Google is taking its homegrown artificial intelligence chips beyond its own data centers, according to a report from The Information. The Alphabet unit is courting smaller cloud providers—often called "neoclouds"—with its Tensor Processing Units (TPUs), and may even help fund facilities that would otherwise stock up on Nvidia graphics processing units (GPUs).
What Are TPUs and Why Do They Matter?
Tensor Processing Units are custom-designed chips that Google created specifically for AI and machine learning tasks. Unlike general-purpose processors, TPUs are optimized for the matrix calculations that power neural networks. Google has used them internally for years to run services like Search, Gmail, and Google Photos, and they are also available through its Google Cloud platform.
Nvidia's GPUs, by contrast, have become the industry standard for AI training and inference. The company's H100 and upcoming Blackwell chips are in high demand, powering everything from OpenAI's ChatGPT to Meta's recommendation algorithms. Nvidia's market capitalization has surged past $2 trillion as a result, making it one of the most valuable companies in the world.
Google's TPUs have long been seen as a potential alternative, but they were largely confined to Google's own ecosystem. That is now changing.
What the Report Says
The Information reports that Google is actively pitching TPUs to smaller cloud providers that typically rely on Nvidia hardware. These "neocloud" providers—companies like CoreWeave, Lambda, and others—have grown rapidly by offering GPU access to AI startups and enterprises that cannot secure enough Nvidia chips from the major cloud providers.
Google may also help fund the construction of TPU-equipped data centers for these providers, according to the report. That would lower the financial barrier for smaller players to adopt Google's chips instead of Nvidia's.
The move comes as competition in the AI chip market intensifies. SpaceX's AI revenue could hit $28 billion a year from data centers, highlighting the massive opportunity in this space. Meanwhile, UK regulators are set to oversee cloud services from Microsoft, Google, Amazon, and Oracle, adding a layer of scrutiny to the market.
What It Means for Investors
For everyday investors, this story is about competition in the AI hardware market—and what it could mean for the companies involved.
For Nvidia: Any shift away from Nvidia GPUs could eventually pressure its dominant market share. Nvidia currently commands an estimated 80% or more of the AI chip market. If Google's TPUs gain traction among smaller cloud providers, it could create a viable alternative for AI workloads, potentially slowing Nvidia's growth. However, Nvidia's lead is substantial, and its software ecosystem—CUDA—remains a powerful moat.
For Alphabet (Google): This move could strengthen Google's position in the cloud computing market, where it trails Amazon Web Services and Microsoft Azure. By offering TPUs to smaller providers, Google may gain a new revenue stream and increase the adoption of its chips. It also reduces the industry's reliance on a single supplier—Nvidia—which could appeal to customers seeking diversification.
For smaller cloud providers: Access to Google's TPUs could give these companies a differentiated offering. They could pitch themselves as alternatives to the big three cloud providers, potentially attracting customers who want to avoid vendor lock-in or who prefer Google's chip architecture for specific AI tasks.
For the broader market: The AI chip market is still in its early stages. Aegis and McMaster recently launched a $3.71 million fast-charge storage project for AI data centers, underscoring the infrastructure buildout underway. As more players enter the chip space—including AMD, Intel, and startups like Cerebras and Groq—competition could drive down costs and accelerate innovation.
What to Watch Next
Investors should watch for official announcements from Google about its TPU partnerships. The Information's report is based on unnamed sources, so confirmation from the company would add credibility. Also keep an eye on Nvidia's earnings calls for any mention of competitive pressure from custom chips like TPUs.
Another key factor is the regulatory environment. Alberta has been courting Google, Meta, and Amazon for AI data centers, showing how governments are competing to attract this infrastructure. Any policy changes could affect the economics of building TPU facilities.
Finally, watch for signs that Google is expanding its chip strategy beyond TPUs. The company has also developed custom video coding chips and is reportedly working on a server chip for its data centers. A broader push into hardware could make Alphabet a more formidable competitor in the semiconductor space.
For now, Google's TPU pitch to smaller cloud providers is a reminder that the AI chip market is not a one-horse race. While Nvidia remains the leader, alternatives are emerging—and that could reshape the landscape for years to come.
