New York, February 26, 2026, 10:15 EST — Regular session
- NVDA drops roughly 3.6% at the open, even after topping earnings estimates and projecting strong revenue.
- Nvidia’s forecast leaves out any China data-center chip sales, though the company noted it holds U.S. licenses allowing shipments of “small amounts” of H200 chips.
- All eyes turn to Nvidia’s March GTC keynote, where traders hope for sharper signals on what’s coming next for demand.
Nvidia dropped 3.6% to $188.42 Thursday morning, pulling back from Wednesday’s $195.56 close despite topping estimates. Shares bounced between $188.41 and $194.21 in early action, according to Investing.com data. (Investing)
Traders kept things subdued at the open—S&P 500 and Nasdaq barely budged as the market took stock of AI risks. With big tech steering the indexes lately, Nvidia’s results remain the reference point, regardless of how quiet the broader action gets. (Reuters)
Traders had set up for a muted move this time. Options pricing was flagging the smallest post-earnings swing for Nvidia in three years. Susquehanna’s Chris Murphy noted, “the element of surprise has diminished” with more nuanced coverage and positioning. (Reuters)
Nvidia delivered January-quarter revenue of $68.13 billion and adjusted earnings per share of $1.62—both topping Wall Street’s expectations. The company now projects $78 billion in revenue this quarter, give or take 2%, based on LSEG figures. The forecast factors in zero data-center chip sales to China, Nvidia said, though it did secure U.S. licenses this month allowing “small amounts” of H200 chips to go to Chinese customers. Two customers accounted for 36% of fiscal 2026 sales as hyperscalers like Meta plan to spend at least $630 billion on capital expenses in 2026. With hyperscale firms ramping up in-house chip development and competitors like AMD and Google stepping up, CFO Colette Kress says Nvidia will keep funding the AI ecosystem. CEO Jensen Huang insists, “This new way of doing computing is not going to go back.” Ken Mahoney at Mahoney Asset Management called much of the beat “baked in to the cake.” (Reuters)
Nvidia reported a record $62.3 billion in data center revenue for the quarter, pushing full-year revenue up to $215.9 billion. The company handed $41.1 billion back to shareholders during fiscal 2026, leaving $58.5 billion still available under its current buyback authorization. Nvidia announced a $0.01 quarterly dividend, set for payment on April 1 to shareholders of record as of March 11. Beginning in the first quarter of fiscal 2027, Nvidia plans to fold stock-based compensation—employee pay via shares and options—into its non-GAAP results. Management also flagged the Vera Rubin platform as the company’s next major move. (SEC)
On Wall Street, those major cloud buyers are dubbed “hyperscalers”—companies with the muscle to pour tens of billions into data centers annually. Nvidia’s fortunes are closely hitched to how much they’re willing to spend. If their budgets get shaky, the shares often react immediately, questions coming after.
The flip side hasn’t disappeared. Customer concentration carries risks, and export regulations keep China unpredictable. Plus, if AI expansion loses steam, major buyers rolling out their own chips could undercut Nvidia’s pricing strength.
The strain was evident from the start. Nvidia shares ticked up premarket, then reversed as the session wore on—traders grumbled that the earnings call lacked fresh guidance on revenue and dismissed chatter about a potential cash return. Deutsche Bank’s Jim Reid labeled the move “perhaps a sign of investors’ increased anxiety over AI valuations.” (Reuters)
Investors are eyeing a few things right now: whether China licensing shifts, if spending by customers gains momentum heading into mid-year, and if supply plus product timing remain constrained enough to sustain the run. The next major hurdle for the stock may hinge less on the last quarter’s numbers and more on how much fresh disclosure Nvidia offers.
Nvidia’s GTC AI conference lands in San Jose from March 16 to 19. Huang is set for the keynote on the 16th. (Nvidia)
