AMD guided its fourth-quarter non-GAAP gross margin to about 55%, from 56% in the third quarter, on revenue about $3.6 billion higher. By our arithmetic, the extra revenue carries a gross margin of about 52%. On the call, CFO Jean Hu said the growing AI business dilutes the average, and Lisa Su said memory costs are rising.
On the earnings call on 3 November, Su said AMD is “working with our customers to ensure that those costs are shared as we go forward”, referring to memory and other components facing “inflationary pressures”.
Business Pill · The Memory Wall
An evergreen Business Engineer explainer on the memory wall in AI chips. It teaches the general idea only and does not discuss AMD’s quarter.
The key insight: By our arithmetic, AMD’s 55% fourth-quarter guide implies about 52% gross margin on the extra $3.6 billion of revenue, against 56% on the third quarter; Hu tied the dilution to the AI business, and Su said AMD is working with customers to share rising memory and component costs.
A TD Cowen analyst asked how HBM costs affect AMD’s model as Instinct ramps, and whether AMD can pass higher input costs to its largest customers. About 22:38 UTC, Su answered: “So the way to think about it is as we ramp, we have had, let’s call it, long term agreements with some of our suppliers as well as some of our customers. But given where the memory market is, and not just the memory market, by the way, other component costs are also seeing some inflationary pressures”.
She continued that AMD is “working with our customers to ensure that those costs are shared as we go forward”, and said: “And even with some of the inflationary pressures that are out there, we certainly believe that we are offering leading TCO for the AI workloads.” Asked whether AMD is meeting server CPU demand, she said: “I do believe there’s more demand”, and named “DDR5 and all of the other components” among the parts AMD needs.
AMD’s non-GAAP gross margin in the third quarter of 2026 (release), the fourth-quarter guide of about 55% given by CFO Jean Hu on the 3 November call, and the margin on the extra fourth-quarter revenue, which is our arithmetic.
What Hu Said About Margin
Hu guided “the fourth-quarter non-GAAP gross margin to be approximately 55%”. Answering the same analyst, she said the AI business is “a little bit of dilutive to corporate average, but gross margin dollars are increasing very significantly”.
Asked by a Bernstein analyst for an entrance and exit rate for gross margins in 2027, she said the Data Center mix is the primary driver, and: “So the mix between those two businesses actually determine what the company’s gross margin will be. But right now, I really don’t have that visibility.”
Memory and PCs
Memory costs also reached AMD’s PC business. Su said “PC sell-through slowed late in the quarter as higher memory and component costs weighed on demand”, and that “In 2027, we are planning for PC market unit shipments to decline by a double-digit percentage”. Hu guided client and gaming revenue down more than 20% sequentially in the fourth quarter; from $3.722 billion, that is below about $2.98 billion, by our arithmetic.
The Arithmetic
AMD’s third-quarter non-GAAP gross profit was $7.483 billion on $13.377 billion of revenue, according to its release. A 55% margin on the $17 billion fourth-quarter midpoint gives about $9.35 billion. The increase, about $1.87 billion, on about $3.62 billion of extra revenue is an incremental gross margin of about 52%, by our arithmetic, against 56% on the third quarter as a whole.
How Each Side Reads These Figures
For readers who see pricing power, Su said AMD is working with customers to share rising costs, has long-term agreements with some suppliers and customers, and still offers leading total cost of ownership, in her words.
For readers who see margin pressure, the fourth-quarter guide is a point lower than the third quarter, Hu said the AI business dilutes the average, and she said she lacks visibility on 2027 gross margin. We cite no analysts’ views beyond their questions.
As we read it, the step-down is small in the average and larger at the margin: the new revenue earns several points less than the old.
As we read it, Su described cost-sharing as work in progress, not settled terms: long-term agreements with some suppliers and customers, and talks to share costs going forward.
As we read it, memory reaches AMD twice, in HBM for Instinct GPUs and in DRAM prices that Su said slowed PC sell-through.
Lisa Su, AMD earnings call, 3 November 2026 (our transcript)
“we are also working with our customers to ensure that those costs are shared as we go forward”
Three Implications
FOR AI CUSTOMERS Su said AMD is working with customers to share rising memory and component costs, while arguing it still offers leading total cost of ownership.
FOR AMD SHAREHOLDERS Hu guided fourth-quarter gross margin to about 55% and said she lacks visibility on 2027 gross margin because it depends on the mix of AI and server CPUs.
FOR PC MAKERS Su said AMD is planning for PC unit shipments to fall by a double-digit percentage in 2027, after memory costs slowed sell-through.
The framework puts it this way: “HBM growth is outpacing accelerator demand, making memory the rate-limiting factor in the AI scaling equation.”
As we read it, Su’s answer shows the commercial side of that constraint: a chip supplier negotiating who pays for memory. The framework was published in December 2025 and does not discuss AMD’s margins.
Quotations come from our own machine transcript of the call audio, as rebroadcast live; it is not an official AMD transcript and times are approximate. AMD did not say how much of the margin step-down comes from memory costs rather than product mix, and did not give HBM prices or contract terms. The analyst’s per-gigawatt revenue range is his description, not an AMD figure, and we do not use it.
The $9.35 billion, $1.87 billion, $3.62 billion, 52% and $2.98 billion figures are our arithmetic from AMD’s release and Hu’s approximate guidance. We did not contact AMD.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growthstory. This week it started being priced as a credit story.
AMD guided fourth-quarter gross margin to about 55%, implying about 52% on the extra $3.6 billion of revenue, by our arithmetic, as Su said AMD is working with customers to share rising memory and component costs and Hu said the growing AI business dilutes the average.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. Quotations from Lisa Su and Jean Hu come from our own machine transcript (faster-whisper) of AMD’s third-quarter 2026 earnings call of 3 November, as rebroadcast live; it is not an official transcript and times are approximate. Third-quarter figures come from AMD’s release filed as Exhibit 99.1 to its 8-K of 3 November. The fourth-quarter gross profit, the incremental margin and the client and gaming bound are our arithmetic. We did not contact AMD.
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.
