Did AMD say “checkmate” to Nvidia?

When it comes to semiconductor stocks, companies like Nvidia are likely to be on the radar. (NVDA -3.00%) and advanced microdevices (AMD -4.76%) is the first thing that comes to mind. Both companies specialize in designing advanced chips called graphics processing units (GPUs), which are essential components for training generative AI models and machine learning.

For the better part of the past two years, Nvidia has emerged as the undisputed leader in the chip space thanks to its extensive lineup of GPUs compared to those offered by AMD. But there’s another side to Nvidia’s computing and networking business that’s rarely talked about. But it’s this very feature that allows Nvidia to maintain its market-leading position when it comes to data center GPUs.

Below, we explain what makes Nvidia such a force, detail AMD’s latest moves to compete more closely with its rivals, and how the chip king is responding .

Nvidia has a hidden gem in its business…

GPUs in data centers perform sophisticated calculations 24 hours a day. But how do these chipsets work and what do they actually do?

For Nvidia, the answer lies in Computing Unified Device Architecture (CUDA), a software platform layered on top of GPUs. The company’s chips could technically run on other platforms, but developers would be limited in what they could accomplish. Therefore, using CUDA in conjunction with Nvidia GPUs is the perfect solution to take advantage of its wide range of capabilities.

By building an end-to-end suite of tightly integrated hardware and software, Nvidia has captured a 90% share of the data center GPU market. Nevertheless, AMD seems to have a firm grip on this, which could be the catalyst for even faster growth for Nvidia’s rival.

Image source: Getty Images.

…but AMD is upping the ante

There’s no doubt that Nvidia has a huge lead over AMD, but there are some subtle signs that the company may be losing market dominance. As we discussed in another article, Nvidia’s data center GPU sales are showing noticeable signs of slowing down. And at the same time, AMD’s GPU business has expanded significantly and is now growing at a similar rate to Nvidia’s business.

One of the reasons for AMD’s rapid growth is the overwhelming success of its MI300X accelerator, with customers including Microsoft, Oracle, and Meta Platforms. Each of these big tech companies is also a major customer for Nvidia, but most notably, they have begun to diversify their GPU clusters and move toward lower-cost alternatives offered by AMD.

In addition to AMD’s accelerators, the company also offers a software platform called ROCm. So far, CUDA has been more widely adopted than ROCm, but the difference in the trends in data center operations offered by Nvidia and AMD means that ROCm will be more popular as AMD looks to gain incremental market share with GPUs. I think it could be a sign that you’re ready to make the leap.

Nvidia’s latest deal is a long-standing chess move

I’m sure I’m not the only one who noticed AMD’s rapid growth in the data center GPU market last year. In late December, Nvidia completed its $700 million acquisition of a company called Run:ai. The company is an Israel-based startup that describes its expertise as “efficient cluster resource utilization for AI workloads.”

I think the acquisition of Run:ai is a very wise move as it highlights Nvidia’s approach to maintaining a tightly integrated ecosystem for customers where AI workloads are trained. I am. As a result, I think it will be more difficult for customers to migrate to alternatives such as AMD’s ROCm.

While I remain bullish on AMD and believe the company is making a number of important strategic moves, I do not believe ROCm is a checkmate move against Nvidia. For now, Nvidia remains safely on its throne as the king of the GPU space.