Nvidia‘s (NVDA 0.12%) shares soared nearly 2,000% over the past five years and turned the chipmaker into the world’s third-most-valuable company. That growth was largely driven by its brisk sales of data center GPUs for processing AI tasks.
The market’s appetite for those high-end GPUs could continue to outstrip Nvidia’s available supply for the foreseeable future. Its revenue soared 126% in fiscal 2024 (which ended this January), and analysts expect it to maintain a compound annual growth rate (CAGR) of 37% from fiscal 2024 to fiscal 2027 as the artificial intelligence (AI) market expands.
Image source: Getty Images.
The bulls believe Nvidia still has plenty of room to run, even as AMD (AMD 0.50%) ramps up its production of its own AI GPUs and major customers like OpenAI and Microsoft (MSFT -0.17%) start developing their own in-house chips. However, a new threat might be emerging: Some of the world’s largest tech companies have formed a new open-source alliance to prevent Nvidia from dominating the AI market with its closed-source programming platform.
Nvidia’s software platform is one of its greatest strengths
To develop apps optimized for Nvidia’s GPUs, software developers need to access the chipmaker’s closed-source CUDA (Computer Unified Device Architecture) programming platform. Those developers would need to access other programming platforms — like Intel’s OneAPI and AMD’s MIOpen — to port their applications to other brands of GPUs.
Since Nvidia already dominates the data center GPU market, there’s not much of an incentive for developers to rewrite their applications for other GPU brands. Many of Nvidia’s critics believe the symbiotic relationship between its GPUs and CUDA platform supports its near-monopolization the GPU market and locks in its software developers.
Can a new open-source alliance challenge CUDA?
That’s why it wasn’t too surprising when Intel (INTC 0.91%), Qualcomm, Samsung, Arm, Alphabet‘s (GOOG 0.21%) (GOOGL 0.04%) Google Cloud, Broadcom‘s Vmware, Fujitsu, and Imagination Technologies recently teamed up to create the UXL (Unified Acceleration) Foundation.
The UXL Foundation plans to use Intel’s OneAPI as a foundation to create an open-source alternative to Nvidia’s CUDA platform. It doesn’t plan to release its first “mature” platform until the second half of this year, but its arrival could potentially loosen Nvidia’s iron grip on the GPU software and hardware markets.
But Nvidia’s investors shouldn’t worry about the UXL Foundation yet
These tech titans clearly want to challenge Nvidia’s dominance of the AI market, but the UXL Foundation probably won’t make much near-term progress for three simple reasons.
First, Nvidia still develops the world’s most powerful data center GPUs. Even though its customers are dealing with supply constraints for its newest GPUs, they won’t abruptly switch over to Intel or AMD’s discrete GPUs just to save a few dollars.
That’s why the vast majority of Google Cloud’s GPUs are still Nvidia chips, even though it’s been gradually developing and testing its own first-party chips in recent years. Therefore, a battle between the CUDA and UXL platforms won’t even matter unless one of the GPU underdogs takes a meaningful bite out of Nvidia’s dominant market share.
Second, AMD — Nvidia’s closest competitor in the discrete GPU race — hasn’t joined the UXL Foundation yet. Another missing tech giant is OpenAI’s top backer, Microsoft, which recently started buying AMD’s AI GPUs. Therefore, AMD and Microsoft might be quietly working on more ways to promote AMD’s MIOpen as an open-source alternative to CUDA and UXL. That fragmentation could, ironically, weaken the open-source push against CUDA and indirectly help Nvidia.
Lastly, the UXL Foundation plans to expand its open-source programming platform to Nvidia’s GPUs in the future. That roadmap might make it easier for developers to optimize their software for a wider range of GPUs, but it also seems to be a blatant admission that Nvidia will remain the 800-pound gorilla of the GPU market.
Investors shouldn’t overreact to the news
Nvidia’s stock dipped after the UXL announcement, but I think it’s a lot of sound and fury that signifies nothing. These big tech companies are teaming up to create an open-source alternative to CUDA, but that alliance is missing two major players and isn’t co-developing any new chips that could actually challenge Nvidia. Simply put, this headline-grabbing team-up won’t stir up any meaningful headwinds for Nvidia’s near-term growth.
Suzanne Frey, an executive at Alphabet, is a member of The Motley Fool’s board of directors. Leo Sun has positions in Qualcomm. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Microsoft, Nvidia, and Qualcomm. The Motley Fool recommends Broadcom and Intel and recommends the following options: long January 2023 $57.50 calls on Intel, long January 2025 $45 calls on Intel, long January 2026 $395 calls on Microsoft, short January 2026 $405 calls on Microsoft, and short May 2024 $47 calls on Intel. The Motley Fool has a disclosure policy.
