Unpopular Opinion: Jensen Huang Is Making Nvidia Its Own Worst Enemy

In the mid-1990s, the advent and proliferation of the internet completely changed the game for businesses by opening sales and marketing channels that hadn't previously existed. For decades, investors have been waiting for the next technological leap forward that would rival what the internet did for corporate America -- and artificial intelligence (AI) looks to be the answer.

Although a laundry list of publicly traded companies has benefited from the AI revolution, none has increased in value quite like Nvidia (NASDAQ: NVDA). Since the beginning of 2023, Nvidia has tacked on $3.9 trillion in market value.

But while multiple aspects of Nvidia's operating model have hit home with investors, there's also the possibility that CEO Jensen Huang is turning his company into its own worst enemy.

Image source: Nvidia.

Nvidia's competitive advantages are undeniable

Nvidia's ascent to Wall Street's most valuable public company reflects its many competitive advantages. This all begins with its famed graphics processing units (GPUs).

GPUs are the brains of AI data centers, responsible for fueling accurate, split-second decision-making. By some analyst estimates, Nvidia's GPUs account for 90% (or more) of those currently deployed in AI-accelerated data centers.

The popularity of this hardware stems from its superior compute capabilities. Beginning with Hopper (H100) and extending through successor GPUs, Blackwell and Blackwell Ultra, no external competitors have come particularly close to outperforming Nvidia's hardware on a compute basis. All three generations of GPUs have contended with order backlogs, leading to exceptional pricing power and a mid-70% gross margin for the world's most valuable public company.

Building on this point, Jensen Huang has not been shy about depressing the proverbial accelerator and solidifying his company's place atop the compute pedestal. Nvidia is aggressively investing in research and development with the goal of bringing an advanced GPU to market annually. The Vera Rubin chip, powered by an all-new processor, is set to make its debut later this year as the successor to Blackwell Ultra.

Tying everything together is Nvidia's CUDA software platform, which never receives enough credit. CUDA is the toolkit developers use to maximize the potential of their Nvidia GPUs, including the building and training of large language models. CUDA acts like an anchor that keeps customers loyal to Nvidia's array of products and services.

With Nvidia's customer list looking like a who's who of the most influential tech companies, it's really no surprise that it's become Wall Street's most consequential public company.

Is CEO Jensen Huang setting Nvidia up for eventual disaster?

But not even the stock market's most successful companies are free of headwinds. For instance, the historical precedent of next-big-thing technologies enduring bubble-bursting events early in their expansion is one of several reasons for investors to be skeptical of Wall Street's largest publicly traded company.

However, it's a decision by CEO Jensen Huang that could hamstring his company in the years to come.

As noted, Huang is overseeing a rapid GPU innovation cycle. The purpose of introducing an advanced chip annually is to maintain Nvidia's compute advantage, as well as charge a significant price premium over external competitors. The desire of businesses to spend freely to ramp their AI infrastructure, coupled with world-leading chip fabricator Taiwan Semiconductor Manufacturing being unable to keep pace with insatiable GPU demand, has favored Huang's approach... thus far.

Image source: Getty Images.

But with every new generation of GPU that Nvidia debuts, it risks rapidly depreciating the value of prior-gen chips. Understandably, businesses expect their hardware investments to lose value over time. However, Nvidia's aggressive innovation timeline can potentially accelerate the depreciation of prior-gen chips, such as Hopper and Blackwell.

The depreciation of older chips poses several challenges. Most companies expect these GPUs to be worth "X" five or six years from now when they consider upgrading their hardware. If Huang's aggressive innovation cycle causes these chips to be worth, let's say, 30% to 50% of "X," businesses are likely to delay a planned upgrade.

Ongoing software advances for CUDA have also sustained exceptionally high enterprise utility for the company's Hopper chip. If improvements to CUDA continue to allow Hopper and other prior-gen chips to outperform their initially designed utility, there's also less of an incentive to upgrade to the newest GPUs.

Another factor to consider is that as the aggregate number of available GPUs expands and Taiwan Semiconductor Manufacturing increases its monthly chip-on-wafer-on-substrate capacity, the GPU scarcity factor that's fueled Nvidia's premium pricing power is going to slowly fade.

While there's certainly logic in maintaining Nvidia's compute advantage and pricing power, the inevitable end of GPU scarcity, coupled with the ongoing depreciation of older GPUs, would be expected to adversely impact demand for the company's newest chips in the coming years. Though Huang's ambitious goal for Nvidia is well-intentioned, he may end up inadvertently hurting his company and its shareholders.

Sean Williams has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Nvidia and Taiwan Semiconductor Manufacturing. The Motley Fool has a disclosure policy.