Roughly three decades ago, the advent of the internet and its mainstream proliferation began changing the business world forever. Though this transformation didn’t occur overnight, investors have been waiting quite some time for the next game-changing innovation to come along and bolster the long-term growth prospects of corporate America. After an extensive wait, artificial intelligence (AI) appears to have answered the call.
Software and systems are being empowered by AI to make decisions, reason, and evolve, all without the aid of human intervention. With use cases in most industries around the globe, it’s perhaps no surprise that the analysts at PwC are forecasting a $15.7 trillion benefit to the global economy from AI by 2030.
No public company has taken the bull by the horns more during the early stages of the AI revolution than Nvidia (NASDAQ: NVDA). The company’s Hopper (H100) and Blackwell graphics processing units (GPUs) have quickly become the standard for high-compute enterprise data centers responsible for training large language models and running generative AI solutions.
But with lofty growth expectations already baked into Nvidia’s nearly $3.2 trillion market cap, there are three very specific reasons to believe Nvidia stock is poised to underwhelm following the release of its fiscal 2025 fourth-quarter operating results after the closing bell on Feb. 26.
To give credit where credit is due, no chip company is particularly close to matching the computing speed of Nvidia’s successor Blackwell chip, or even unseating the Hopper GPU at this point.
However, the company has also benefited immensely from AI-GPU scarcity. Overwhelming demand for Nvidia’s hardware, coupled with limited supply, has helped it book orders well in advance, as well as charge a premium price for its AI-GPUs. Whereas Advanced Micro Devices was netting in the neighborhood of $10,000 to $15,000 for its Instinct MI300X AI-accelerating chips in early 2024, Nvidia was commanding up to $40,000 for its Hopper chip. The result was a gross margin that peaked at 78.4% in the first quarter of fiscal 2025.
The concern for Nvidia on and after Feb. 26 is that computing speed isn’t everything in the AI space. It’s dominated because it’s been able to supply the in-demand hardware businesses are asking for. But as AI-GPU scarcity wanes, so will the company’s pricing power and its supercharged gross margin.
While most investors have focused on direct competitors, such as AMD, the bigger threat is the possibility of losing out on valuable data-center real estate from its top customers by net sales. Most members of the “Magnificent Seven” are internally developing AI chips of their own. Even though these GPUs are unlikely to surpass the computing potential of Nvidia’s Hopper and/or Blackwell chips, they’re substantially cheaper and not backlogged. This is a recipe for future orders of Nvidia’s hardware from America’s most-influential businesses to disappoint.
