OpenAI Plans to Buy Google AI Chips, Diversifying Beyond Nvidia in Strategic Hardware Shift

9 Sources

Share

OpenAI plans to purchase Google's Tensor Processing Units in a move to diversify AI hardware supply beyond its heavy reliance on Nvidia. The deal strengthens commercial ties between the AI competitors as tech giants collectively spend up to $600 billion on AI infrastructure in 2026.

OpenAI Shifts Hardware Strategy with Google TPU Purchase

OpenAI plans to buy Google's AI chips, according to

2

citing a person familiar with the matter reported by The Information. The planned purchase of Tensor Processing Units marks a significant shift for the ChatGPT maker, which has been one of Nvidia's largest customers. Financial terms, quantities, and delivery schedules remain undisclosed, with both

1

declining to comment on the report.

Source: Decrypt

Source: Decrypt

The move comes as technology companies ramp up spending on data centers and semiconductors to meet surging AI infrastructure demand.

3

reports that Alphabet, Amazon, and Microsoft are collectively expected to spend as much as $600 billion on AI infrastructure during 2026, attracting investor scrutiny about when these investments will generate sufficient financial returns.

Growing Commercial Ties Between AI Competitors

The planned purchase deepens commercial ties between OpenAI and Alphabet's Google, which compete directly in AI products and search. OpenAI recently announced that its frontier AI models would become available on Google Cloud, following a similar move with Amazon Web Services.

4

notes that Google Cloud also began offering OpenAI's GPT-6 models to select customers in private preview last week.

OpenAI began renting Google's TPU chips through Google Cloud in June 2025, according to

5

. However, the company initially described it as early testing with no plans to deploy them at scale. A direct purchase would mean owning the hardware instead of paying Google by the hour, representing a further step in its use of hardware from a competing company.

Reducing Reliance on Nvidia's Dominant Position

OpenAI has historically relied heavily on Nvidia's graphics processing units to train AI models and run inference computing. The company uses these GPUs to develop and operate its AI systems, including the models that power ChatGPT and other services. In September 2025, Nvidia signed a letter of intent to invest up to $100 billion in OpenAI as it deployed at least 10 gigawatts of Nvidia systems, though the actual investment ended up at $30 billion.

Source: AIM

Source: AIM

Google's TPU chips have gained traction as an alternative to Nvidia's market-leading AI accelerators as companies seek cheaper options and chips suited to running AI models. Google began selling TPUs directly to some customers this year, after previously renting them exclusively through Google Cloud. The processors are built specifically for machine learning workloads, designed for the mathematical operations that AI models run on.

OpenAI's Multi-Pronged Hardware Approach

The Google TPU purchase represents just one element of OpenAI's diversified hardware strategy.

4

reports that OpenAI unveiled Jalapeño, its first custom AI chip built with Broadcom, in June 2026. The company plans to deploy this processor before the end of 2026 as part of a multi-generation effort to develop specialized computing hardware designed for inference workloads.

OpenAI is actively hiring an engineer to lead the optimization of its frontier models on TPUs, indicating serious intent to adapt its models to Google's hardware. Richard Ho, who leads OpenAI's hardware program, spent nearly nine years at Google and was one of the leaders of its Cloud TPU project before joining OpenAI, providing internal expertise for the transition.

Two weeks after the Nvidia investment agreement, AMD gave OpenAI the right to buy up to 160 million AMD shares, roughly 10% of the company. The right is tied to OpenAI deploying 6 gigawatts of AMD chips, further demonstrating the company's strategy to diversify AI computing resources across multiple suppliers.

Google's Expanding TPU Business

Google's TPU business could benefit significantly from OpenAI's demand. Alphabet's CFO Anat Ashkenazi has said the company expects to book a small share of TPU sales revenue by the end of 2026, with most revenue coming in 2027.

2

notes that demand has grown so much that Google's own researchers have had to wait for access to TPU capacity.

Source: TechRadar

Source: TechRadar

Google unveiled two eighth-generation TPUs in April: the 8t for training AI models and the 8i for inference. Anthropic, OpenAI's main rival, signed an agreement with Google and Broadcom in April for multiple gigawatts of next-generation TPU capacity expected to come online starting in 2027. Oracle and Google also recently signed a deal to bring TPUs to Oracle Cloud Infrastructure from 2027, expanding the supply chain for Google's processors beyond its own cloud platform.

Implications for AI Hardware Markets

The reported deal signals a maturing AI hardware market where leading developers are pursuing multiple sources of computing capacity rather than depending on a single supplier. This shift could reshape competitive dynamics in the AI accelerator space, potentially reducing Nvidia's pricing power while validating Google's long-term investment in custom chip development.

Investors will watch closely whether OpenAI's hardware diversification improves its cost structure for training and deploying increasingly expensive AI models. The strategy of combining purchases from Nvidia, Google, and AMD with custom chip development from Broadcom suggests OpenAI is optimizing for both performance and supply chain resilience as it scales operations.

Alphabet reports its quarterly results today, with investors looking for signs that its heavy investment in AI infrastructure continues to drive growth in its cloud business and nascent TPU sales operation.

© 2026 TheOutpost.AI All rights reserved