Nvidia Stock Price Targets Pop After $78B Forecast as RBC, Truist Lift Views

February 26, 2026
Nvidia Stock Price Targets Pop After $78B Forecast as RBC, Truist Lift Views

NEW YORK, Feb 26, 2026, 10:22 (EST)

  • RBC bumped its Nvidia price target up to $250, previously $240. Truist also raised its target, moving to $283 from $275.
  • Nvidia put its first-quarter revenue estimate at roughly $78 billion, making clear the projection does not include any data-center compute revenue from China.
  • HSBC’s Frank Lee stuck with his Buy call, but he cut the target to $310, citing valuation and the growing presence of custom AI chips.

Thursday saw RBC Capital and Truist Financial bump up their price targets on Nvidia. The AI chip giant had just delivered another record-setting quarter, along with a revenue forecast that outpaced most expectations.

Nvidia’s role as the market’s go-to gauge for data-center appetite has only grown, thanks to heavy GPU spending by major cloud players looking to power AI models. Custom chips, though, are grabbing attention, with investors scrutinizing whether they could deliver similar results for less money—and those revisions are now squarely in focus.

Nvidia shares slipped roughly 3% out of the gate. According to RBC, the market’s lukewarm response hints at rising concerns that revenue growth could be topping out, despite Nvidia’s upbeat guidance. 1

Nvidia expects first-quarter revenue to hit $78.0 billion, give or take 2%, according to guidance late Wednesday. The company made a point of excluding any Data Center compute revenue from China in its forecast. “Computing demand is growing exponentially — the agentic AI inflection point has arrived,” CEO Jensen Huang said, alluding to AI capable of handling tasks with less human involvement. 2

RBC bumped up its target to $250 and stuck with an Outperform rating after results and guidance topped not just its own forecasts but the Street’s. According to the bank, visibility now reaches further out to 2027. The next-gen “Rubin” platform remains on schedule, and margins are staying resilient even with pricier memory.

Several brokers have lifted their targets too. Investing.com noted Rosenblatt and Bernstein both now see $300 as the new mark. Wolfe Research kept its Outperform call, stuck with a $275 target, and flagged Nvidia’s gross margin forecast—about 75%. The firm also highlighted that the current guidance doesn’t factor in any China data-center revenue.

Truist bumped its price target up to $283 from $275 and stuck with its buy call, according to MarketBeat, signaling roughly 45% upside based on where the shares were trading at the time. MarketBeat figures showed Wall Street’s average target sitting at $271.32, with the bulk of analysts still rating the stock as a buy. 3

HSBC’s Frank Lee isn’t wavering on his Buy call, though he trimmed his price target for the stock to $310 from the prior $320, according to TipRanks. The adjustment came after Lee reduced the P/E multiple in his valuation model. “We believe that the demand for GPUs remains intact and expect GPUs to account for the bulk of hyperscalers’ capex,” Lee said, referring to the largest cloud data-center players and their capital spending. 4

Lee dismissed claims that ASICs—these are specialized chips—are making Nvidia’s GPUs obsolete just yet. He pointed to Google’s Gemini 3 model, which was trained on its own TPU chip, as an example. But Lee also highlighted Nvidia’s move into server CPUs, noting what he called an agreement with Meta for a sizable deployment of Nvidia’s Grace CPUs. The “Vera” chips, he added, are expected on deck after that.

But those new targets stack up against already lofty expectations. RBC, for one, is still working from a base case where hyperscaler spending eases off only gradually. Both RBC and HSBC flagged persistent uncertainty around exports to China; stricter regulations or a faster drop in spending could quickly alter the growth outlook. MarketBeat, for its part, highlighted risks tied to valuation, insider selling, and competition, noting the possibility that some AI workloads could move to in-house silicon instead of Nvidia’s chips.