Over the last month, Wall Street has sent a stern warning to the investing community that stocks can, in fact, move lower. It took just 16 trading sessions for the benchmark S&P 500 to dip into correction territory -- a decline of at least 10% from a recent closing high -- with the Nasdaq Composite falling into correction at an even faster pace.
While stock market corrections tend to weigh on most sectors and industries, few of Wall Street's most-influential businesses have been hit harder than artificial intelligence (AI) juggernaut Nvidia (NVDA 1.07%), which is now 23% below its record-closing high of $149.43 (achieved on Jan. 6).
Image source: Getty Images.
There's no denying that Nvidia's graphics processing units (GPUs) have played a central role in fueling the AI revolution. Its Hopper (H100) GPU and next-generation Blackwell GPU architecture are the brains responsible for split-second decision making, training large language models, and supporting generative AI solutions.
Most investors, along with an overwhelming majority of Wall Street analysts, view this 23% dip in Nvidia stock as a surefire buying opportunity -- but I'm not among them. What follows are a half-dozen reasons I'm keeping my distance from Wall Street's AI darling.
1. Internal competition is picking up
To begin with, I'm not convinced that Nvidia will be able to hang onto its monopoly like market share in AI-data centers much longer. Even if its Hopper and Blackwell GPUs easily sustain their computing speed superiority, there's more to earning AI-data center real estate than just speed. Price and accessibility also matter.
While most investors are focused on direct competitors to Nvidia ramping up their production (e.g., Advanced Micro Devices), I view internal competition as a much greater threat to its growth potential and bottom line. Many of its top customers by net sales are internally developing AI chips to use in their data centers. Despite their hardware falling short of Blackwell's computing potential, internally developed chips are considerably cheaper and not backlogged like Nvidia's.
Long story short, I fully expect Nvidia to begin losing data center real estate to its own top customers in the coming quarters.
2. Margins tell the tale
This leads to my next point: AI-GPU scarcity has played a critical role in Nvidia's success. Demand overwhelming supply allowed Nvidia to charge upwards of $40,000 per Hopper chip early last year. For the sake of comparison, the cost of AMD's Instinct MI300X AI-accelerating chip clocked in at $10,000 to $15,000 per unit.
NVDA Gross Profit Margin (Quarterly) data by YCharts.
Thanks to this combination of exceptional pricing power and AI-GPU scarcity, Nvidia's gross margin surged to more than 78% in the fiscal first quarter of 2025 (ended April 28, 2024). But in each subsequent quarter, Nvidia's gross margin has declined. The company reported a generally accepted accounting principles (GAAP) gross margin of 73% during the fiscal fourth quarter (ended Jan. 26, 2025), and projected a GAAP gross margin of 70.6% (+/- 50 basis points) for the fiscal first quarter of 2026.
We're witnessing Nvidia's biggest competitive advantage -- AI-GPU scarcity -- disappear before our eyes. As it does, it's going to further erode Nvidia's gross margin.
3. The regulatory landscape isn't friendly
A third reason I'm not itching to own shares of Nvidia is the regulatory environment surrounding artificial intelligence.
From October 2022 through President Joe Biden's departure from the Oval Office two months ago, his administration restricted the export of high-powered AI chips and AI-related equipment to China. Nvidia specifically developed toned-down version of its Hopper chip for the Chinese markets (the A800 and H800), and these GPUs were subsequential added to the restricted export list.
Under President Donald Trump, these restrictions are unlikely to be eased. President Trump favors national security and wants America to be an AI leader. In fiscal 2025, China accounted for approximately 13% of Nvidia's $130.5 billion in total sales. This is roughly $17 billion in yearly revenue that may be at risk for Wall Street's AI darling.
Image source: Getty Images.
4. Tariffs can muck up the works
To add fuel to the fire, President Trump's tariff policy is creating uncertainty.
The president has implemented a 20% cumulative tariff on select imports from China, and has threatened to impose a tariff in the neighborhood of 25% on various semiconductor imports. This has the potential to drive up the infrastructure costs associated with building and/or sustaining an AI-accelerated data center.
The potentially bigger concern is what deteriorating trade relations between the world's top-two economies by gross domestic product might do to Nvidia's sales and bottom line. If President Trump's tariffs draw a steep divide with China, businesses in the world's No. 2 economy may choose to bypass Nvidia altogether and support domestic chipmakers.
5. Next-big-thing investments have a checkered past
Even though Wall Street offers no guarantees, I'm a student of history. This is to say that I'm a strong believer in history rhyming.
Since the internet began proliferating in the mid-1990s, every next-big-thing technology/innovation has eventually navigated its way through a bubble-bursting event. This is a roundabout way of saying that investors persistently overestimate how quickly a new technology will gain utility and be widely adopted by businesses and/or consumers. Ultimately, lofty expectations aren't met, leading to the bursting of a bubble.
NASDAQ: NVDA
Key Data Points
While artificial intelligence has exhibited every sign of being a long-term game-changing technology, it's nowhere close to being a mature trend at the moment. Most businesses lack a well-defined game plan to generate a positive return on their AI investments, and they haven't optimized their existing AI solutions.
If and when the AI bubble bursts, arguably no company will take a more direct hit than Nvidia, which generated north of 88% of its full-year sales in fiscal 2025 from its data center segment.
6. Nvidia's valuation has plenty of room to deflate
The sixth and final reason I'm still not a buyer of Nvidia stock following its 23% tumble from its January closing high is its valuation.
I don't doubt some folks will raise an eyebrow at this statement. Nvidia's rapid profit growth over the last couple of years has the company's stock valued at 20 times forward-year earnings per share (EPS). This is more or less in-line with the forward price-to-earnings (P/E) ratio of the benchmark S&P 500 and well below where things stood a year ago.
NVDA PS Ratio data by YCharts. PS Ratio = price-to-sales ratio.
However, Nvidia is still valued at 22 times sales. Despite being well below its peak price-to-sales (P/S) ratio of 42.39, set in June 2024, many of the "Magnificent Seven" components are valued at P/S ratios of 12 or below.
Historically, P/S ratios in the neighborhood of 30 to 40 have marked the top for businesses leading a next-big-thing trend. This is yet another instance where historic precedent suggests Nvidia stock has much further to fall.
