On February 3, 2023, Anthropic, an AI safety and research company, announced it had selected Google Cloud as its cloud provider. The collaboration is structured so the two companies can co-develop AI computing systems, with Anthropic planning to use Google Cloud’s GPU and TPU clusters to train, scale, and deploy its AI models.
Focus on Claude and safety
Anthropic is developing Claude, an AI assistant built on the company’s work toward safe, steerable AI. The company has introduced techniques such as Constitutional AI aimed at producing AI systems that are easier to trust and understand. Anthropic says the Google Cloud partnership will provide the infrastructure performance and scale needed as it makes its systems available to a larger set of users.
What Anthropic says
Dario Amodei, Anthropic’s CEO, stated that partnering with Google Cloud supports the company’s next phase, when it will deploy its AI systems more broadly. He described the partnership as providing the cloud infrastructure performance and scale Anthropic requires.
Why this matters
Strategic ties between cloud providers and AI research firms accelerate the training and deployment of large models while making infrastructure and safety capabilities central to that expansion. For Anthropic, access to Google Cloud’s specialized compute resources is intended to enable wider distribution of Claude alongside continued development of safety and interpretability features.
Timing
The partnership announcement was made on February 3, 2023.
Notable detail
Anthropic specifically mentioned using Google Cloud’s GPU and TPU clusters to support the training, scaling, and deployment of its AI systems.
