Nvidia has reported the biggest quarter in its history, and the first since Jensen Huang resigned as chief executive. Revenue for the three months to 25 October came in at $115.8 billion, up 103% on a year earlier and well ahead of the roughly $109 billion analysts had pencilled in. The company also guided to about $130 billion for the current quarter, and interim chief executive Colette Kress raised its growth outlook for next year.
For anyone following robots and AI, this is the number that tells you whether the money behind the boom is still flowing. Right now, it is.
The numbers
The figures below come from Nvidia’s own results release, and CNBC and The Verge reported the same headline numbers on Tuesday evening.
- Revenue: $115.8 billion, up 20% on the previous quarter and 103% on the year. Analysts surveyed by LSEG expected $109.1 billion.
- Data centre: $107.6 billion, up 110% on the year and about 93% of all sales.
- Profit: net income of $65.3 billion on a GAAP basis, with a gross margin of 74.8%. Adjusted earnings were $2.68 a share against $2.48 expected.
- Edge computing: $8.2 billion, up 42% on the year. This is the division that holds robotics, in-car computing and gaming PCs.
- Guidance: $130 billion, plus or minus 3%, for the fourth quarter, which includes an extra week. Nvidia says it is assuming no data-centre compute revenue from China.
CNBC reports that Kress also lifted the outlook for the financial year ending January 2028 to about 75% revenue growth, up from roughly 70% before.
The new chips are shipping
The headline product is Vera Rubin, the successor to Blackwell. Nvidia says its Vera Rubin NVL72 racks have begun shipping in what it calls the fastest product ramp in its history, and that they are already running at CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure and Nebius. Kress said the platform is “off to an explosive start”. Nvidia said the move to Rubin was the reason gross margin slipped slightly from the previous quarter.
Nvidia also pointed to demand from the AI labs. Kress said on the call that Anthropic’s annualised revenue reached $81 billion at the end of October, and Nvidia expects frontier labs such as OpenAI and Anthropic to account for around 15% of its data-centre sales this year. That ties neatly to Anthropic’s planned stock market debut, which is due to price today.
What it means for robots
Robotics is a small part of Nvidia’s revenue and sits inside the $8.2 billion edge line, so this report is not a robotics report. It matters because Nvidia’s chips, simulation tools and robot models underpin much of the industry. The release lists new Isaac tools, datasets and runtime for general-purpose robots, and notes that Agile Robots introduced a reference humanoid powered by Nvidia’s Jetson Thor and built on Isaac GR00T. We covered that company’s $500 million funding round and Nvidia’s GR00T 2 robot model earlier this month.
The simple reading is that while the AI infrastructure build-out keeps accelerating, the companies training robot brains have a well-funded supplier. The more cautious reading is that robotics still depends on a business whose fortunes are tied to a handful of enormous customers. Nvidia disclosed that in the last three quarters, four direct customers each accounted for between 11% and 13% of total revenue.
The questions behind the beat
The results land after a bruising month. CNBC reports that Nvidia’s shares are down 24% in November and just had their sixth straight weekly fall, the longest run since 2012, amid broader worries about an AI slowdown. The investor questions were less about last quarter than about what comes next.
- Leadership. Board chair Harvey Jones opened the call by saying the board “does not intend to reinstate” Huang as chief executive, and that the company would not take questions on the circumstances of his resignation or the independent investigation. Nvidia has hired the executive search firm Heidrick & Struggles to find a permanent successor. Our earlier report covers how Kress came to be interim chief executive.
- Commitments. Nvidia disclosed $118 billion of guarantees as of 25 October, mostly linked to data-centre financing including an OpenAI project in Ohio. It says its maximum exposure is $105 billion and that it has not made any payments under the guarantees to date. It also reported $32 billion of commitments to neoclouds through 2032, which only take effect if those companies secure financing and the capacity is up and running. Kress said Nvidia would be “rigorously evaluating” future commitments.
- China and export rules. The US Commerce Department tightened rules on Monday to stop Chinese firms renting advanced chips in third countries, and Nvidia has formed a board committee on national security and export controls. Nvidia’s guidance assumes no China data-centre compute revenue at all.
How the market reacted
Nvidia’s shares rose by close to 4% in after-hours trading, according to CNBC, and neocloud stocks such as CoreWeave, Nebius and Iren gained 8% to 10%. Kress told CNBC it was “business as usual” and that customers “continue to be with us”. Amazon’s Andy Jassy said separately that he did not foresee a day when Amazon was not a partner of Nvidia’s.
Not everyone is convinced. CNBC notes that Morgan Stanley trimmed its 2027 and 2028 data-centre sales estimates by about 10% after the leadership news, and that analysts see the real impact of the change at the top arriving after the current financial year, because contracted revenue is largely in the books until then.
What to watch next
- Anthropic’s pricing and first trade. CNBC reports the company is likely to price above its $70 to $78 range, and Nvidia is among the investors with a stake.
- The chief executive search. The board has not given a timetable.
- Export rules. Any further US or Chinese moves could change the China picture quickly.
- Robot releases. Nvidia has said the GR00T 2 code is due in December.
For the wider picture on the money pouring into AI hardware, see our coverage of UK chip start-up Fractile’s $720 million raise. Figures in this story are Nvidia’s own or as reported by the outlets named, and analyst estimates vary by data provider.