Anthropic, the AI safety and research company, will leverage Google’s cutting-edge AI chips in a multi-year deal potentially worth tens of billions of dollars to power the training of its Claude chatbot, Reuters has learned.The agreement marks a significant win for Google as it seeks to establish its Tensor Processing Units (TPUs) as the industry standard for demanding AI workloads and underscores anthropic’s rapid growth and escalating computational needs.
The partnership addresses a critical bottleneck in AI development: access to sufficient and advanced computing power. As large language models like Claude grow in complexity, the cost and availability of specialized hardware become paramount.This deal allows Anthropic to accelerate its AI research and deployment, competing more effectively wiht industry leaders like OpenAI and meta, while simultaneously solidifying Google’s position in the burgeoning AI infrastructure market.
Anthropic will utilize Google Cloud’s TPUs – specifically designed for machine learning – over the coming years. While the exact financial terms remain undisclosed, people familiar with the agreement estimate the total value could exceed $10 billion, potentially reaching tens of billions depending on Anthropic’s scaling needs and future TPU generations.
The move comes as demand for AI chips surges, fueled by the rapid advancement of generative AI. Nvidia currently dominates the market, but google is aggressively pushing its TPUs as a competitive alternative, emphasizing their performance and cost-effectiveness for specific AI tasks. Anthropic’s decision to adopt TPUs represents a major endorsement of Google’s technology.
anthropic, founded by former OpenAI researchers, is focused on building reliable, interpretable, and steerable AI systems. Claude is designed to be a helpful, harmless, and honest AI assistant, and the increased computational power will enable Anthropic to refine its model and expand its capabilities. The company recently secured a $4.1 billion investment led by Amazon.