amazon announced that, in order to strengthen its ai infrastructure footprint, it will make an additional investment of up to $25 billion in anthropic. over the past several years, amazon has already invested $8 billion in this ai startup. in a statement released on monday, anthropic said it plans to spend more than $100 billion over the next decade on amazon web services’ technology offerings, including both existing and next-generation versions of amazon’s self-developed ai chip, trainium. at the same time, anthropic has secured up to 5 gigawatts of compute capacity for training and deploying its claude large language model.
in a statement, amazon ceo andy jassy said: “anthropic’s commitment to run its large language models on amazon web services’ trainium chips over the next decade underscores the successful collaboration between the two companies in the field of custom chips. we will continue to provide customers with the technology and infrastructure they need for generative ai research and development.”
the official announcement indicates that this investment includes an immediate infusion of $5 billion, with the remaining up to $20 billion to be disbursed later based on “specific commercialization milestone targets.” the initial investment is valued at approximately $380 billion, reflecting anthropic’s latest enterprise valuation.
anthropic revealed that it will complete the deployment of nearly 1 gigawatt of trainium 2 and trainium 3 chip compute capacity by the end of this year.
currently, major global cloud providers are racing to expand their ai compute reserves. in february of this year, amazon disclosed that it expects capital expenditures of about $200 billion for the full year, with the bulk of the funding going toward ai infrastructure.
just two months before this latest investment, the e-commerce giant had announced a planned investment of up to $50 billion in openai, anthropic’s main competitor. both ai firms are vying to demonstrate their strength to investors as they prepare for potential initial public offerings (ipos) as early as this year. in recent months, openai executives have repeatedly accused anthropic of having insufficient compute capacity, calling it a strategic misstep.
on monday, anthropic stated that demand for the claude model among enterprise customers and developers continues to rise, while usage among general users has also surged, putting pressure on its infrastructure and affecting both the stability and performance of its model services. this new partnership with amazon will rapidly expand its available compute capacity.
in a statement, anthropic ceo dario amodei said: “user feedback shows that the claude model has become an indispensable tool in daily work, so we must build out the supporting infrastructure to keep pace with this rapid growth in demand. our collaboration with amazon will enable us to continue pushing the frontiers of ai research while providing claude services to a broad base of users—including 100,000 enterprise customers leveraging amazon web services for development.”
founded in 2021, anthropic’s founding team consists of researchers and senior executives who previously left openai. the company is best known for its claude series of large ai models, and its commercial rollout targeting enterprises has already begun to yield results, with annualized revenue exceeding $30 billion.
in 2023, anthropic selected amazon web services as its preferred cloud service provider; in 2024, it further designated aws as its core partner for model training. at the same time, the company has also entered into cooperation agreements with rival cloud providers such as microsoft and google.
last november, microsoft announced an investment of up to $5 billion in anthropic, while anthropic committed to purchasing $30 billion worth of compute capacity on microsoft azure. earlier this month, anthropic expanded its partnerships with google and broadcom, securing several gigawatts of additional compute resources.
