OpenAI plans to buy Google’s AI chips, The Information reported late on 27 October 2026, in a story headlined “OpenAI, a Longtime Nvidia Customer, Plans to Buy Google’s AI Chips.” The financial terms and other details of the planned purchase could not be learned, according to the report as relayed by First Squawk.
The Information’s article is behind a paywall and we have not read its full text. OpenAI’s own careers page does show a job that points the same way: an engineer to “lead the optimization of OpenAI’s state-of-the-art frontier models on TPUs.”
Short explainer · AI chips brief
A vertical news Short narrated from the published headlines and opening lines of this piece and six related FourWeekMBA AI chip pieces.
Business Pill · GPUs
An evergreen Business Engineer explainer on why AI runs on GPUs. It teaches the general idea only and does not discuss OpenAI, Google or TPUs.
The key insight: As we read it, the report and OpenAI’s own job listing point the same way: OpenAI is preparing to run frontier models on Google’s TPUs as well as on the Nvidia GPUs it already uses at scale.
The Information’s report, by Ryan Barwick, was published at 23:59 UTC on 27 October. First Squawk’s relay of it on X said that OpenAI plans to buy Google’s AI chips, “though the financial terms and other details of the planned purchase couldn’t be learned.”
That leaves the size of the purchase, the TPU generation, the delivery dates and whether the chips would sit in OpenAI’s own data centres or in a cloud all unreported in what we could read. TPUs, short for Tensor Processing Units, are the AI chips Google designs; Alphabet said in July that it had delivered TPU systems to customer data centres for the first time in the second quarter.
Timeline of steps that put Google’s TPUs outside Google’s own cloud, as reported: Alphabet said on its 22 July call that it delivered TPU systems to customer data centres for the first time in the second quarter; Oracle and Google said on 27 October that TPUs will be offered on Oracle’s cloud from calendar year 2027; The Information reported late on 27 October that OpenAI plans to buy Google’s AI chips, with terms not learned. Sources: Alphabet Q2 2026 call; Oracle and Google joint release; The Information.
Update, 28 Oct: Bloomberg reports the plan too.Bloomberg reported that OpenAI intends to buy artificial intelligence computing chips from Alphabet’s Google, citing a person familiar with the matter, as part of the ChatGPT maker’s broader efforts to build out its AI infrastructure. That makes it a second outlet reporting the plan, after The Information.
Final terms of any deal aren’t yet known, according to the person, Bloomberg said, and OpenAI and Google declined to comment. Neither report says a deal has been signed. Bloomberg’s report beyond these points is behind its paywall and is not reflected here.
What Was Posted on X
OPENAI PLANS TO BUY GOOGLEโS AI CHIPS, THOUGH THE FINANCIAL TERMS AND OTHER DETAILS OF THE PLANNED PURCHASE COULDNโT BE LEARNED – THE INFORMATION
First Squawk relayed The Information’s report on X: OpenAI plans to buy Google’s AI chips, though the financial terms and other details of the planned purchase couldn’t be learned.
What OpenAI’s Own Job Listing Says
A listing on OpenAI’s careers page, titled “Software Engineer, TPU Compilers Workload” and based in San Francisco, sits on the company’s Inference team. The page shows no posting date; we read it on 28 October.
“We’re looking for an engineer to lead the optimization of OpenAI’s state-of-the-art frontier models on TPUs,” the listing says. “You will own performance across the full forward pass, from model code through kernels to execution on the hardware.”
The listing asks for “practical experience optimizing machine learning workloads on TPUs, through training, inference, or both” and names quantization as one technique to explore. It lists pay of “$380K – $500K + Offers Equity.”
The listing does not mention Google, a purchase or any volume. It shows OpenAI is hiring to run its models on TPUs; it does not by itself confirm the purchase The Information reported.
