[News] OpenAI Reportedly Plans to Buy Google TPUs; Oracle to Bring TPUs to Its Cloud in 2027, Adding Pressure on NVIDIA

Google’s push to expand its AI chip business could gain momentum as OpenAI reportedly considers purchasing its TPUs. According to Reuters, citing The Information, OpenAI, one of NVIDIA’s largest GPU customers, plans to purchase Google’s AI chips, sources say. The potential deal would deepen ties between OpenAI and Alphabet’s Google, despite their competition in AI products and search. Financial terms and other details remain undisclosed.

The reported move builds on an existing relationship between the two companies. Reuters reported in June 2025 that OpenAI had begun renting Google’s AI chips.

The cooperation may now extend beyond chip rentals. As noted by BigGo Finance, OpenAI has posted a job opening for an engineer to lead efforts to optimize its most advanced frontier models for Google’s Tensor Processing Units (TPUs). The recruitment suggests OpenAI is preparing for deeper integration with Google’s chip architecture, the report indicates.

Google’s TPUs are gaining ground as an alternative to NVIDIA’s dominant AI accelerators, as companies look for more cost-effective computing solutions and chips tailored to AI workloads. Google also began offering TPUs for direct purchase to select customers this year, Reuters notes.

Despite its growing interest in Google’s chips, OpenAI maintains substantial financial and infrastructure ties with NVIDIA. According to BigGo Finance, NVIDIA agreed in August to provide more than $100 billion in credit support for OpenAI’s planned data center project in Ohio, which is expected to rely exclusively on NVIDIA’s chip architecture. NVIDIA CEO Jensen Huang also said in a blog post that OpenAI’s existing and planned deployments could represent approximately $600 billion in NVIDIA compute through 2030.

Google Expands TPU Availability Through Oracle Cloud

Meanwhile, Google is expanding its TPU business through a partnership with Oracle. Digital Today, citing The Information, reports that the companies have announced plans to offer Google’s AI chips to Oracle Cloud customers, marking the first time TPUs will be available through a cloud platform outside Google’s own. Oracle plans to introduce the chips to its public cloud service, Oracle Cloud Infrastructure (OCI), in 2027.

The partnership could also reach a substantial scale. Speaking to Data Center Dynamics, Oracle Cloud Infrastructure Senior Vice President Karan Batta described the agreement as a multi-year, multi-gigawatt partnership rather than an experimental deployment. He added that Oracle aims to offer new Google TPU generations as soon as they become available. Meanwhile, according to BigGo Finance, sources indicate that Oracle will purchase the chips directly, with a major customer shared by Google and Oracle expected to be the end user.

Oracle is also expanding its partnership with NVIDIA, with plans to deploy Vera Rubin systems at multi-gigawatt scale, Reuters reports. The cloud provider is adding AI accelerators from Cerebras and SambaNova as well, further diversifying its chip offerings.

Read more

• [News] MediaTek’s Google TPU v9 Shipments May Slip to 1Q28; ASIC Race Heats Up as Taiwan Firms Eye Opportunities
• [News] NVIDIA’s First Investment in a Taiwan-Listed Firm in Spotlight: What the MediaTek Deal Means for the AI Race

(Photo credit: Google)