AMD is receiving assistance with the development of its AI accelerators. According to Bloomberg, Microsoft is offering support to the chip company to reduce its reliance on Nvidia’s GPUs for the Microsoft Azure cloud platform in the future. Bloomberg also reports that AMD would contribute to the development of a Microsoft AI accelerator, although the Redmond company denies this.
The collaboration would mean that AI applications on Azure will not only have to run on Nvidia hardware in the future. However, AMD still has some catching up to do: Nvidia currently holds 80 percent of the server market. In addition, many applications are optimized for the CUDA and tensor cores found in recent Nvidia chips. Microsoft is currently using “tens of thousands” of Nvidia GPUs in data centers.
Additional Motives
In addition to the prospect of cheaper servers, Microsoft has other motives for the collaboration. Nvidia and AMD are currently the only major players in the GPU market, although Intel recently entered this market as well. A lot of AI will undoubtedly be processed in the cloud, but increased competition can also benefit on-premises organizations. Having an alternative for data that needs to be processed by AI in the on-premises data center or at the edge is beneficial.
Furthermore, with the introduction of Windows 12 in 2024, there will be a strong focus on a wide range of AI applications in the operating system. If Microsoft can ensure that both AMD and Nvidia can provide that, it can make the transition to Windows 12 easier.
With competition between AMD and Nvidia intensifying specifically for AI purposes, the AI hardware market is becoming more competitive and complex. Earlier, Google already claimed that its own AI supercomputer was faster and more cost-effective on the self-developed TPUs (Tensor Processing Units) than Nvidia’s A100 GPUs. The competition for Nvidia will become more intense, even if it will be years before it has to concede on a large scale in the server market. Before that, it has invested many years in the development of AI accelerators. These are not only in business GPUs, but also in consumer products since 2018, with the launch of the GeForce RTX 2000 series.
