• NVIDIA (NVDA) reported fiscal Q4 revenue of $68.1 billion, a 73% year-over-year increase, and issued bullish Q1 guidance of $78 billion, yet its stock initially fell 4.5%.
  • Analysts attribute the drop to "sell-the-news" reactions and profit-taking, with concerns over the sustainability of hyperscaler AI capital expenditures.
  • Despite the pullback, Goldman Sachs (GS) reiterated a Buy rating with a $250 price target, citing strong demand-supply imbalances and NVIDIA's competitive edge in AI infrastructure.

A Quarter of Contrasts

NVIDIA's latest earnings report delivered the kind of numbers that typically send shares soaring: fiscal Q4 revenue surged 73% year-over-year to $68.1 billion, with earnings per share jumping 82% to $1.62, both handily beating analyst estimates. The company's guidance for the current quarter—$78 billion in revenue—further underscored the relentless demand for its AI chips, particularly from data centers where revenue grew 75% to $62.3 billion. Gross margins remained robust at 75.2%, and free cash flow hit $34.9 billion, painting a picture of a company firing on all cylinders.

Yet, in a twist that caught many off guard, NVIDIA's stock tumbled 4.5% in after-hours trading following the announcement, dragging the broader semiconductor sector and the S&P 500 lower. According to people familiar with the matter, the initial sell-off reflected a classic "sell-the-news" move, with investors locking in profits after a prolonged rally. One trader noted, "When you see numbers this good, it's hard not to take some chips off the table, especially with valuations where they are." The stock later rebounded slightly in extended trading as analysts digested the details, but the volatility highlighted underlying jitters.

The AI Capex Conundrum

Beneath the surface, concerns are mounting about the sustainability of the AI spending boom that has fueled NVIDIA's meteoric rise. Goldman Sachs, in a note to clients, pointed out that while AI capital expenditures are projected to grow 62% in 2026, this marks a slowdown from the 73% growth expected in 2025. "There's a growing fear that we might be approaching a peak in hyperscaler spending," said an analyst who requested anonymity due to firm policy. "Big Tech companies like Meta (META) and others are pouring billions into AI infrastructure, but at some point, that pace has to moderate."

NVIDIA's guidance, however, suggests the company remains confident in the near term. The recent expansion of its partnership with Meta, which includes commitments for next-generation Blackwell and Rubin AI GPUs, provides a tangible boost. "This deal underscores the ongoing demand from hyperscalers," a company spokesperson said in a brief statement, though they declined to comment on specific financial terms. Efforts to reach other major customers for additional comments were unsuccessful by press time.

Market Dynamics and Future Outlook

The disconnect between NVIDIA's stellar performance and its stock reaction isn't entirely new; the company has seen similar patterns in past quarters where blowout results were met with flat or declining shares. This time, the dislocation is more pronounced, with shares largely stagnant over the past five months despite a 37% rise in forward earnings estimates. Goldman Sachs attributes this to "over-earning" fears, where investors worry that current growth rates are unsustainable. "We see a clearer path to outperform," the firm wrote, maintaining its $250 price target, which implies roughly 30% upside from current levels.

Looking ahead, the focus is shifting to 2027, when some analysts predict AI capex might peak. In the short term, however, demand-supply imbalances and memory constraints continue to support NVIDIA's position. The company's full-stack advantage—encompassing its CUDA software ecosystem and advanced GPU architectures—remains a significant barrier to competitors like AMD (AMD) and Broadcom (AVGO), which are gaining traction but still lag in overall market share. As one industry insider put it, "NVIDIA isn't just selling chips; it's selling the entire AI infrastructure, and that's a hard moat to cross."

Correction: An earlier version of this article misstated the year for fiscal Q4 revenue; it is for 2026, not 2025. The text has been updated.