NVIDIA's revenue accelerated growth of 62% last quarter, and the quarterly guidance exceeded expectations again. Jensen Huang stated that "Blackwell sales far exceeded expectations" | Earnings report insights
In the third quarter, total revenue accelerated year-on-year growth for the first time in two years, with data center revenue also reaching a record high for a single quarter, increasing by 66% year-on-year and nearly 25% quarter-on-quarter; the median revenue guidance for the fourth quarter is expected to grow by 65% year-on-year. The gross margin for the third quarter was slightly below expectations, but is expected to rise to 75% in the fourth quarter, marking the first year-on-year increase in six quarters. Jensen Huang stated that the sales of Blackwell chips far exceeded expectations, with both training and inference computing demands growing exponentially. The stock price rose more than 6% in after-hours trading. The CFO reiterated that new chips are expected to generate $500 billion in revenue over the next few quarters, noting that the A100 GPU shipped six years ago is still operating at full capacity this year. Update
As investors continue to worry about the unsustainable high investments in the artificial intelligence (AI) sector, AI chip leader NVIDIA brings good news: revenue growth accelerated in the last fiscal quarter, with a growth rate returning to over 60%, and revenue from data centers hitting a record high for a single quarter, reflecting the strong and sustained demand for AI infrastructure. Revenue for the current fiscal quarter is expected to maintain a growth rate of over 60%, once again exceeding Wall Street expectations.
NVIDIA CEO Jensen Huang stated in the earnings announcement that the company's latest generation Blackwell architecture chips "far exceeded expectations, and cloud GPUs are sold out," adding, "The demand for training and inference computing continues to accelerate, both showing exponential growth. We have entered a virtuous cycle of AI."
Commentators noted that the cloud GPU sell-out is typically Huang's way of indicating that demand from chip buyers and users—cloud service providers—is strong and that there is no idle capacity. This is clearly part of his recent promotional strategy aimed at reassuring investors concerned about overcapacity. Other comments pointed out that the earnings report showed that as of the end of the last fiscal quarter, NVIDIA held $60.6 billion in cash and cash equivalents, indicating that it still has ample funds to support AI applications in new economic fields.
After the earnings report was released, NVIDIA's stock, which had risen nearly 3%, saw its after-hours gains quickly expand, with after-hours gains exceeding 6%. Analysts believe that the total revenue and gross margin guidance provided by NVIDIA alleviated external concerns about an AI bubble.
On November 19, Eastern Time, NVIDIA announced its financial data for the third quarter of fiscal year 2026 (referred to as Q3), ending on October 26, 2025, as well as performance guidance for the fourth quarter (referred to as Q4).
1) Key Financial Data:
Revenue: Q3 operating revenue was $57.01 billion, a year-on-year increase of approximately 62%, compared to analyst expectations of $55.19 billion, and NVIDIA's own guidance of $52.92 billion to $55.08 billion, with the previous quarter showing a year-on-year growth of 56%.
EPS: Q3 adjusted earnings per share (EPS) on a non-GAAP basis was $1.30, a year-on-year increase of 60%, compared to analyst expectations of $1.26, with the previous quarter showing a year-on-year growth of 54%.
Gross Margin: Q3 adjusted gross margin was 73.6%, a year-on-year decrease of 1.4 percentage points, compared to analyst expectations of 74.0%, with NVIDIA's guidance being 73% to 74%, and the previous quarter at 72.7%, a year-on-year decrease of 3 percentage points.
Operating Expenses: Q3 adjusted operating expenses were $4.215 billion, a year-on-year increase of 38%, compared to analyst expectations of $4.22 billion, with NVIDIA's guidance being $4.2 billion, and the previous quarter showing a growth of 36%.
2) Segment Business Data:
Data Center: Q3 data center revenue was $51.2 billion, a year-on-year increase of 66%, compared to analyst expectations of $49.34 billion, with the previous quarter showing a year-on-year growth of 56%
Gaming and AI PCs: In the third quarter, revenue from gaming and AI PC businesses reached $4.3 billion, a year-on-year increase of 30%, while analysts expected $4.42 billion, with a year-on-year growth of 49% in the previous quarter.
Professional Visualization: In the third quarter, professional visualization revenue was $760 million, a year-on-year increase of 56%, while analysts expected $612.8 million, with a year-on-year growth of 32% in the previous quarter.
Automotive and Robotics: In the third quarter, revenue from automotive and robotics businesses was $592 million, a year-on-year increase of 32%, while analysts expected $620.9 million, with a year-on-year growth of 69% in the previous quarter.
