Nvidia (NASDAQ: NVDA) has posted a solid but unspectacular 17.3% gain so far in 2026, a far cry from the doubling and tripling seen in earlier AI boom years.
The stock’s slower pace reflects growing competition from custom silicon built by major cloud players, plus questions about whether hyperscaler capital spending can remain this aggressive.
Investors also appear less willing to pay the premium multiples that defined Nvidia’s rocket-ship phase, when each quarterly report felt like the opening of a new and exciting chapter.
Despite the quieter price action, the fundamentals beneath the surface look considerably more compelling than the share price might suggest at first glance.
Nvidia recently issued a full-year-ahead revenue forecast for the first time, projecting approximately 70% revenue growth for fiscal 2028, putting it on a path toward $700 billion in annual sales.
CEO Jensen Huang was direct about the reasoning, noting that customer demand actually implies higher growth than 70%, but supply chain limitations are the binding constraint, not a lack of orders.
Memory shortages and component bottlenecks are limiting how quickly Nvidia can build out AI infrastructure, meaning the 70% target is a conservative, supply-limited figure rather than a demand ceiling.
Nvidia’s ACIE segment, covering AI clouds, industrial customers, and sovereign enterprises, delivered 25% sequential growth in the second quarter, outpacing the 13% growth from hyperscale cloud providers.
The ACIE segment now represents roughly half of data center revenue, giving Nvidia a broader and less concentrated customer base than it had during the early stages of the AI buildout.
Capex across the top five hyperscalers, including Amazon, Alphabet, Microsoft, Meta Platforms, and Oracle, is estimated to reach around $800 billion this calendar year and $1.3 trillion in 2027.
Nvidia’s next-generation Vera Rubin platform is expected to capture an even larger share of that spending as big tech deploys its rapidly expanding infrastructure budgets over the coming year.
New demand drivers are also emerging, with agentic AI workloads requiring significantly more inference and reasoning horsepower, while robotaxi fleets from Alphabet and Tesla move from pilot programs to scaled deployment.
Orbital data centers and on-orbit inference represent additional incremental demand that could push Nvidia’s actual performance past its own official guidance as these applications scale.
Wall Street’s consensus earnings per share estimate for Nvidia’s fiscal 2028, which ends in January of that year, sits at $15.52, implying 67% growth versus the fiscal 2027 EPS target of $9.31.
Applying a 25x forward price-to-earnings multiple to that estimate, which is roughly where Nvidia trades today, produces a share price target of approximately $388 from the current price of $225.
If Nvidia’s multiple expands back toward its historical forward P/E average of 33 during the AI era, shares could be worth closer to $512, implying gains of between 72% and 127% from current levels.
The central insight here is that Nvidia’s earnings power alone is large enough to produce a dramatically different stock price, even without any multiple expansion at all.
