Nvidia Beat Its Own Forecast and Investors Sold Anyway
The company guided fourth-quarter revenue to $65.0 billion, roughly $3.8 billion above what the Street was modeling. The stock rose in after-hours trading.
That's the problem with being the most important company in the most important sector of the most important technology cycle in a generation. You don't get credit for clearing the bar. You get punished for not clearing it by enough.
The Expectation Trap
Nvidia's third-quarter fiscal 2026 results, released on November 19, 2025, showed revenue of $57.0 billion, up 62% year-over-year. Data center revenue, which powers AI chip buildouts at Microsoft, Meta, and Amazon, hit $51.2 billion. Gross margin came in at 73.6%. Every number beat consensus. The guide beat consensus. The stock rose in after-hours trading.
The stock traded at a high forward earnings multiple heading into the print, a valuation that assumes strong execution of growth. When a company carries that multiple, a beat is the baseline. What moves the stock is the gap between the beat and what wasn't yet priced in. Nvidia's guide implied revenue growth of roughly 14% quarter-over-quarter. That's strong. The company posted a 22% sequential leap from Q2 to Q3, and through most of 2024 and early 2025 investors had internalized that parabolic trajectory as the new normal.
The sequential growth trajectory looks reasonable when you understand semiconductor cycles. You can't compound at triple-digit rates forever. The law of large numbers eventually asserts itself. A $57 billion revenue base cannot grow at the same pace as a $15 billion one. Equity markets are staffed by people who bought Nvidia in early 2023, watched it rise substantially, and now need it to keep growing to justify the position size.
What the Market Heard
The analyst call focused heavily on supply constraints, particularly around the new Blackwell chip architecture. CEO Jensen Huang noted that demand for Blackwell is "staggering" and that the company is working through production bottlenecks with TSMC. That is both bullish (the product is working, customers want it) and bearish (you can't ship what you can't make, and bottlenecks mean revenue gets pushed right).
Investors heard the second part louder.
China sales were collapsing. Nvidia's China revenue has been falling since the Biden administration tightened export controls on advanced AI chips in late 2023, and that pressure has only increased under subsequent rule changes through 2025 and 2026. The company is designing workaround chips that comply with the new thresholds, but those products carry lower margins and serve a customer base that is now buying on a shorter planning horizon. China used to be 20% of Nvidia's revenue. It is now in the single digits, and the replacement revenue has to come from U.S. and European hyperscalers who are already buying at near-maximum capacity.
The bull case for Nvidia rests on the assumption that AI chip spending is in the second inning. The bear case is that it's in the seventh, and the model-training gold rush that drove 2024 and early 2025 buying is now giving way to slower-growth inference workloads that don't require the same chip density.
Both cases can be true for a while. What cannot be true for long is a valuation that assumes the second-inning story with no room for the seventh-inning risk. The market is pricing Nvidia for perfection. Perfection arrived. It wasn't enough.
The company guided fourth-quarter revenue to $65.0 billion, roughly $3.8 billion above what the Street was modeling. The stock rose in after-hours trading.
That's the problem with being the most important company in the most important sector of the most important technology cycle in a generation. You don't get credit for clearing the bar. You get punished for not clearing it by enough.
The Expectation Trap
Nvidia's third-quarter fiscal 2026 results, released on November 19, 2025, showed revenue of $57.0 billion, up 62% year-over-year. Data center revenue, which powers AI chip buildouts at Microsoft, Meta, and Amazon, hit $51.2 billion. Gross margin came in at 73.6%. Every number beat consensus. The guide beat consensus. The stock rose in after-hours trading.
The stock traded at a high forward earnings multiple heading into the print, a valuation that assumes strong execution of growth. When a company carries that multiple, a beat is the baseline. What moves the stock is the gap between the beat and what wasn't yet priced in. Nvidia's guide implied revenue growth of roughly 14% quarter-over-quarter. That's strong. The company posted a 22% sequential leap from Q2 to Q3, and through most of 2024 and early 2025 investors had internalized that parabolic trajectory as the new normal.
The sequential growth trajectory looks reasonable when you understand semiconductor cycles. You can't compound at triple-digit rates forever. The law of large numbers eventually asserts itself. A $57 billion revenue base cannot grow at the same pace as a $15 billion one. Equity markets are staffed by people who bought Nvidia in early 2023, watched it rise substantially, and now need it to keep growing to justify the position size.
What the Market Heard
The analyst call focused heavily on supply constraints, particularly around the new Blackwell chip architecture. CEO Jensen Huang noted that demand for Blackwell is "staggering" and that the company is working through production bottlenecks with TSMC. That is both bullish (the product is working, customers want it) and bearish (you can't ship what you can't make, and bottlenecks mean revenue gets pushed right).
Investors heard the second part louder.
China sales were collapsing. Nvidia's China revenue has been falling since the Biden administration tightened export controls on advanced AI chips in late 2023, and that pressure has only increased under subsequent rule changes through 2025 and 2026. The company is designing workaround chips that comply with the new thresholds, but those products carry lower margins and serve a customer base that is now buying on a shorter planning horizon. China used to be 20% of Nvidia's revenue. It is now in the single digits, and the replacement revenue has to come from U.S. and European hyperscalers who are already buying at near-maximum capacity.
The bull case for Nvidia rests on the assumption that AI chip spending is in the second inning. The bear case is that it's in the seventh, and the model-training gold rush that drove 2024 and early 2025 buying is now giving way to slower-growth inference workloads that don't require the same chip density.
Both cases can be true for a while. What cannot be true for long is a valuation that assumes the second-inning story with no room for the seventh-inning risk. The market is pricing Nvidia for perfection. Perfection arrived. It wasn't enough.
Sources
Read Next
GIC rates stall as bonds retreat: why deposit accounts still lag
Why Your Employer's Life Insurance Probably Covers Less Than You Think
August Sales Fell 6.9%: What CREA's Numbers Actually Reveal About Buyer Hesitation
Canada's Resource Advantage: What Global Investors Are Betting On in 2026