The perfect storm may be brewing for Nvidia.
For two years, no trend has created more buzz on Wall Street than the rise of artificial intelligence (AI). The ability for AI-driven software and systems to become more proficient at their assigned tasks, as well as learn new skills over time without the aid of human intervention, gives this technology seemingly limitless long-term potential.
In Sizing the Prize, the analysts at PwC estimated AI would increase worldwide gross domestic product (GDP) 26% ($15.7 trillion) by the turn of the decade. A lift of this magnitude means companies up and down the AI landscape can be winners.
However, no company has more directly benefited from the rise of AI than semiconductor behemoth Nvidia (NVDA -3.26%). In less than two years, Nvidia has grown from a $360 billion business that was somewhat important within the tech sector to Wall Street’s most-valuable publicly traded company ($3.64 trillion market cap).
Given the critical role Nvidia is playing in the AI revolution, Wall Street and investors are laser-focused on Nov. 20, which is when the company will lift its proverbial hood and unveil its operating results for the quarter ending on Oct. 27.
While the optimism surrounding Nvidia is thick enough to cut with a knife, I can offer a half-dozen reasons Nvidia stock will hit a brick wall on Nov. 20.
Nvidia’s operating ramp has been textbook
Before digging into the details of why Nvidia’s stock can struggle following the release of its fiscal third-quarter results, let me provide some background that explains why Nvidia has added $3.3 trillion in market value in under two years.
The heart of Nvidia’s growth is its hardware. Orders for the company’s H100 graphics processing unit (GPU), commonly referred to as the “Hopper,” and the successor Blackwell GPU architecture, are backlogged. Businesses are eager to gain first-mover advantages, and Nvidia’s AI-GPUs offer superior computing capabilities.
In addition to strong demand, Nvidia is commanding stratospheric pricing power for its hardware. Whereas competing AI-GPUs are being priced in the $10,000 to $15,000 range, the Hopper has been consistently commanding a price point of $30,000 to $40,000 per chip. Businesses willingly paying a premium for Nvidia’s solutions pushed its gross margin to as high as 78%.
I’d be remiss if I didn’t also mention the key role Nvidia’s CUDA software platform has played in driving sales growth. CUDA is the toolkit developers use to build large language models and maximize the computing potential of their GPUs. In other words, CUDA is the lure that’s keeping Nvidia’s customers within its umbrella of products and services.
With truly staggering sales growth — $27 billion in fiscal 2023 to an estimated $180 billion in fiscal 2026 — it’s not surprising to see investors piling in.
Yet there’s a half-dozen reasons to believe the perfect storm is brewing for Nvidia, which needs nothing short of flawless execution to sustain its nearly parabolic move higher.
Opinion: The party ends for Nvidia on Nov. 20
When Nvidia delivers its fiscal Q3 operating results in four days, there’s a good chance it’ll top consensus revenue and profit forecasts. Over the last seven quarters, Nvidia has handily outpaced earnings-per-share (EPS) expectations. But simply surpassing what was expected of the company in its October quarter is unlikely to be enough to support additional upside for a variety of reasons.
To start with, external competition has officially arrived. After Nvidia secured an estimated 98% of GPU shipments to data centers in 2022 and 2023, based on a study by TechInsights, it’ll likely cede some of this market share to Advanced Micro Devices this year. AMD is quickly ramping up production of its MI300X AI-GPU, and recently introduced the MI325X, which should go into production before the end of the year. With Nvidia’s GPUs backlogged and businesses eager to gain first-mover advantages, it wouldn’t be surprising to see companies scoop up AMD’s AI hardware instead.
However, a bigger threat to Nvidia might just be internal competition. Many of its top customers by net sales, including Microsoft, Meta Platforms, Amazon, and Alphabet, are internally developing AI chips for use in their data centers. The cost to develop/produce these internal chips, coupled with their ease of access, could make them more favorable than Nvidia’s hardware in future quarters.
To add to this second point, AI-GPU scarcity has played a critical role in lifting Nvidia’s pricing power. As more chips become available and scarcity wanes, Nvidia’s pricing power and gross margin will be negatively impacted.
The third issue to consider is that Nvidia is still constrained by its supply chain. Taiwan Semiconductor Manufacturing, which produces chips for Nvidia, is aiming to boost its chip-on-wafer-on-substrate (CoWoS) monthly production capacity to 80,000 wafers. CoWoS is necessary for the packaging of high-bandwidth memory in AI-accelerated data centers. Being at the mercy of its suppliers means Nvidia isn’t able to meet all of its orders and could cause it to lose valuable data center real estate to external and internal competitors.
Fourth, export regulations offer little reason to cheer. In 2022, U.S. regulators clamped down on Nvidia’s ability to export its AI-GPUs to China, the world’s No. 2 economy by GDP. The following year, Nvidia’s toned-down AI chips, the A800 and H800, which were designed specifically for China, were also added to the export restriction list. With President-Elect Donald Trump promising to get tough on trade with China, Nvidia’s prospects of generating big bucks from exports of AI hardware to China are quickly dimming.
Insider trading activity provides the fifth puzzle piece of the perfect storm for Nov. 20. Over the last 47 months, no executive or director has purchased a single share of Nvidia stock on the open market. While there are plenty of reasons to sell stock, not all of which are nefarious, the only reason to buy shares of a company is if you think they’re headed higher. The actions of Nvidia’s executives and directors make it clear that Nvidia stock isn’t viewed as a good value.
The sixth and final reason I’ll be looking for Nvidia stock to hit a brick wall on Nov. 20 is history. Over the past three decades, investors have consistently overestimated how quickly a new technology or innovation will gain utility and/or adoption. This leads to lofty growth expectations that eventually (key word!) come up short.
Despite ample demand for the Hopper and Blackwell, what’s sorely lacking from Nvidia’s customers are well-defined plans to monetize their AI investments. In other words, businesses are buying, but there’s no clear use case. All technologies need time to mature, and artificial intelligence is unlikely to be the exception to the unwritten rule. In my view, this puts Nvidia stock in a precarious position come Nov. 20.