Skip to content
audiotale.net
Sign in
NewsNews

Nvidia Q3 FY2027 results: $115.8B revenue, $130B guide, no China data center sales assumed

Nvidia's first report since Jensen Huang's exit shows revenue up 103% to $115.8 billion and a $130 billion outlook that assumes no Data Center compute sales to China.

audiotale.net team

· 6 min read

Written by AI from the sources linked below. How we use AI

On this page
  1. Key points
  2. What happened
  3. Background
  4. The numbers
  5. How it works
  6. What it means
  7. What we don't know yet
  8. What to watch next
  9. Sources

Nvidia reported revenue of $115.8 billion for its third fiscal quarter of 2027 (the three months to 25 October 2026), up 103% on a year earlier, and told investors to expect $130.0 billion, plus or minus 3%, in the fourth quarter. It is the company's first earnings report since Jensen Huang stepped down as chief executive, and the first in which interim CEO Colette Kress said Nvidia is no longer shipping licensed Hopper data center products to China and assumes no Data Center compute revenue from China in its outlook. For developers and AI teams, the numbers show where the compute money is going: Vera Rubin racks have started shipping, and demand from AI clouds, industry and enterprises is growing faster than demand from the big hyperscalers.

Key points

  • Revenue was $115.8 billion, up 20% from the previous quarter and 103% from a year ago; Data Center revenue was $107.6 billion, up 110% year on year (Nvidia).
  • Nvidia guides the fourth quarter to $130.0 billion, plus or minus 3%, in a quarter with an extra week, with gross margin of 71.0% to 72.0% (Nvidia).
  • The outlook assumes no Data Center compute revenue from China; Data Center Hopper shipments to China were $2.4 billion in the quarter (Nvidia CFO commentary, as reproduced by MarketScreener).
  • Sales to "AI Clouds, Industrial, & Enterprise" customers rose 195% to $47.5 billion, against $60.1 billion (up 71%) from hyperscalers (CFO commentary, as reported by MLQ.ai and MarketScreener).
  • Nvidia says it began shipping Vera Rubin NVL72, "the fastest product ramp in the company's history".

What happened

Nvidia published its results on 17 November 2026 and held its call at 2 p.m. Pacific. The company's own numbers, from its press release:

  • Revenue: $115.8 billion, against $96.2 billion in the second quarter and $57.0 billion a year earlier. Nvidia had forecast about $108.0 billion, plus or minus 2%, in August (MLQ.ai).
  • Data Center: $107.6 billion, up 21% from the previous quarter and 110% from a year ago.
  • Edge computing: $8.2 billion, up 14% on the quarter and 42% on the year. Together with Data Center this accounts for the full $115.8 billion.
  • Margins and profit: GAAP gross margin 74.8% (non-GAAP 74.9%); GAAP diluted earnings per share $2.70, non-GAAP $2.68; GAAP net income $65.3 billion.
  • Shareholder returns: $28.1 billion returned in the quarter through repurchases and dividends; about $227.3 billion remained under the buyback authorisation. In September the board added $150 billion to it. The next dividend is $0.25 a share, payable 23 December 2026.

Nvidia also announced an agreement to acquire Hugging Face, listed among the quarter's highlights. Kress indicated on the call that the deal is still on track, according to CNBC's account carried by MLQ.ai. We do not have the terms.

Background

The report followed an unusual week for the company. According to MLQ.ai's account of the call, board chair Harvey Jones said Huang resigned on 12 November "after disclosure of personal behavior inconsistent with the company's code of conduct, standards, and cultural norms", and that the board "does not intend to reinstate Jensen as CEO". Jones said the board has retained Heidrick & Struggles to search for a permanent chief executive and would not take questions about the circumstances of the resignation. We cover the leadership change only as it bears on the business; the company has said little else, and we do not report unconfirmed claims.

Nvidia also announced on 17 November that its board has set up a national security and export control oversight committee. That sits alongside the China change below.

The Verge reports that Nvidia received White House approval in December last year to ship Hopper-generation H200 chips to China, and that Reuters reported this month that about 600,000 H200 chips were stuck in Hong Kong after Beijing withheld approval for large-scale imports. We have not independently verified the Reuters figure.

The numbers

Revenue has roughly doubled in a year and is guided higher again.

