Google on Thursday announced that its most powerful Tensor Processing Unit (TPU) yet — Ironwood — will be made widely available in the coming weeks, marking a major step in its bid to dominate the rapidly expanding artificial intelligence infrastructure market.
Initially introduced in April for testing and deployment, the seventh-generation Ironwood TPU is built in-house and designed to manage a wide range of AI tasks — from training large models to powering real-time chatbots and AI agents.
According to Google, a single Ironwood pod can connect up to 9,216 chips, eliminating “data bottlenecks for the most demanding models” and giving customers the “ability to run and scale the largest, most data-intensive models in existence,” CNBC reported.
The company said Ironwood is over four times faster than its predecessor, underscoring its potential to compete head-on with Nvidia’s GPU offerings, which currently dominate the AI computing landscape.
“TPUs have been in the works for a decade,” Google said, adding that AI startup Anthropic plans to use up to one million Ironwood TPUs to power its Claude model.
The launch places Google squarely in the middle of a high-stakes race alongside Microsoft, Amazon, and Meta, all racing to build the backbone of AI computing.
Unlike traditional GPUs, Google’s custom silicon promises advantages in price, performance, and energy efficiency, potentially making it an attractive alternative for AI-focused companies.
Alongside Ironwood, Google is introducing a suite of upgrades to make its cloud infrastructure cheaper, faster and more flexible as it competes with Amazon Web Services (AWS) and Microsoft Azure.
The timing aligns with robust financial momentum in Google’s cloud division. The company reported third-quarter cloud revenue of $15.15 billion, up 34% year-on-year — its fastest growth rate yet. Microsoft’s Azure saw a 40% jump, while AWS rose 20%.
Google also revealed it had signed more billion-dollar cloud deals in the first nine months of 2025 than in the previous two years combined.
To meet the surging AI demand, the company increased its capital spending forecast to $93 billion, up from $85 billion.
“We are seeing substantial demand for our AI infrastructure products, including TPU-based and GPU-based solutions,” CEO Sundar Pichai said in the company’s earnings call.
“It’s been one of the key drivers of our growth over the past year, and we continue to see very strong demand ahead,” he added.