OpenAI, the company behind ChatGPT, is reportedly preparing to buy Google's Tensor Processing Units (TPUs), according to The Information. The report, published Tuesday, suggests the AI giant is looking to add Google's custom chips to its arsenal, a notable shift for a company that has been one of Nvidia's biggest customers for graphics processing units (GPUs).
TPUs are Google's custom-designed chips built specifically for AI workloads. Unlike Nvidia's GPUs, which are general-purpose processors that happen to excel at AI, TPUs are tailored from the ground up for machine learning tasks. They are a key part of Google's cloud infrastructure and are also used to power its own AI models.
The potential purchase would deepen ties between two companies that are, in many ways, rivals. OpenAI and Google compete head-to-head in AI products, from chatbots to search. Yet they also increasingly depend on each other: OpenAI needs massive computing power, and Google has both the chips and the cloud infrastructure to provide it.
Why OpenAI is diversifying its chip supply
For years, Nvidia has dominated the AI chip market, with its GPUs becoming the de facto standard for training and running large language models. But that dominance has created a bottleneck. Demand for Nvidia's chips has far outstripped supply, leading to long lead times and soaring costs. Big tech companies have been pouring billions into data centers and custom silicon to secure enough computing power for AI, and this deal fits squarely into that scramble.
Reuters reported in June 2025 that OpenAI had already started renting Google's AI chips. Buying them outright would represent a deeper commitment. For OpenAI, adding TPUs alongside Nvidia hardware could diversify its supply chain and potentially cut costs. Renting chips from cloud providers is expensive, and owning hardware can be more economical over the long run, especially for a company with OpenAI's scale.
The move also gives OpenAI leverage in negotiations with Nvidia. By signaling it has alternatives, OpenAI may be able to secure better pricing or priority access to Nvidia's latest chips. This is a common strategy in industries where a single supplier holds enormous power.
What it means for investors
For everyday investors, this story is a reminder that the AI boom is not just about Nvidia. While Nvidia has been the clear winner so far, the landscape is shifting. Companies like Google, Amazon, and Microsoft are all developing their own AI chips, and cloud providers are increasingly offering alternatives to Nvidia's GPUs. This could mean more competition in the AI hardware market, which might eventually put downward pressure on prices.
For Google's parent company, Alphabet, selling TPUs to OpenAI would be a win. It would turn a competitor into a customer, generating revenue from a business that is otherwise trying to take market share from Google's search and AI products. It also validates Google's chip design efforts, which have often been overshadowed by Nvidia's dominance.
For OpenAI, the deal is part of a broader strategy to secure the resources it needs to keep scaling. The company has been in the news for its massive fundraising efforts, including reports of a $30 billion bridge round at a $1.25 trillion valuation ahead of a possible IPO. It has also been expanding its revenue streams, such as rolling out ads on ChatGPT to 89 countries. Securing a reliable and cost-effective chip supply is critical to that growth story.
The broader trend here is the massive increase in data center spending. Companies across the tech sector are building out infrastructure at an unprecedented pace, and that spending shows no signs of slowing. This has implications for a wide range of companies, from chipmakers to cloud providers to the firms that build and cool data centers. For instance, Oracle has announced plans to offer Google TPUs on its cloud by 2027, and Oracle is expanding its AI server lineup beyond Nvidia with chips from Google and SambaNova. Even China's lab-grown diamond makers are pivoting to AI chip cooling, a sign of how far the ripple effects extend.
For investors, the key takeaway is that the AI chip market is becoming more complex. Nvidia's dominance is not guaranteed to last forever, and the companies that can secure diverse and cost-effective computing power will be better positioned. As always, it's important to focus on the long-term fundamentals rather than getting caught up in the day-to-day headlines.
Neither OpenAI nor Google has officially confirmed the reported purchase, so the details could change. But the direction is clear: the race for AI computing power is intensifying, and even the biggest players are looking for every advantage they can get.
