For more than two years, no trend has captivated the attention of investors quite like the evolution of artificial intelligence (AI). The ability for software and systems to reason, act, and potentially even evolve to learn new skills, all without the aid of human intervention, gives this technology a seemingly limitless ceiling.
According to PwC's Sizing the Prize, AI can add $15.7 trillion to the global economy by 2030 through a combination of productivity improvements and various consumption-side effects. This is a massive addressable market, with no company benefiting more in the early going than semiconductor colossus Nvidia (NVDA +1.63%).
In short order, Nvidia's Hopper (H100) graphics processing units (GPUs) and Blackwell GPU architecture grabbed hold of a monopoly like share of the AI-GPUs used in high-compute data centers. Overwhelming demand for the company's hardware, coupled with AI-GPU scarcity, led to exceptional pricing power and a generally accepted accounting principles (GAAP) gross margin that hit 78.4% during the fiscal first quarter of 2025 (ended April 28, 2024).
But over the last week, Wall Street's optimism for its AI darling has noticeably soured -- and, in my view, this is likely just the beginning.
Image source: Getty Images.
More than a half-dozen Wall Street analysts cut their Nvidia price target last week
While commentary from Wall Street analysts is often fluid, their price targets act as anchors that help investors gauge their longer-term sentiment toward a company. Last week, eight Wall Street analysts cut their price target on Nvidia:
- Gil Luria of D.A. Davidson lowered his firm's price target from $125 to $120.
- Vivek Arya of Bank of America Securities lowered his firm's price target from $200 to $160.
- Harsh Kumar of Piper Sandler lowered his firm's price target from $175 to $150.
- Srini Pajjuri of Raymond James lowered his firm's price target from $170 to $150.
- Timothy Arcuri of UBS lowered his firm's price target from $185 to $180.
- Kevin Cassidy of Rosenblatt lowered his firm's price target from $220 to $200.
- Chris Caso of Wolfe Research lowered his firm's price target from $180 to $150.
- Jim Kelleher of Argus Research lowered his firm's price target from $175 to $150.
A majority of these price target adjustments occurred after Nvidia released a regulatory filing on April 15 where it disclosed a potential hit of up to $5.5 billion tied to exports of its high-powered H20 chips to China. Nvidia noted it would need to obtain a special license to export the H20 chip to the world's second-largest economy by gross domestic product.
However, this filing shouldn't be a surprise. For three years prior to Donald Trump taking office the Joe Biden administration clamped down on the export of AI-accelerating chips and equipment to China. The Biden administration added the A800 and H800 chips to the restricted export list, which are AI-GPUs Nvidia had developed specifically for the Chinese market.
While some investors might be looking at this slip-up as the ideal opportunity to snag shares of Wall Street's Ai darling on the cheap, there's a realistic possibility we're witnessing the beginning of a wave of analyst price target cuts for Nvidia.
Nvidia's GAAP gross margin tells the tale
Although Wall Street analysts tend to be reactive rather than proactive most of the time, the telltale figure that strongly suggests price target cuts are going to become something of the norm for Nvidia is its GAAP gross margin.
NVDA Gross Profit Margin (Quarterly) data by YCharts.
As noted, a perfect storm of demand and limited supply allowed Nvidia to charge north of $40,000 for its Hopper chip early last year. This resulted in the aforementioned GAAP gross margin of 78.4%. But in each subsequent quarter, Nvidia's GAAP gross margin has declined:
- Q1 2025: 78.4%
- Q2 2025: 75.1%
- Q3 2025: 74.6%
- Q4 2025: 73%
- Q1 2026 (est.): 70.6% (+/- 50 basis points)
This decline reflects both an increase in competition within AI-accelerated data centers, as well as a decrease in AI-GPU scarcity.
While most investors remain laser-focused on direct competitors to Nvidia's data-center dominance, its biggest threat has come from within. Most members of the "Magnificent Seven," which happen to be some of Nvidia's top customers by net sales, are internally developing AI-GPUs and solutions of their own. Even though these chips won't be sold externally, and they're likely to be slower than Nvidia's hardware on a compute basis, they're considerably cheaper than Nvidia's AI-GPUs, and there's no backlog.
Nvidia is poised to lose data center real estate to these internally developed chips with each passing quarter. In turn, I expect this to steadily minimize AI-GPU scarcity and reduce the pricing power that's fueled the parabolic climb of Nvidia stock.
Image source: Getty Images.
Nvidia's product cycle is a problem
A second issue that's likely to cause Wall Street's price targets for Nvidia to move lower is its product replacement cycle.
Normally, innovation is everything in the tech sector. Nvidia's Hopper and Blackwell chips are the clear leaders in compute speed in AI-accelerated data centers. It's why we've witnessed Nvidia's full-year sales skyrocket from $27 billion in fiscal 2023 to north of $130 billion in fiscal 2025.
However, Nvidia's expediency of innovation, and its desire to maintain its lead in compute capacity, may not sit well with its current or prospective customers.
After introducing the Hopper, Nvidia began rolling out Blackwell late last (calendar) year. Next in line is the Vera Rubin GPU architecture, which is expected to debut in 2026, followed by the Vera Rubin Ultra in the second-half of 2027. Both Vera Rubin chips will feature the all-new Vera processor.
But bringing a new AI-GPU to market annually comes with a cost. Previous buyers are likely to see the value of their AI hardware investments rapidly depreciate.
Additionally, the improvements observed in compute capacity, high-bandwidth memory, and energy efficiency from one year to the next is likely to diminish over time. In other words, there's less of an incentive for Nvidia's clients -- especially its biggest customers -- to invest aggressively in the latest hardware.
History is not in Nvidia's corner
Last but certainly not least, history is likely to drag Wall Street's price targets on the company considerably lower.
NASDAQ: NVDA
Key Data Points
For much of the last 30-plus years, investors have had a next-big-thing technology or innovation to captivate their attention. Though each of these trends offered juicy addressable markets, the one factor they've all shared was an eventual bubble-bursting event. Beginning with the internet in the mid-1990s, every next-big-thing trend has navigated its way through a bubble.
The reason bubbles form is simple: investors overestimate the early stage adoption and/or utility of a new technology or innovation. At some point, lofty investor expectations fail to be met, which leads to the bubble bursting.
Although Nvidia has enjoyed stellar demand for its AI hardware, what can't be ignored is that most businesses investing in AI have yet to optimize their solutions or generate a positive return on their investments. We're witnessing all the same telltale signs that investors have overestimated the early adoption and real-world use cases of artificial intelligence.
The good news is that many of these game-changing trends went on to be successful. But the undeniable reality is that every next-big-thing technology/innovation needs plenty of time to mature. AI isn't anywhere close to being a mature technology, as of yet.
Considering that Nvidia has benefited more than any other public company from the rise of artificial intelligence, an eventual bubble-bursting event would, in all likelihood, hit it harder than other AI stocks.
