Nvidia (NVDA +3.03%) shares are up 1,390% since the artificial intelligence boom began in January 2023, and Wall Street still thinks the stock is undervalued. Among 65 analysts, the median target price is $300 per share, implying 37% upside from the current share price of $218.Nvidia shareholders recently got good news from Wall Street. Consensus earnings estimates have recently been revised higher, such that analysts now expect earnings to increase at 44% annually over the next three years. In March, the consensus estimate said earnings would increase at 33% annually over that period.
What changed? Wall Street analysts once again underestimated how much money hyperscalers would spend on AI infrastructure. Here are the important details.
Image source: Getty Images.
Nvidia dominates the market for AI infrastructure across GPUs, CPUs, and networking equipment
Nvidia is a full-stack accelerated computing company that develops graphics processing units (GPUs), central processing units (CPUs), and networking equipment, supported by a robust ecosystem of software tools. That approach lets the company optimize performance and power efficiency in ways most competitors cannot, which explains why Nvidia systems are the gold standard in artificial intelligence.
Most readers probably know that Nvidia GPUs account for a large percentage of data center accelerator sales (around 90%, according to HPC Wire). But readers may be less familiar with the company’s prowess in other categories. Nvidia recently became the largest networking company in the world, and it’s on pace to become the largest CPU supplier by the end of this year.
Of course, there’s been a lot of talk about application-specific integrated circuits (ASICs), chips purpose-built for specific workloads like artificial intelligence. Some investors are worried that custom silicon will eventually displace Nvidia. But those fears are unwarranted. ASICs perform certain tasks more cheaply than Nvidia GPUs, but they are less flexible and lack the robust software development ecosystem that backs Nvidia chips.
“Nvidia isn’t going anywhere anytime soon,” according to Meera Pandit, global market strategist at J.P. Morgan. “Only Nvidia chips can handle any AI workload. Custom hardware is a safe and efficient bet for known workloads like inference, but there’s an obsolescence risk as AI evolved.”
Today’s Change(3.03%) $6.59Current Price$224.09Key Data PointsMarket Cap$5.4TMarket cap calculated using publicly traded shares outstanding only. Does not include unlisted, private, or dual-class non-traded shares. Implied market cap may vary.Day’s Range$220.20 – $225.1052wk Range$164.07 – $236.54Volume396.5KAvg Vol149MGross Margin74.15%Dividend Yield0.12%
Wall Street raised its hyperscaler capital expenditure (capex) spending forecast for 2026
Currently, 26% of hyperscaler capital expenditures (capex) go straight to Nvidia’s bottom line, according to research from J.P. Morgan. That astonishing metric underscores the essential role Nvidia plays in the AI infrastructure market. And assuming the company maintains its pricing power and market share, earnings growth should more or less match capex growth going forward.
Here’s the good news for shareholders: Wall Street has consistently underestimated how much hyperscalers will spend on AI infrastructure. “At the start of both 2024 and 2025, consensus estimates implied capex growth of roughly 20% for the year,” writes Goldman Sachs. “In reality, it exceeded 50% in both years.”
The same thing happened in 2026. Last June, the consensus estimate said capex spending among the five largest hyperscalers — Alphabet, Amazon, Meta Platforms, Microsoft, and Oracle — would total $361 billion this year. But Wall Street has since raised its forecast by over 100%, such that the consensus estimate now says their capex spending will total $733 billion in 2026.
Similarly, investors have reason to think Wall Street is making the same mistake with 2027. The consensus estimate currently says capex spending among the top five hyperscalers will grow 28% to $939 billion next year. But that would be a major slowdown compared to capex growth of 56% in 2024, 73% in 2025, and the projected capex growth of 90% in 2026.
Here’s the big picture: Wall Street currently expects capex spending among the five largest hyperscalers to grow at 41% annually through 2028. Meanwhile, the consensus estimate says Nvidia’s earnings will increase at 44% annually over the same period. It makes sense that those figures are roughly equivalent.
However, if analysts are underestimating hyperscaler capex, which is plausible given their track record, it stands to reason that they are also underestimating Nvidia’s future earnings. And if earnings grow faster than expected over the next few years, the efficient market hypothesis predicts the stock price will rise. That makes Nvidia a worthwhile long-term investment.