Rosenblatt Securities raised its price target on Nvidia (NVDA) to $325 from $300 on March 18, maintaining a buy rating, as analyst Kevin Cassidy updated his financial model following management meetings with analysts.

The new target implies roughly 77% upside from the March 16 closing price of $183.22.

This upgrade makes Rosenblatt one of the most bullish voices on Wall Street on Nvidia, heading into what the firm sees as a pivotal year for AI infrastructure spending.

What’s behind Rosenblatt’s new Nvidia target

Cassidy’s revised model now projects over $1 trillion in combined revenue from Nvidia’s Blackwell and Rubin product lines between 2025 and 2027. Rosenblatt added roughly $25 billion in additional revenue expectations for the second half of fiscal 2028, incorporating contributions from Rubin Ultra and Vera standalone products.

The $325 price target reflects a valuation of 25 times fiscal 2028 estimated earnings per share.

The central thesis is that Nvidia’s competitive position extends well beyond its dominant share in AI training. Cassidy argues that the company’s full-stack advantage, spanning CUDA software, NVLink networking, and rack-scale systems, positions it to lead in AI inference as well, which is increasingly where the growth is shifting as more enterprises move from building AI models to actually deploying them at scale.

At the GTC 2026 conference, CEO Jensen Huang signaled $1 trillion in combined orders for Blackwell and Vera Rubin through 2027, reinforcing the multi-year demand backdrop underpinning Rosenblatt’s estimates.

The Nvidia Q4 results that set the stage

Rosenblatt’s upgrade follows a strong earnings report in late February. Nvidia’s Q4 results showed revenue of $68.13 billion, up 73% year over year, beating the Street consensus of roughly $66.2 billion by nearly $2 billion.

Non-GAAP earnings per share came in at $1.62, topping the $1.53 estimate by 9 cents.

Data-center revenue, which now accounts for more than 91% of total sales, grew 75% year over year to $62.3 billion.

More Nvidia:

  • Nvidia stock gets major reality check on ‘$100B’ number
  • Nvidia CEO delivers blunt 7-word rebuttal on software stocks
  • Bank of America resets Nvidia price target after earnings

Purchase commitments surged 90% quarter over quarter to $95.2 billion, which Rosenblatt views as a strong signal of sustained customer demand for next-generation platforms.

First-quarter fiscal 2027 guidance came in at $78 billion, roughly 7% above the Street consensus of $72.6 billion at the time, with 75% gross margins intact. Notably, Nvidia stated that its Q1 guidance does not assume any data center compute revenue from China.

Q4 fiscal 2026 at a glance

  • Revenue: $68.13 billion, up 73% year over year
  • Non-GAAP EPS: $1.62, ahead of $1.53 estimate
  • Data center revenue: $62.3 billion, up 75% year over year
  • Purchase commitments: $95.2 billion, up 90% quarter over quarter
  • Q1 FY2027 guidance: $78 billion, vs. $72.6 billion consensus
Nvidia’s first-quarter fiscal 2027 guidance came in at $78 billion, exceeding the Street consensus.

Morris/Bloomberg via Getty Images

Where Wall Street stands after GTC

Rosenblatt is not alone in turning more bullish after GTC. Several other firms have raised their targets in recent weeks.

  • Wolfe Research reiterated an outperform rating with a $275 target, highlighting Nvidia’s $1 trillion revenue visibility across 2026 and 2027.
  • Bank of America raised its target to $300 from $275, also maintaining a buy, after analyst Vivek Arya updated his model following the Q4 earnings beat, raising his estimates across fiscal years 2027 through 2029.
  • Truist raised its target to $287, also with a buy rating, citing three inference demand drivers it described as combining forces, per Investing.com.

The Street consensus price target currently sits around $267, reflecting broad but measured optimism.

Rosenblatt’s $325 sits well above that range, making it among the highest targets on the Street, alongside Tigress Financial at $360.

The AI inference inflection point

Much of the bull case now hinges on inference rather than training. Cassidy’s note, echoed by Wolfe and others, centers on Nvidia’s upcoming Vera Rubin platform, expected in the second half of 2026, which the company says will deliver up to a 10x reduction in inference token cost compared to Blackwell.

In addition, Nvidia’s Dynamo inference software, announced at GTC, is designed to intelligently route workloads across GPU, ASIC, and CPU clusters, a capability Rosenblatt believes widens Nvidia’s moat against custom silicon from Amazon, Google, and AMD.

Sovereign AI contracts are also contributing to the growth outlook. Nvidia has announced major deals in Saudi Arabia, the UAE, India, and Japan as governments invest in domestic AI infrastructure.

Key risks analysts flag

Not every analyst shares Rosenblatt’s conviction. The bears point to a handful of structural risks that could limit upside from current levels.

Hyperscalers including Amazon, Google, and Microsoft are all developing proprietary AI accelerators aimed at reducing their dependence on third-party chips. AMD continues to gain ground in certain segments. Power infrastructure constraints in the U.S. could slow data center buildouts.

And China export restrictions remain a wildcard, with Nvidia itself excluding China revenue from its current guidance.

Nvidia’s next earnings report is expected on May 27, 2026, which will give investors the first clear look at whether Q1 guidance translates into another beat-and-raise quarter.

Related: Bank of America resets Nvidia price target after earnings