This article was originally published on Fool.com. All figures quoted in US dollars unless otherwise stated.
Roughly 30 years ago, the advent of the internet changed the growth trajectory for businesses across the globe. Although it took a few years for the internet to mature as a technology and for businesses to fully understand how to harness its potential, it's had a notably positive impact on long-term growth trends.
Since the mid-1990s, Wall Street has been waiting patiently for the next leap forward for corporate America. Over the last two years, artificial intelligence (AI) appears to have answered the call.
AI-driven software and systems have the ability to become more proficient at their assigned tasks, as well as learn new skill sets without human intervention. This capacity to learn and evolve over time is what gives this technology seemingly limitless potential and utility in most industries around the globe.
While the AI ecosystem is vast, which should allow numerous businesses to thrive, no company been a more direct beneficiary of the rise of AI than cutting-edge semiconductor stock Nvidia (NASDAQ: NVDA). Since 2023 began, Nvidia's market cap skyrocketed from $360 billion to north of $3.6 trillion, which makes it the largest publicly traded company, as of this writing.
Nvidia's operating expansion has been virtually textbook
Less than two years ago, when Nvidia lifted the hood on fiscal 2023 (Nvidia's fiscal year ends in late January), the company reported $27 billion in full-year sales. In the current fiscal year (2025), it's pacing closer to $129 billion in full-year revenue, with Wall Street calling for almost $192 billion in sales next year.
This otherworldly growth is a function of Nvidia's AI-graphics processing units (GPUs) being the preferred choice for businesses running high-compute data centres. The analysts at TechInsights pegged Nvidia's share of GPU shipments to data centres at 98% in 2022 and 2023. Based on the company's two-year sales ramp, it'd be a fair assumption that Nvidia's H100 GPU (commonly known as the "Hopper") and successor Blackwell GPU architecture aren't having any issues finding buyers.
Nvidia has also been able to take advantage of the law of supply and demand. With orders for the Hopper and next-generation Blackwell chip backlogged, it's been able to meaningfully increase the price for its hardware. The roughly $30,000 to $40,000 price tag for the Hopper represents a 100% to 300% premium to what Advanced Micro Devices is netting for its MI300X chips for AI-accelerated data centres.
Credit also needs to be given to Nvidia's CUDA software platform for its virtually textbook operating expansion. CUDA is the toolkit developers use to maximise the potential of their Nvidia GPUs, including building large language models. This platform has helped keep clients loyal to Nvidia's umbrella of products and services.
Although everything seems to be going perfectly for Nvidia, as its stock performance would suggest, one metric in the company's recently released third-quarter operating results (for the quarter ended Oct. 27) tells a different story.
This lone metric strongly suggests Nvidia stock has hit its peak
As expected, Nvidia's headline figures look as good as advertised. Quarterly sales jumped 94% from the year-ago quarter to $35.08 billion, while net income surged 109% to $19.3 billion. Both were well ahead of Wall Street's consensus forecast.
But there is one exceptionally important figure that's showing signs of weakness, and it foreshadows the very real possibility of the top being in for Nvidia stock.
When Nvidia lifted the hood on its first-quarter operating results in May, it reported a scorching-hot gross margin of 78.4%. The dramatic increase the company has enjoyed in its gross margin is a function of AI-GPU scarcity and its aforementioned exceptional pricing power.
NVDA Gross Profit Margin (Quarterly) data by YCharts.
The above chart isn't yet reflective of Nvidia's fiscal third-quarter gross margin of 74.6%.
However, the tide is turning. After delivering a 78.4% gross margin in Q1 2025, Nvidia reported a gross margin of 75.1% in Q2 2025 and 74.6% in the latest quarter. For the fiscal fourth quarter, it's forecasting a gross margin of 73% to 73.5%, +/- 50 basis points.
Although Nvidia's gross margin is notably higher now than it was prior to the AI revolution taking shape, this steady decline we're witnessing is evidence that AI-GPU scarcity is waning and competition is picking up.
Most of Wall Street is focused on the external competition Nvidia will have to contend with. AMD has been steadily ramping up production of its MI300X AI-GPUs and recently unveiled its next-gen MI325X chip, which it intends to put into production before the end of the year. AMD is a brand-name company with a rich history and a considerably cheaper AI-GPU than Nvidia's Hopper and Blackwell chips. Businesses wanting to gain first-mover advantages may be compelled to skip the potentially long wait for Nvidia's hardware and choose AMD.
But the bigger issue for Nvidia may very well be the competition it's facing from within. Many of the company's top customers by net sales are members of the "Magnificent Seven," and they're all internally developing AI-GPUs to deploy in their data centres. Even if these chips come up short to Nvidia's hardware in terms of computing capabilities, they're still markedly cheaper and more easily accessible.
Anything that reduces AI-GPU scarcity is going to adversely impact Nvidia's pricing power and its gross margin.
History would like a word, too
Based on the trend we're seeing from Nvidia's gross margin, its shares have likely topped. But history would like a word, as well.
Even though the internet transformed the business world three decades ago, it took time for the technology to mature and for businesses to properly utilise it to generate a positive return on their investment. Every next-big-thing innovation for 30 years, including the internet, has worked its way through an early innings bubble-bursting event.
What this tells us is that investors consistently overestimate how quickly a new technology will be adopted by consumers and/or businesses on a broad basis. It also suggests investors are overly optimistic about the utility of cutting-edge technologies. Though artificial intelligence can be every bit the game-changer that Wall Street expects it to be, it's going to take time for businesses to figure out how best to utilise AI. This is what leads to lofty expectations eventually falling short.
Investors are seeing this dynamic play out right in front of their eyes. Even with sizable investments in AI data centres from some of Wall Street's most-prominent companies, most of these businesses lack a clear game plan to generate a positive return on their AI investments any time soon. The utility aspect of AI isn't well understood at the moment -- and that's a scary thing for a company whose gross margin is in a decisive decline.
This article was originally published on Fool.com. All figures quoted in US dollars unless otherwise stated.
