Anthropic plans to use the latest TPU chip 'at scale' to help train its generative artificial intelligence models, including its Claude large language model (LLM).
The company, which counts Google as an investor, will use the v5e variant of Google's custom AI chip family. The new chip is now generally available on Google Cloud.
"This expanded partnership with Anthropic, built on years of working together, will bring AI to more people safely and securely, and provides another example of how the most innovative and fastest growing AI startups are building on Google Cloud," Thomas Kurian, CEO of Google Cloud, said.
Google acquired a 10 percent stake in Anthropic for $300 million back in late 2022. Last month, the tech giant invested another $500m and promised to invest a further $1.5bn for an unknown stake.
Anthropic has built one of the largest Google Kubernetes Engine (GKE) clusters in the AI sector, uses Google's fully managed PostgreSQL-compatible database AlloyDB, and Google's BigQuery data warehouse.
However, the company is not an exclusive user of Google Cloud, as OpenAI is with Microsoft Azure. In September, rival Amazon said that it would invest up to $4bn in Anthropic, and become its primary cloud provider.
The two companies said that they would also collaborate on the development of future Trainium and Inferentia chips.
