Anthropic and Amazon have struck a massive new deal. Amazon will invest a fresh $5 billion in Anthropic, bringing its total investment in the AI company to $13 billion. In return, Anthropic has committed to spending over $100 billion on AWS over the next decade, securing up to 5 gigawatts of computing capacity to train and run Claude.
The deal is one of the largest AI infrastructure commitments ever announced and signals just how deeply intertwined Anthropic's future is with Amazon's cloud platform.
How the Deal Is Structured
This is not a straightforward cash investment. Like many of the mega-deals shaping the AI industry, it is structured partly as cloud infrastructure services rather than direct capital. Amazon provides both funding and compute capacity, while Anthropic commits to spending that money and much more on AWS services over the next ten years.
At the center of the arrangement are Amazon's custom chips: Graviton, a low-power CPU, and Trainium, an AI accelerator chip designed to compete with Nvidia. The deal specifically covers Trainium2 through Trainium4 chips, even though Trainium4 is not yet available. The latest generation, Trainium3, was released in December. Anthropic has also secured the option to buy capacity on future Amazon chips as they become available.
This chip commitment is strategically significant. It means Anthropic is locking in access to next-generation AI hardware years before it reaches the market a move that provides supply certainty in an industry where compute capacity is one of the most valuable and constrained resources.
Echoes of the OpenAI-Amazon Deal
The deal mirrors an agreement Amazon struck with OpenAI just two months ago. In that arrangement, Amazon joined a $110 billion funding round contributing $50 billion that valued OpenAI at $730 billion pre-money. That deal was also structured partly as cloud services rather than straight cash.
Amazon is now the largest single investor in both of the AI industry's two dominant companies a position that gives it extraordinary leverage over the compute infrastructure underpinning the entire enterprise AI market. Whether backing both sides of the OpenAI-Anthropic rivalry creates conflicts of interest or simply smart portfolio diversification is a question investors and regulators are likely to examine closely.
Anthropic's Growing Scale
The Amazon investment comes at a moment of extraordinary momentum for Anthropic. The company's annualized revenue recently hit $30 billion, surpassing OpenAI's reported $24-25 billion figure. Claude Code has become the dominant AI coding tool among enterprise users, and the company recently launched Claude Design for visual creation.
Anthropic has also been navigating complex political dynamics, including a Pentagon supply-chain risk designation and a thawing relationship with the broader Trump administration following briefings on its Mythos model.
VCs have reportedly been offering Anthropic capital at an $800 billion valuation or higher, though the company has so far declined those offers. The Amazon deal could serve as a precursor to a formal funding round at or above that level.
The 5 Gigawatt Question
The scale of compute Anthropic is securing is staggering. Five gigawatts of computing capacity is enough to power multiple massive data center campuses the kind of infrastructure that entire countries are struggling to build. For context, OpenAI's Stargate project aims to deploy up to $500 billion in AI infrastructure. Anthropic is now locking in compute capacity on a comparable scale through its AWS partnership.
The energy requirements alone are significant. Five gigawatts is roughly equivalent to the output of five large nuclear power plants a reminder that the AI industry's appetite for power is becoming one of its most pressing constraints.
What It Means
The deal cements Anthropic's position as one of the two companies at the center of the AI revolution, alongside OpenAI. It also cements Amazon's position as the infrastructure provider that both companies depend on a role that may prove even more valuable than the AI models themselves.
For the broader AI industry, the deal is another signal that the era of trillion-dollar commitments has arrived. The companies building the most capable AI models are now making decade-long infrastructure bets measured in hundreds of billions of dollars. Whether those bets pay off will depend on whether AI delivers the transformative economic value its proponents promise but for now, the scale of the wagers keeps growing.
