OpenAI Is Buying Google’s AI Chips, a Hit to Nvidia’s Biggest Customer Bond

OpenAI reportedly buying Google's custom AI chips would quietly shift leverage in one of the most lucrative vendor relationships in tech, and the fallout for NVIDIA's pricing power and Alphabet's chip ambitions is only beginning to take shape.

Published October 28, 2026, 7:22am ET · 3 min read

A close-up view shows two hands, one on each side, plugging gray and black USB-like cables into a sleek, silver-gray server unit. The server unit features a prominent glowing white outline of a human brain, depicted with a network of interconnected lines and dots. In the background, blurred server racks with illuminated green and multi-colored lights suggest a large data center environment.
The symbolic act of connecting to an AI processing unit highlights the critical role of specialized chips in advancing artificial intelligence, mirroring the recent strategic shifts in the tech industry. © 24/7 Wall St.

OpenAI has spent more on NVIDIA (NASDAQ:NVDA | NVDA Price Prediction) graphics processors than almost any other company, and it now plans to buy Google’s own AI chips, according to The Information. The report came out the night of October 27, 2026. It cites one person familiar with the matter and mentions Google’s tensor processing units (TPUs), custom chips built for AI work.

Neither company has confirmed the plan. Google did not immediately respond to a request for comment, and OpenAI declined to comment.

Alphabet (NASDAQ:GOOG, NASDAQ:GOOGL) closed at $320.58, up 1.19%, a gain of $3.78, and then added 0.93% after hours to reach $323.57. NVIDIA rose 0.9% to $217.94 and stayed nearly flat after hours at $218.14.

Signed Orders Versus a Single-Source Report

No terms, volumes, timing, or chip generation have been disclosed. Until one of the companies confirms it, this remains a reported plan, since nobody has shown a signed order.

OpenAI has used TPUs before. It began renting Google’s AI chips in June 2025, and the report adds that it now owns them. A year earlier, OpenAI said it had “no specific plans to use TPUs at present.”

OpenAI has been moving in this direction for a while. It recently said it would offer its frontier models through Google Cloud. In June, it announced its first custom chip, and a current job listing asks for “significant experience optimizing real workloads on TPUs for training or inference.”

A buyer this large wants a second supplier because it gains leverage on price, protection against supply shortages, and software that is not tied to a single vendor.

How Much Alphabet and NVIDIA Have at Stake

Alphabet started selling TPUs directly to some customers this year, although its chief financial officer has said most of that revenue will be recognized in 2027. Google already rents TPUs to Anthropic, Apple (NASDAQ:AAPL) and Citadel Securities. An OpenAI purchase would show that TPUs can work as a chip business selling to outside buyers, on top of a cloud unit whose revenue grew 82% to $24.77 billion last quarter.

For NVIDIA, the first measure is scale: its data center unit brought in $89.02 billion in a single quarter. Its financial ties to OpenAI are also extensive. On August 17, 2026, NVIDIA agreed to provide up to $105 billion in credit support for an OpenAI data center in Ohio that starts at 4.25 gigawatts. NVIDIA CEO Jensen Huang reportedly said OpenAI committed to spend about $600 billion on NVIDIA servers through 2030.

NVDA price target

Why NVIDIA’s Pricing Power Faces New Pressure

OpenAI still calls itself one of NVIDIA’s biggest customers. On the same day as the report, its computing chief said, in a joint announcement with NVIDIA and Oracle (NYSE:ORCL), “We are bringing Vera into production on OCI to support these workloads at scale.”

Huang has said he is “very confident that they’re going to be customers and partners of ours for a very long time.” As the seller, he would naturally say that, but no evidence shows OpenAI cutting its orders.

The risk to NVIDIA is a slow loss of pricing power, because each credible alternative improves the buyer’s position in negotiations. NVIDIA trades at a price-to-earnings ratio of 44, and that multiple assumes NVIDIA keeps its 71.1% gross margin. TPU software still lags behind NVIDIA’s, so the purchase could be smaller or slower than the headline implies.

What Alphabet’s Valuation Gap Means for Investors

Alphabet has fallen 5.79% over the past week and is up only 2.62% this year, while NVIDIA is up 17.13%.

At a P/E of 14 and a market value of $3.874 trillion, Alphabet costs far less than NVIDIA at $5.263 trillion. That gap suggests the market gives little premium to Alphabet’s position as a credible second AI chip supplier.

GOOGL price target
GOOGL earnings explorer

The next test comes when Alphabet reports after the close on October 28, 2026. If management names outside TPU buyers or raises its 2027 TPU revenue outlook, the case gets stronger. If the call skips TPUs and the stock closes below $320.58, the market will have viewed this report as noise.

Contact [email protected] for any questions or corrections.

Omor Ibne Ehsan

Omor Ibne Ehsan is a writer at 24/7 Wall St. He is a self-taught investor with a focus on growth, cyclical, and dividend equities that have strong fundamentals, value, and long-term potential. He also has an interest in high-risk, high-reward investments such as penny stocks.

All articles →