OpenAI May Buy Google's AI Chips as $600 Billion Spending Boom Challenges Nvidia

OpenAI reportedly plans to purchase artificial intelligence chips directly from Google, potentially giving Nvidia's biggest rival a valuable customer as technology companies prepare to spend hundreds of billions on AI infrastructure.

According to The Information, which cited a person familiar with the matter on October 27, OpenAI intends to buy Google's tensor processing units, or TPUs. Financial terms, quantities and delivery schedules remain undisclosed.

The possible purchase would deepen an unusual relationship between two AI competitors. OpenAI relies heavily on Nvidia's GPUs, while Google increasingly offers its own processors to external customers seeking alternatives.

The report comes as Amazon, Microsoft and Alphabet plan combined AI infrastructure spending of as much as $600 billion this year, raising questions about which chipmakers will capture that investment.

Google Could Win One of Nvidia's Biggest AI Customers

OpenAI already has experience with Google's infrastructure.

Reuters reported in June 2025 that the ChatGPT developer began renting Google AI chips, while OpenAI recently announced plans to distribute its advanced models through Google Cloud.

Purchasing processors directly would represent a different commercial arrangement, potentially giving OpenAI greater control over how its computing infrastructure is deployed.

The company is also recruiting an engineer to optimize frontier models for TPUs, according to The Information. However, the listing does not establish that a hardware purchase has been completed.

Google's chips are designed for intensive machine-learning calculations, offering an alternative to Nvidia's more broadly programmable GPUs.

The comparison of custom AI chips explained why specialized processors can offer efficiency advantages for particular workloads without necessarily replacing GPUs everywhere.

Google's expanding supplier network also includes Marvell, whose TPU-related agreement could generate up to $120 billion in cumulative sales through 2033 if its conditions are met.

Oracle Adds Google TPUs as Nvidia Competition Expands

The timing is particularly significant because Oracle announced plans to offer Google's TPU infrastructure through its cloud platform.

Oracle has simultaneously expanded its Nvidia relationship, demonstrating that major cloud providers are adopting multiple chip technologies rather than relying exclusively on one supplier.

OpenAI is pursuing a similar strategy.

In June, the company unveiled its first custom AI processor developed with Broadcom, with deployment planned before the end of 2026.

That development followed increasing interest in Broadcom's inference chips, which are designed to handle the enormous computing requirements of running AI models.

Meanwhile, Anthropic has already diversified across Nvidia, Google and Amazon infrastructure, with Broadcom financing potentially supporting billions in additional TPU capacity.

The competition is not simply about processor performance. Customers must also consider software compatibility, availability, electricity consumption and total operating costs.

$600 Billion AI Spending Boom Raises Stakes for Nvidia

The potential shift comes amid extraordinary investment commitments.

Amazon, Microsoft and Alphabet could collectively spend nearly $600 billion on AI infrastructure during 2026, as examined in Coinpaper's recent AI spending analysis.

Nvidia remains deeply connected to that expansion.

Its strategic relationship with OpenAI includes plans to deploy at least 10 gigawatts of Nvidia systems, alongside a previously announced investment framework worth up to $100 billion.

However, OpenAI's growing supplier relationships suggest that future demand may be distributed across more companies.

The question of whether enormous AI investments generate sufficient returns is also becoming more important as infrastructure costs accelerate.