News

A big reason why Nvidia (NASDAQ: NVDA) is currently the overwhelming leader in the AI accelerator market is related to software. The company's CUDA software platform, which enables developers to write applications that tap into Nvidia's GPUs for accelerating workloads, has been widely used across academia and industry for nearly two decades. CUDA has become the de facto standard, and it only works with NVIDIA's products.

When the AI boom kicked off in late 2022, Nvidia's GPUs and CUDA platform represented the path of least resistance. There are capable AI accelerators from competitors, including Intel (NASDAQ: INTC) and Advanced Micro Devices, and cloud giants like Alphabet are increasingly designing their own custom AI chips. However, the CUDA software platform effectively locks customers into using Nvidia's GPUs.

An industrywide revolt

Intel has been working on oneAPI, an open specification that's at the heart of the company's efforts to kick-start an alternative to CUDA, for the past few years. oneAPI is multi-architecture, meaning that a single code base can be used across CPUs, GPUs, and other types of chips, and multi-vendor, meaning it aims to support chips from any company. This is in stark contrast to CUDA, which is locked to Nvidia's GPUs.

Intel is far from the only company eager to see Nvidia's AI accelerator dominance erode. In late 2023, Intel came together with Arm Holdings, Alphabet's Google, Qualcomm, Samsung, and a few others to form the Unified Acceleration (UXL) Foundation. The core goal of this union is to build an open software ecosystem for all accelerators. Intel's oneAPI is at the center of this effort.

For companies like Google, which buys mountains of Nvidia's GPUs to power AI workloads for itself and its cloud customers, having alternatives to Nvidia could help cut down costs considerably. For companies like Qualcomm and Samsung, breaking CUDA's iron grip on the AI accelerator market makes it more likely that their own products will gain traction.

Samsung has concrete plans to enter the AI accelerator market in early 2025 with the launch of its Mach 1 chips. While Samsung's first effort will be focused on AI inference, not AI training, the software ecosystem will ultimately determine whether the chip is a success.

Meanwhile, Qualcomm is focused on integrating AI into all of its chips, an effort that will require software support to pay off. "We're actually showing developers how you migrate out from an Nvidia platform," said Qualcomm's head of AI and machine learning in an interview with Reuters.

Intel may benefit most of all

Intel is going after the AI market in multiple ways. The company sells its Gaudi line of AI accelerators, and it's been baking AI chips into its recent CPUs.

While the industrywide effort to break CUDA's dominance will make Intel's products more attractive, it's the company's semiconductor foundry business that may benefit the most. Intel plans to overtake foundry leader TSMC technologically in early 2025 with its Intel 18A process, and it has aspirations to become the world's second-largest foundry by 2030. The more competitive the AI accelerator market is, and the more chips that have AI capabilities built-in, the more potential customers for Intel's foundry.

While a more competitive AI chip market may hinder Intel's market share in AI accelerators, the company's foundry opportunity is enormous. Global foundry spending topped $100 billion in 2022, and it's expected to more than double to $231 billion by 2032. In the long run, Intel's foundry could be raking in tens of billions of dollars in revenue annually.

It will take time for the anti-CUDA efforts to gain traction. UXL is aiming to get the specifications for its open alternative to a mature state by the end of the year, an eternity in the fast-moving world of AI. It will take even more time to get developers on board in numbers that are big enough to matter.

But the reward for this effort will be a more competitive AI chip market. For Intel, a hoard of AI accelerators gunning for Nvidia's dominance could help supercharge the company's foundry business in the years ahead.

Should you invest $1,000 in Intel right now?

Before you buy stock in Intel, consider this:

The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Intel wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.

Stock Advisor provides investors with an easy-to-follow blueprint for success, including guidance on building a portfolio, regular updates from analysts, and two new stock picks each month. The Stock Advisor service has more than tripled the return of S&P 500 since 2002*.

*Stock Advisor returns as of March 25, 2024

Suzanne Frey, an executive at Alphabet, is a member of The Motley Fool's board of directors. Timothy Green has positions in Intel. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Nvidia, and Qualcomm. The Motley Fool recommends Intel and recommends the following options: long January 2023 $57.50 calls on Intel, long January 2025 $45 calls on Intel, and short May 2024 $47 calls on Intel. The Motley Fool has a disclosure policy.

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