Broadcom booked $16.7 billion of AI semiconductor revenue in its quarter ended 2 August 2026, its results release says, against $89,023 million of Data Center revenue at Nvidia in the quarter ended 26 July, per Nvidia’s 10-Q. That puts Broadcom’s AI line at 18.8% of Nvidia’s Data Center revenue, up from 12.7% a year earlier, by our arithmetic.
Over the same year, AMD’s Data Center segment slipped from 7.9% to 7.5% of Nvidia’s, and Marvell’s data center revenue from 3.6% to 2.4%, by our arithmetic from their filings. Broadcom is the only one of the three whose AI or data center revenue grew faster than Nvidia’s.
Business Pill · Vertical Integration
An evergreen Business Engineer explainer on vertical integration. It teaches the general idea only and does not discuss Broadcom, AMD, Marvell, Nvidia or their filings.
The key insight: As we read the four filings, the ground Nvidia’s rivals gained in the past year went to Broadcom: its AI revenue rose from 12.7% to 18.8% of Nvidia’s Data Center line, while AMD’s Data Center segment slipped from 7.9% to 7.5% and Marvell’s data center revenue from 3.6% to 2.4%, by our arithmetic.
“Demand for our custom AI accelerators and networking continues to be very strong”, Broadcom chief executive Hock Tan said in the release of 2 September. “Q3 AI semiconductor revenue of $16.7 billion grew 221% year-over-year, and 54% quarter-over-quarter”, the release says.
A year earlier, for the quarter ended 3 August 2025, Broadcom said “Q3 AI revenuegrowth accelerated to 63% year-over-year to $5.2 billion”. For the current quarter, the release says Broadcom expects “AI semiconductor revenue to accelerate to $21.7 billion, up 236% year-over-year”, $5.0 billion more than the third quarter, by our arithmetic.
Broadcom’s AI line covers both custom AI accelerators and networking, as Tan’s comment describes; the release figures we read do not split the two.
Broadcom’s AI semiconductor revenue, AMD’s Data Center segment and Marvell’s data center revenue, each as a share of Nvidia’s Data Center revenue, in mid-2025 and mid-2026 quarters. Shares are our arithmetic from the companies’ filings; the lines are not like for like.
Against Nvidia’s Data Center Line
Nvidia’s 10-Q lists Data Center revenue of $89,023 million for the quarter ended 26 July 2026 and $41,096 million a year earlier, growth of 117%. Broadcom’s AI revenue grew 221% over a similar period, so its share of Nvidia’s figure rose by 6.1 points, by our arithmetic.
In dollars, the gap still widened. Nvidia’s Data Center revenue was $72.3 billion larger than Broadcom’s AI revenue in the latest quarter, against $35.9 billion a year earlier, by our arithmetic.
AMD and Marvell
AMD’s 10-Q for the quarter ended 27 June 2026 lists Data Center revenue of $6,718 million, against $3,240 million a year earlier, growth of 107% by our arithmetic. That is 7.5% of Nvidia’s Data Center revenue, down from 7.9%, as our piece on AMD’s data center line set out.
Marvell’s results release lists data center revenue of $2,171.5 million for the quarter ended 1 August 2026, up 46% from $1,490.5 million. Its data center end market includes AI systems and servers as well as ethernet switching, storage and data center interconnect, the release says.
Broadcom’s AI revenue was 2.5 times AMD’s Data Center revenue in the latest quarters, up from 1.6 times a year earlier, by our arithmetic. The three companies together came to about 29% of Nvidia’s Data Center figure, against about 24% a year earlier.
A J.P. Morgan Asset Management chart of frontier labs’ future compute agreements by chip supplier is in our piece on the frontier-lab deals.
The Structural Read
As we read Broadcom’s release, its AI line is described as custom AI accelerators and networking, so the 18.8% is a revenue comparison, not a chip-share figure; the filings we read do not split the two.
As we read the dollar figures, the ratio and the gap moved in opposite directions: Broadcom’s share of Nvidia’s Data Center revenue rose 6.1 points while the dollar gap doubled from $35.9 billion to $72.3 billion, by our arithmetic.
Hock Tan, Broadcom, 2 September 2026
“Demand for our custom AI accelerators and networking continues to be very strong”
Three Implications
FOR BROADCOM Its guided $21.7 billion of fourth-quarter AI semiconductor revenue is $5.0 billion above the third quarter, by our arithmetic from its release.
FOR NVIDIA Its 17 November results replace the $89,023 million second-quarter figure used in these ratios, per its release timing.
The framework argues: “The within-layer split is what makes silicon strategically interesting again.”
As we read it, Broadcom’s AI revenue reaching 18.8% of Nvidia’s Data Center line, by our arithmetic, is one revenue measure of that split. The framework predates these filings and does not discuss these figures.
What Is Not Established
The four lines are not like for like. Broadcom’s AI figure includes networking, Marvell’s data center figure includes storage and general-purpose servers, and Nvidia’s Data Center line includes its own networking. The fiscal quarters also end on different dates, up to about a month apart.
The filings we read do not give a custom-accelerator revenue figure for Broadcom or Marvell on its own, so this is not a measure of custom-chip market share. The ratios use Nvidia’s second-quarter figure; Nvidia’s third-quarter results are due on 17 November, per the company. We did not contact the companies for this piece.
This Week In The AI Supercycle
AI Lost Its Central Banker, the Lab Oligopoly, and the Future of Open Source
This week’s issue: the Nvidia leadership change, read as the AI build-out losing its central banker: who now finances the GPU landlords, backstops the data centres and sponsors open models, and why two labs are becoming the bank.
Broadcom’s $16.7 billion of quarterly AI semiconductor revenue was 18.8% of Nvidia’s $89,023 million of Data Center revenue, up from 12.7% a year earlier, by our arithmetic from their filings. AMD’s Data Center segment slipped to 7.5% of Nvidia’s, from 7.9%, and Marvell’s data center revenue to 2.4%, from 3.6%.
A note on sourcing. Broadcom’s AI semiconductor revenue and quotes are from its results releases for the third quarters of fiscal 2026 and 2025, and Marvell’s data center revenue from its release for the second quarter of fiscal 2027, all filed with the SEC as 8-K exhibits. AMD’s and Nvidia’s segment revenue are from their 10-Q filings, and Nvidia’s results date from its 28 October release.
Shares, ratios, gaps and growth rates not stated in the filings are our arithmetic. The companies were not contacted for this piece.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