A Longtime Nvidia Customer
The Information’s headline calls OpenAI a longtime Nvidia customer. Oracle’s 27 October release with Nvidia said OpenAI trained and served GPT-6 Astra in Abilene on an Oracle Cloud Infrastructure supercluster with more than 100,000 Nvidia Blackwell GPUs, as we reported the same day.
On that day CNBC also reported that OpenAI’s own Jalapeño chip will run on Oracle’s cloud, as our Oracle piece noted.
Connecting the Coverage
Google’s external TPU hardware revenue as estimated by one former MediaTek employee in an expert-network interview posted on X by investor Rihard Jarc on 26 October 2026: about $8 billion in 2026, $51 billion in 2027 and around $104 billion in 2028. These are one person’s estimates, not Alphabet’s figures, and are unverified. The OpenAI purchase is The Information’s report.
On 27 October we reported that Oracle will rent out Google’s TPUs on its cloud from 2027. Oracle and Google’s joint release called it “the first time Google TPUs will be offered on a hyperscale cloud platform outside Google,” with availability in calendar year 2027. What’s new: a named AI lab, OpenAI, is now reported as planning to buy the chips; whether it would own them or use them through a cloud is not reported.
Also on 27 October we reported that an ex-MediaTek expert sees Google TPU sales at $51B in 2027. That figure is one former employee’s estimate in an expert-network interview, not Alphabet’s; the same estimate put external TPU hardware revenue at about $8 billion this year and around $104 billion in 2028.
As we read it, the listing describes TPU work across the whole stack: it asks for someone to “own performance across the full forward pass, from model code through kernels to execution on the hardware.”
As we read it, the report landed on the same day that Oracle and Google said TPUs will be offered on Oracle’s cloud, which the joint release called “the first time Google TPUs will be offered on a hyperscale cloud platform outside Google.”
As we read it, the missing numbers are the story’s main gap: The Information’s relayed sentence says the financial terms and other details “couldn’t be learned.”
“lead the optimization of OpenAI’s state-of-the-art frontier models on TPUs”
Three Implications
FOR NVIDIA The Information’s headline calls OpenAI a longtime Nvidia customer; the Oracle and Nvidia release said OpenAI trained GPT-6 Astra on more than 100,000 Nvidia Blackwell GPUs.
FOR GOOGLE Alphabet said in July that “the vast majority of the revenues from these agreements will be realized in 2027”; it reports its third quarter on 28 October.
FOR ENGINEERS OpenAI’s listing names pay of “$380K – $500K + Offers Equity” for the TPU optimization role.
The framework describes the chip layer as having “three competitive categories within one layer: generalist substrate (NVIDIA), hyperscaler custom (TPU, Trainium, MTIA), merchant custom challengers (Cerebras, Groq, SambaNova).”
As we read it, The Information’s report puts OpenAI in a position to buy from the first two categories at once. The framework was not written about this report.
What Is Not Established
We have not read The Information’s article beyond its headline, byline and the relay of one sentence, so we cannot say how many chips, at what price, over what period, or who The Information’s sources are. OpenAI and Google had not commented in anything we read.
The job listing shows TPU optimization work; it does not state a purchase. Whether OpenAI would buy the chips outright or use them through a cloud is not established in what we read.
This Week In The AI Supercycle
Execution Fault Lines, Concentrated Demand, and Geopolitical Firewalls
This week’s issue: the AI build is not short of demand for its parts. It is short of cheap money, and its end revenue rests on a narrow base of large, often venture-funded buyers.
The Information reports that OpenAI, a longtime Nvidia customer, plans to buy Google’s AI chips, with terms not learned. OpenAI’s own careers page is hiring an engineer to lead the optimization of its frontier models on TPUs.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. The planned purchase is The Information’s reporting; its article is paywalled and we read only its headline, byline and First Squawk’s relay of one sentence. The job-listing quotes are from OpenAI’s careers page as we read it on 28 October. The TPU revenue figures are one former MediaTek employee’s estimates, not Alphabet’s.
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.
