Why Nvidia stock is soaring despite broader market sell-off?

Nvidia stock (NASDAQ: NVDA) is up roughly 1.2% on Monday, an unusual sight on a day when the Dow is down over 800 points, and stagflation fears have most of Wall Street on the back foot.

The resilience isn't random, as a combination of fresh analyst conviction, strong recent earnings, and a major product event one week away is keeping institutional buyers engaged even as macro sentiment turns ugly.

Morgan Stanley's conviction

The institutional setup started on March 4, when Morgan Stanley's Joseph Moore reinstated Nvidia as his top semiconductor pick with a $260 price target, implying roughly 47% upside from last week's close.

Moore's core argument: Nvidia's stock has essentially gone nowhere for two quarters while the underlying business kept growing, creating what he described as an "unexpectedly favorable entry point."

The valuation data support that reading.

Nvidia now trades at roughly 18 times projected 2027 earnings, a relatively contained multiple for a company that reported 73% year-over-year revenue growth in its latest quarterly results.

For context, the stock traded at comparable levels in mid-2025 before staging a significant run. It hasn't looked this inexpensive relative to its growth rate in over a year.​

Supply-chain signals reinforce the bull case.

Nvidia's total supply-related commitments nearly doubled from $50.3 billion at the end of Q3 to $95.2 billion by the Q4 close.

Hyperscalers aren't just buying chips as they are locking in capacity well ahead of schedule.

Moore reads that as evidence that AI infrastructure spending is durable, not approaching a peak.​

Nvidia stock: GTC 2026 is one week away

Nvidia's annual GTC conference opens March 16 in San Jose.

Jensen Huang delivers the keynote to an expected 30,000 attendees with hundreds of thousands more expected to watch online.

He has already teased "a chip that will surprise the world," which is doing quiet work in Monday's tape.

The main item on the roadmap is Vera Rubin, which is Nvidia's next-generation combined GPU-CPU platform arriving in the second half of 2026.

The architecture promises a tenfold reduction in inference costs, meaning running AI models in real-world products gets significantly cheaper.

Lower inference costs typically expand Nvidia's market rather than cannibalize it.​

Beyond Vera Rubin, hardware analysts expect GTC to include a first public preview of Feynman, the generation after Rubin.

If Huang confirms early details, cloud buyers gain multi-year procurement visibility, the kind of forward planning that turns one-time deals into long-term revenue streams for Nvidia.​

The earnings backdrop gives the conference added weight. Q4 revenue came in at $68.1 billion, up 73% year over year, well above the $65.9 billion estimate.

The company guided for about $78 billion in revenue next quarter, well above market expectations. Full-year fiscal 2026 revenue reached $215.9 billion, with the data-center segment contributing about $197.3 billion.