Amazon is making one of its most aggressive artificial intelligence moves yet, committing up to $25 billion in additional investment in AI startup Anthropic, while securing a long-term cloud and computing commitment that could reshape the competitive landscape for generative AI infrastructure.
The expanded partnership signals how central AI has become to Amazon’s future strategy, particularly through its cloud division, Amazon Web Services, which continues to battle rivals for dominance in the rapidly expanding AI infrastructure race.
Under the new agreement, Amazon will immediately inject $5 billion into Anthropic, with an additional $20 billion available if agreed commercial milestones are met.
This builds on the $8 billion Amazon has already invested in the startup, significantly deepening its stake in one of the most closely watched AI companies in the world.
In return, Anthropic has committed to spending more than $100 billion on AWS services over the next decade, creating one of the largest long-term cloud infrastructure partnerships in the AI sector.
At the centre of the deal is computing power.
Anthropic, the company behind the Claude AI models, will increasingly rely on Amazon’s custom-built AI chips, known as Trainium, to train and deploy its systems.
The agreement includes plans for up to five gigawatts of computing capacity over time, with one gigawatt expected to come online by the end of this year using Trainium2 and Trainium3 chips.
This scale of infrastructure is critical for training advanced large language models, which require enormous computing resources to process data and improve performance.
The partnership also strengthens Amazon’s position in AI hardware. By securing a major long-term customer for its custom silicon, the company is betting that in-house chip development will help it compete more effectively against rivals who rely heavily on third-party processors.
Already, more than 100,000 customers are using Anthropic’s Claude models through AWS, highlighting how deeply integrated the startup has become within Amazon’s cloud ecosystem.
The collaboration also aims to simplify access for enterprise users. Businesses will be able to use Claude directly through their AWS accounts, allowing them to integrate AI tools with existing billing systems, security controls, and monitoring infrastructure without needing separate agreements.
Amazon has been investing heavily in expanding its global data centre footprint, with capital expenditure expected to reach around $200 billion this year. A significant portion of that spending is directed toward AI infrastructure, reflecting the rapid growth in demand for advanced computing.
Amazon CEO Andy Jassy said the partnership demonstrates the progress both companies have made in developing custom AI systems. He emphasised that Amazon’s silicon strategy is designed to deliver higher performance at lower cost, a key advantage in the increasingly competitive AI market.
Anthropic CEO and co-founder Dario Amodei said demand for its Claude models is accelerating rapidly, with users increasingly relying on the system for everyday work.
He added that the expanded partnership with Amazon would allow the company to scale its research and infrastructure to meet growing global demand.
Following the announcement, Amazon’s stock rose in after-hours trading, reflecting investor optimism about its deepening role in AI infrastructure and its dominance in cloud computing.
As competition intensifies among major tech players, this deal positions Amazon not just as a cloud provider but as a central force powering the next generation of artificial intelligence systems.
