Nvidia Faces Growing Competition from AMD in AI Chip Market

Nvidia (NASDAQ: NVDA) dominates the market for graphics processing units (GPUs) used in data centers, particularly for training and implementing artificial intelligence (AI) models. CEO Jensen Huang predicts that data center operators will invest $1 trillion in AI infrastructure over the next five years, highlighting the significant opportunity in this sector.

Last year, Nvidia held an estimated 98% share of the data center GPU market. However, increased competition, especially from Advanced Micro Devices (NASDAQ: AMD), threatens its position.

AMD’s Expanding Role in the AI Chip Landscape

AMD is becoming a vital player in the AI chip field. Its chips are integral to popular electronic devices, including Sony‘s PlayStation 5 and Tesla‘s vehicle infotainment systems. The company offers a range of products, including GPUs, CPUs (central processing units), and NPUs (neural processing units), particularly favored by personal computer manufacturers.

Notably, AMD’s Ryzen AI 300 series NPUs are expected to power over 100 platforms from companies like Microsoft, Asus, Lenovo, and HP. With a focus on processing AI tasks directly on the device, these chips improve user experience by decreasing dependence on external data centers.

AMD’s competition with Nvidia is intensifying in the data center GPU space. The company will announce its financial results for the third quarter of 2024, which ended on September 30, on October 29. Here are two key metrics investors should closely watch.

Image source: Getty Images.

Crucial Metric #1: Data Center Revenue

AMD began shipping its MI300X data center GPU at the end of last year, which competes against Nvidia’s flagship H100. The MI300X has already attracted major clients like Microsoft, Oracle, and Meta Platforms, with some users reporting performance and cost benefits over the H100.

The MI300 series achieved $1 billion in sales in just two quarters, marking the fastest ramp-up in AMD’s product history. Looking ahead, AI GPUs will significantly influence AMD’s data center revenue growth.

In the first quarter of 2024, AMD reported $2.3 billion in data center revenue, an 80% increase year-over-year. Then in the second quarter, the company achieved $2.8 billion, reflecting an impressive 114% rise compared to the previous year.

This trend of increasing revenue positions AMD favorably as it approaches its Q3 report, indicating that the MI300 is capturing market share. With the data center segment contributing nearly half of AMD’s total revenue, it will be a focal point in their upcoming announcement.

Crucial Metric #2: Sales Guidance for Data Center GPUs

The MI300X is just the first step for AMD in AI data center GPUs. At a recent “Advancing AI” event, CEO Lisa Su highlighted upcoming products, including the MI325X and the innovative MI350X.

Scheduled to begin shipping in Q1 2025, the MI325X will compete with Nvidia’s H200 GPU (the H100’s successor). Su has stated that it features 80% more high-bandwidth memory and offers 30% improved inference performance compared to the H200. Thus, while AMD’s products are launching after Nvidia’s, they offer technological advantages that could attract customers.

Nevertheless, Nvidia is gearing up to release GPUs based on its new Blackwell architecture in late fiscal 2025. The Blackwell GPU systems promise up to 30 times the inference performance of H100 systems and improved cost efficiency, placing pressure back on AMD’s forthcoming offerings.

AMD is also developing its own architecture, named CNDA (Compute DNA) 4, which serves as the basis for the MI350X GPU. This product aims to provide up to 35 times greater performance than its predecessors. However, fans will have to wait until the second half of 2025 for its release. Investors should keep an eye on updates during the Q3 earnings call.

As for sales forecasts, AMD initially expected to sell $3.5 billion in data center GPUs for 2024, but this estimate has been increased to $4.5 billion after positive first and second-quarter performances. A further revision could indicate ongoing strong demand for the MI300 and reinforce perceptions that AMD is a viable alternative to Nvidia.

Additionally, increasing demand for the MI300 will likely boost sales of the MI325X and MI350X in 2025, which could bode well for AMD’s stock as we head into the new year.

Is Now the Right Time to Invest in AMD?

Before considering an investment in Advanced Micro Devices, take note of this:

The Motley Fool Stock Advisor team has identified what they believe are the 10 best stocks to buy right now… and Advanced Micro Devices isn’t one of them. The selected stocks have the potential for significant returns in the coming years.

For context, consider that when Nvidia was listed on April 15, 2005, investing $1,000 at that time would have grown to $845,679!

Stock Advisor offers investors a clear strategy for success, including portfolio guidance, regular updates from analysts, and two new stock picks each month. Since its inception in 2002, Stock Advisor has more than quadrupled the returns of the S&P 500.

See the 10 stocks »

*Stock Advisor returns as of October 21, 2024

Randi Zuckerberg, a former director of market development and spokesperson for Facebook, is a member of The Motley Fool’s board of directors. Anthony Di Pizio has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Advanced Micro Devices, HP, Meta Platforms, Microsoft, Nvidia, Oracle, and Tesla. The Motley Fool recommends options related to Microsoft. The Motley Fool has a disclosure policy.

The views and opinions expressed herein are those of the author and do not necessarily reflect those of Nasdaq, Inc.