3) Performance Guidance:
Revenue: Fourth quarter revenue is expected to be $65 billion, with a fluctuation of 2%, i.e., $63.7 billion to $66.3 billion, with the median analyst expectation at $61.98 billion.
Gross Margin: The adjusted non-GAAP gross margin for the fourth quarter is expected to be 75.0%, with a fluctuation of 50 basis points, i.e., 74.5% to 75.5%, with the median analyst expectation at 74.6%.
Operating Expenses: Adjusted operating expenses for the fourth quarter are expected to be $5 billion, with analysts expecting $4.59 billion.
Total Revenue Accelerates Year-on-Year Growth for the First Time in Two Years, Data Center Revenue Hits Record High
The financial report shows that NVIDIA's third quarter revenue reached a record $57 billion, exceeding the company's entire guidance range and more than 3% higher than analyst expectations. The year-on-year revenue growth rate increased from 56% in the previous quarter to 62%, marking the first acceleration in revenue growth compared to the previous quarter in two years since the fourth quarter of fiscal year 2024.
The data center business, which contributed nearly 90% of third quarter revenue, also set a record for single-quarter revenue, exceeding analyst expectations by nearly $19 billion, or about 4%, with a year-on-year growth rate increasing from 56% in the previous quarter to 66%. Compared to the first quarter, data center revenue grew by approximately $10.1 billion, a quarter-on-quarter increase of 24.6%.
In the third quarter, the adjusted EPS growth rate also accelerated year-on-year, increasing from 54% in the previous quarter to 60%, marking the highest growth rate within this fiscal year.
Revenue Guidance Median Expected to Increase by 65% Year-on-Year, Gross Margin Rises to 75%
From the performance guidance, NVIDIA's entire fourth quarter revenue guidance range is above the median analyst expectations. Based on the median of the guidance range of $65 billion, NVIDIA expects fourth quarter revenue to set a new single-quarter high, nearly 5% higher than the median analyst expectation, and is expected to grow by over 65% year-on-year, indicating a potential to maintain growth above 60%However, some very optimistic Wall Street analysts previously expected revenue to be $75 billion, which is more than 15% higher than the midpoint of NVIDIA's guidance range.
NVIDIA's gross margin for the third quarter fell to 73.6%, which was lower than analysts' expectations of 74.0%, but the gross margin for the fourth quarter
For the fourth quarter, NVIDIA expects the gross margin midpoint to increase to 75.0% from 73.5% in the same period last year, marking the first year-on-year growth in gross margin in six quarters since the second quarter of fiscal 2025.
Wedbush Securities analyst Matt Bryson pointed out that NVIDIA achieved two goals that investors previously worried it could not reach: first, predicting total revenue to exceed $60 billion, and second, expecting a gross margin of around 75%.
Kunjan Sobhani, a senior technology analyst at Bloomberg Industry Research, commented that NVIDIA's third-quarter performance and fourth-quarter guidance both exceeded market expectations, with the guidance particularly high compared to expectations, indicating that it is ahead of most buy-side institutions' forecasts. This suggests that the production capacity of the GB300 chip will be significantly increased.
CFO: The A100 GPU shipped six years ago is still operating at full capacity this year
Institutions such as Citigroup and JP Morgan have previously held an optimistic view on NVIDIA's third-quarter performance. This confidence stems from NVIDIA's customer base. Microsoft, Amazon, Alphabet, and Meta together account for more than 40% of NVIDIA's sales, and it is expected that the total AI spending of these companies will grow by 34% to $440 billion over the next 12 months.
More importantly, Jensen Huang revealed at the GTC conference in October this year that NVIDIA has secured chip orders worth $500 billion for the calendar years 2025 and 2026, including the new generation Rubin chip, which will begin mass production next year.
During the earnings call on Wednesday, NVIDIA's Chief Financial Officer (CFO) Colette Kress reiterated that NVIDIA expects some of its most advanced chips to generate $500 billion in revenue over the next few quarters, and there is a massive overall opportunity of trillions of dollars in the data center infrastructure sector.
Kress stated that the number of users and gross margin of OpenAI are encouraging signs, and the revenue of its competitor Anthropic is also similarly encouraging.
Kress mentioned that in the third quarter, the GB300 contributed more to NVIDIA's revenue than its predecessor GB200, achieving a "cross-over." The A100 GPU shipped six years ago is still operating at full capacity this year.
Kress revealed that sales of the customized AI chip H20 for the Chinese market were only $50 million in the third quarter. She said that due to geopolitical issues and increasing market competition, orders for the H20 AI chip will never be very large. The company is working to find ways to export "more competitive data center computing products" to China
