Brockman: OpenAI Gave AMD a Year of Chip Gains in 5 Months

OpenAI president Greg Brockman said at The New York Times DealBook Summit on 18 November that, working with AMD, OpenAI’s models delivered “a year’s worth of full generational improvements in five months” in chip design, in a clip posted by Wall St Engine.

Our number: a year’s worth of gains in five months is about 2.4 times the usual pace, by our arithmetic (12 months divided by 5).

Business Pill · Lead Time Explained

An evergreen Business Engineer explainer on lead time. It teaches the general idea only and does not discuss Brockman’s remark.

The key insight: OpenAI’s president says its models gave AMD a year’s worth of chip-design improvements in five months, about 2.4 times the usual pace by our arithmetic, a claim made on stage with no published data yet.

What Brockman Said

Brockman framed it as an example of AI moving into “economically valuable domains”. His words, as we transcribed them from the clip: “we’re starting to see in chip design, we’re announcing today that working with AMD, we’re able to get a year’s worth of full generational improvements in five months just using our models”.

He added: “That’s on top of what the humans were doing, like a year’s worth of improvements with just AI, just AI capability which is continuing to improve”.

Brockman said OpenAI's models gave AMD a year's worth of full generational improvements in five months, about
Brockman said OpenAI’s models gave AMD a year’s worth of full generational improvements in five months, about 2.4x the usual pace by our arithmetic. Source: NYT DealBook Summit clip.

The AMD Deal Behind It

The remark sits on top of a commercial tie. In AMD’s 6 October 2025 release, the companies announced “a 6 gigawatt agreement” covering multiple generations of AMD Instinct GPUs, with “The first 1 gigawatt deployment of AMD Instinct MI450 GPUs” set to begin in the second half of 2026.

AMD also said it “has issued OpenAI a warrant for up to 160 million shares of AMD common stock”, vesting as deployments scale toward 6 gigawatts and tied to AMD share-price targets. The release described the two companies “sharing technical expertise to optimize their product roadmaps”.

Why It Matters for the Chip Race

Faster design cycles matter because AMD is competing with a much larger Nvidia: Nvidia’s data-center revenue was about 12.7 times AMD’s on the latest reported quarters, per our analysis of the two companies’ filings. The remark also follows Brockman’s push to make cyber a bigger OpenAI business than coding, and lands as Nvidia’s stock and analyst targets reset after its first results without Jensen Huang.

The Structural Read

As we read it, the remark positions OpenAI’s models as a design tool for its own chip supplier, on top of being a buyer of up to 6 gigawatts of AMD GPUs.

As we read it, if design cycles shorten, the binding constraint shifts further toward manufacturing capacity, which the claim does not address.

Greg Brockman, OpenAI president, at the NYT DealBook Summit

“That’s on top of what the humans were doing, like a year’s worth of improvements with just AI”

Three Implications

FOR AMD AMD’s release ties OpenAI’s warrant for up to 160 million AMD shares to deployments scaling toward 6 gigawatts.

FOR OPENAI Brockman cited chip design as an example of AI moving into economically valuable domains.

FOR READERS The claim is an executive’s remark; OpenAI and AMD had not published the underlying data when we checked.

The Business Engineer Lens

This story maps onto the Business Engineer framework The AI Supercycle.

The analysis puts it this way: “Allocation power at the foundry is now more consequential than chip design itself.”

As we read Brockman’s claim, faster AI-assisted design would sharpen that point: a chip designed sooner still needs foundry capacity. The analysis does not discuss this remark.

What Is Not Established

Brockman said OpenAI was “announcing today” the AMD result, but when we checked OpenAI’s news feed at about 15:35 UTC on 18 November we found no post on it, and we did not find an AMD statement. He did not say which chip, which design stages, or how the improvement was measured. AMD did not comment in what we read, and we did not contact either company.

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.

Read the full issue →

Business Engineer Framework

The AI Supercycle

A Business Engineer analysis of the cycles running inside the AI buildout, from capability to capital to geopolitics.

Read the Map of AI →

The Bottom Line

OpenAI’s president says its models compressed a year of AMD chip-design gains into five months, about 2.4 times the usual pace; the claim is his, and neither company has published the underlying data yet.

97,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.

A note on sourcing. Brockman’s words are our transcription of a 46-second clip credited on screen to the New York Times DealBook Summit, posted by Wall St Engine; we did not watch the full session. Deal terms are from AMD’s 6 October 2025 release. The 2.4x figure is our arithmetic. We did not contact OpenAI or AMD.

Sources: Wall St Engine on X, clip of Greg Brockman at the NYT DealBook Summit, 18 Nov 2026 · AMD press release, 6 Oct 2025

Scroll to Top
FourWeekMBA