerings) across sectors like healthcare, autonomous systems, and financial services.”
Source:: Network World