Nvidia quarterly revenue, with the fourth-quarter outlookUS$ billion
Nvidia quarterly revenue, with the fourth-quarter outlook (US$ billion)
US$ billion
Q3 FY2657.0
Q2 FY2796.2
Q3 FY27115.8
Q4 FY27 outlook130.0

Source: NVIDIA, third quarter fiscal 2027 results

Inside Data Center, the mix is shifting away from the biggest cloud providers:

Nvidia Q3 FY27 revenue by customer group and segmentUS$ billion
Nvidia Q3 FY27 revenue by customer group and segment (US$ billion)
US$ billion
Hyperscale60.1
AI Clouds, Industrial, & Enterprise47.5
Edge Computing8.2

Source: NVIDIA CFO commentary, as reproduced by MarketScreener and MLQ.ai

Year on year, hyperscale revenue rose from $35.1 billion to $60.1 billion (up 71%), while the AI clouds, industrial and enterprise group went from $16.1 billion to $47.5 billion (up 195%), per the CFO commentary as reproduced by MarketScreener. By segment, Compute & Networking was $106.0 billion and Graphics $9.8 billion.

Margins are heading the other way: guidance for the fourth quarter is 71.0% to 72.0%, down from 74.8%. MLQ.ai notes the third-quarter margin slipped from 75.0% in the second quarter as the Rubin ramp began. Operating expenses are expected at about $11.0 billion on a GAAP basis ($10.7 billion non-GAAP).

The company's commitments grew too. According to MarketScreener's reproduction of the CFO commentary, inventory rose to $39.7 billion from $31.6 billion in the previous quarter, and supply commitments to $390 billion from $279 billion, "primarily related to the multi-year procurement of memory supply". We could not check those figures against the original commentary, so treat them as reported.

How it works

Nvidia's customer split matters because of how AI infrastructure is financed. Hyperscalers buy with their own cash. The AI clouds and enterprise group includes neoclouds and companies building their own clusters, and they often need outside financing. Asked about guaranteeing customers' data center financing, Kress said the guarantees take effect only once a facility is ready and "If all of that begins, and unfortunately they are delayed or they have a challenge in terms of building that, our guarantee doesn't exist anymore", adding "Is it new? No. Did we change anything? No" (as quoted by MLQ.ai from GamesBeat). Kress's guidance for next year, she said, "considers data center readiness, the strength and durability of end customer demand, and our customers access to capital".

The product cycle is also moving. Nvidia says Vera Rubin NVL72 racks are running at CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure and Nebius, and are coming online soon at SpaceXSI. It also says Google Cloud's A5X instances and Oracle's Vera Rubin systems are generally available. For readers following our hardware coverage, the company recently introduced a science-focused GPU; see our Rubin HPC article.

What it means

For AI builders. Nvidia says its roadmap is on track and Vera Rubin racks are shipping to cloud providers, so Rubin-class capacity should start appearing in the cloud. The frontier labs are a large buyer: Kress said they should account for about 15% of Data Center sales this year, and said Anthropic's annualised revenue reached $81 billion at the end of October (as reported by CNBC through MLQ.ai; we have not confirmed that figure with Anthropic).

For China-exposed companies. Nvidia's licensed Hopper shipments to China were $2.4 billion in the quarter, and revenue from customers headquartered in China including Hong Kong was $11.1 billion, per MLQ.ai's reading of the commentary. From now on, Nvidia's guidance includes none of the former. Kress did not say why, according to The Verge.

For investors, the shape of the quarter. Revenue beat the company's own forecast and LSEG's average estimate of $109.10 billion, per CNBC through MLQ.ai; the $130 billion outlook was above the $125.1 billion consensus. Shares rose about 3.6% in after-hours trading, MLQ.ai reports.

What we don't know yet

  • Why Nvidia stopped China shipments. Kress gave no reason, per The Verge.
  • The permanent CEO. No timeline beyond the board's wish to appoint "as quickly as possible" without rushing.
  • Hugging Face terms and timing. The release lists an agreement but we have no price or closing date.
  • The fiscal 2028 outlook. MLQ.ai reports Kress now expects revenue growth of about 75%, up from 70% in August. We have not seen the transcript.

What to watch next

  • Delivery volumes of Vera Rubin NVL72 and how gross margins move through the fourth quarter.
  • Whether the board names a permanent chief executive.
  • The status of H200 chips held outside China and any new export rules.
  • Regulatory review of the Hugging Face acquisition.

Sources

Written by

audiotale.net team

The official account of the people who run audiotale.net: announcements, conversation starters and the newsroom's drafts. Posts written with AI say so. Every other account here is a real person.