Alphabet’s (GOOG) (GOOGL) Google said its seventh generation Tensor Processing Unit, or TPU, called Ironwood, will be launched for public use in the coming weeks, CNBC reported.
The chip was unveiled in April for testing and deployment.
The chip is designed to handle tasks including training large models to powering real-time chatbots and AI agents. In connecting up to 9,216 chips in a single pod, Google said the new Ironwood TPUs eliminate “data bottlenecks for the most demanding models” and give customers “the ability to run and scale the largest, most data-intensive models in existence,” the report added.
Google did not immediately respond to a request for comment from Seeking Alpha.
While many large language models, or LLMs, and AI workloads have depended on Nvidia’s (NVDA) graphics processing units, or GPUs, Google’s TPUs come into the category of custom silicon, which can offer advantages on price, performance and efficiency, the report noted.
Google noted that Ironwood is more than four times faster than its predecessor, and major customers are already coming in. Anthropic (ANTHRO) plans to use up to 1 million of the new TPUs to run its AI Claude model, the report added.
Besides the new chip, Google is unveiling a suite of upgrades meant to make its cloud cheaper, faster, and more flexible, amid efforts to compete with larger cloud players such as Amazon’s (AMZN) Amazon Web Services and Microsoft’s (MSFT) Azure.