Google has introduced two fresh AI computing systems, helping it make a major effort to meet Nvidia’s challenge in the booming AI hardware industry, as Alphabet’s shares ticked up by about 3% on Nov. 9.
Google is adding two specialized systems to its latest generation of Tensor Processing Unit (TPU) processors, the 8th generation, with this announcement. One is intended to safeguard sensitive data during the AI processing, and the other is intended to process bigger and more complex workloads of the AI. The move indicates that Google is looking to go beyond being a big buyer and user of AI chips. It also aims to become a significant player in a space that Nvidia has carved out an advantageous niche in.
The announcement is yet another indication that Google’s custom AI hardware division is playing a bigger role in Alphabet’s expansion plans as an investor.
Google Adds Two New Systems to Its AI Chip Lineup
Google has released two new systems, TPU 8tx and TPU 8i Max, to its collection of the eighth generation of TPU systems, bringing the total to four systems released over the past year. The TPUs are specialized processors that are developed by Google for its AI training and application needs, and are a direct competitor to the ever-popular Nvidia GPUs in the area of AI, under Google’s own AI infrastructure and software ecosystem.
TPU 8tx is geared towards safeguarding sensitive operational data for government and regulated enterprises, and TPU 8i Max is optimized for challenging AI workloads like coding assistance, mathematical reasoning, and drug discovery. The expansion is indicative of the demand for more sophisticated AI systems and agents, Google says.
The New Systems Target Two Growing AI Challenges
The TPU 8tx aims to safeguard delicate data during its processing by AI systems. It provides built-in security, ensuring that private data and AI models are not exposed during processing. Google says the system uses fifth-generation AMD EPYC processors coupled with Google’s TPU technology to offer up to twice the performance of its predecessor with little to no compromise in performance.
With the TPU 8i Max, the focus is on computing power: The two TPU 8i chips are packed together into a single chip, doubling the processing power, on-chip memory, and memory bandwidth. It can be used to power AI coding assistants that analyze whole software projects and research systems that handle complex scientific data.
Google’s AI Chip Business Is Becoming a Bigger Opportunity
Google’s custom chip business is also growing beyond just for Google. TPUs were previously primarily used to power products like Search, YouTube, and Gemini. The company is now offering its AI hardware to other companies via Google Cloud as a competitor to Nvidia processors, with a particular focus on the company’s new JetPack AI development platform. In Google Cloud, the company is offering its AI hardware to other enterprises in place of Nvidia processors, particularly its newest AI development platform called JetPack.
The strategy is taking off. Google Cloud revenue increased by 103% YoY to $30.71 billion in Q3 2026, which included sales of TPUs. Google claims that it now has almost 100 customers using its TPU business, ranging from financial services to retailers, to AI research companies and universities. This is another revenue stream for Google, not just from using the chips in its own services, but also from others utilizing them.
Nvidia Still Has a Major Advantage to Defend
Google’s latest announcement will make Nvidia a more competitive player, but the AI chip marketplace is still far from being a done deal. Nvidia’s well-known GPUs, software, ecosystem, and strong ties with AI developers help drive the company’s future. The company has a solid market position with its popular GPUs, software, ecosystem, and relationship with AI developers. The issue is that Google has a big head start in its control of chips, data centers, cloud services, and its own Gemini models, enabling it to develop devices that run its own AI workloads.
But if more businesses are going to move to TPUs, software compatibility, performance, availability, and cost are going to be important. Google has been trying to solve some of these problems by enabling developers to run supported PyTorch code on TPUs without having to rewrite the entire application, as part of TorchTPU. But for companies, increased competition may result in more hardware options and lower prices, but Nvidia’s dominance won’t come overnight.
Alphabet’s Stock Reaction Reflects Investor Confidence
Google’s announcement led Alphabet’s stock price to open over 3% higher on Nov. 9, according to Reuters. The gains come as investors have seen an interest in Google’s strategy of monetizing its investments in AI technologies.
But the more interesting question is whether Google can appeal to more TPU buyers and be competitive with Nvidia outside of Google Cloud. Google has made further strides with its new chips towards making its custom AI hardware a viable business option for businesses developing and operating sophisticated AI systems.
