Why This AI Stock Could Outperform Nvidia Over the Next 5 Years

Nvidia (NASDAQ: NVDA) has led the artificial intelligence (AI) investing sector since 2023. While 2025 hasn't been quite as good a year for its shareholders as the previous two were, the stock is still up around 30% year to date. The chipmaker is still a dominant force in the AI realm, but observers are raising the possibility that its competitors could start taking meaningful market share away from it.

Should Nvidia start to lose its edge over the competition, it may no longer be the best AI stock to own. But over the next five years, which stock could overtake it?

Image source: Getty Images.

Different approaches to the AI computing problem

Nvidia makes graphics processing units (GPUs), which are computing units that can process many thousands of calculations in parallel. This makes them perfectly suited for tasks that require a lot of computing power and that can be broken down into many smaller tasks that can be handled independently. Most artificial intelligence workloads fall into that category. Nvidia's GPUs have been the most popular parallel processors available since before the AI infrastructure buildout started in 2023, but that may change.

One issue Nvidia is having right now is that demand is outpacing its supply. In its earnings release for its fiscal 2026 third quarter (which ended Oct. 26), CEO Jensen Huang noted that the company was "sold out" of cloud GPUs. While some may view this as a good problem to have, I think it's a bit of an issue. Its clients' needs for AI accelerator chips aren't going to decrease just because Nvidia can't fully supply them. Instead, they are going to look for alternative solutions.

That could open up the door for competing chipmakers to steal some significant market share, particularly if buyers find that the alternative products provide similar performance with lower costs. Two chipmakers that appear poised to benefit are AMD (NASDAQ: AMD) and Broadcom (NASDAQ: AVGO).

AMD is a direct competitor to Nvidia, as it also makes GPUs. It recently forecast that it would grow its data center revenue at a 60% compound annual rate over the next five years. For reference, its data center division's revenue growth was a mere 22% during Q3. For AMD to grow as fast as predicted, it will need to eat into Nvidia's market share, which it already might be doing.

Broadcom is taking a different approach to the AI computing problem. Rather than designing another general-purpose GPU, it's partnering directly with hyperscalers to design custom AI computing units -- called application-specific integrated circuits -- that meet their precise needs. Among the best examples of these chips are Alphabet's Tensor Processing Units (TPUs).

TPUs and other custom AI computing units aren't designed to handle a wide variety of workloads like GPUs are. However, when these computing units only see the one type of workload that they are configured for, they can deliver higher performance at lower costs. The trade-off of lower price and higher performance for less flexibility is a no-brainer for many data center operators, and Broadcom's approach could eat into Nvidia's market share as AI workloads become more streamlined.

These companies have the potential to steal some of Nvidia's market share over the next five years, but all the chip designers all have one thing in common.

Taiwan Semiconductor will rise regardless of which chip designers are winning

None of these companies fabricates its computing units in-house. Instead, the "fabless" chipmakers outsource that work to a handful of other companies that specialize in fabrication. In the chip production realm, Taiwan Semiconductor Manufacturing (NYSE: TSM) captures nearly all of the highest-end work. It's not only the largest chip manufacturer by revenue in the world, but also a leader in its ability to deliver the most advanced chips. Broadcom, AMD, and Nvidia often work with Taiwan Semiconductor for their cutting-edge chips, and that's unlikely to change in the future, as that industry is nearly impossible to disrupt.

With Nvidia projecting between $3 trillion and $4 trillion in data center infrastructure spending by 2030, all three of these fabless chip companies could have a significant market opportunity to capture. I'm not sure which will be the ultimate winner among them. Still, I know that TSMC will be a winner regardless, giving it a strong chance of outperforming Nvidia over the next five years, particularly if Nvidia loses a meaningful amount of market share to its peers.

Keithen Drury has positions in Alphabet, Broadcom, Nvidia, and Taiwan Semiconductor Manufacturing. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Nvidia, and Taiwan Semiconductor Manufacturing. The Motley Fool recommends Broadcom. The Motley Fool has a disclosure policy.