Amazon is reportedly mulling the idea of further investment in artificial intelligence (AI) company Anthropic.
The companies are also partnering up, along with Palantir, to provide advanced AI models to US defense and intelligence agencies.
As reported by The Information and citing a "person familiar with the matter," Amazon and Anthropic are reportedly in talks for a major financial investment that could see Amazon investing "multiple billions" in the AI start-up.
According to the report, the investment is contingent on Anthropic using Amazon Web Services' (AWS) own-developed silicon that is hosted on AWS to train its AI models.
Currently, Anthropic uses Nvidia chips for training.
Nvidia has led the pack in developing GPUs throughout the AI boom, and counts cloud providers among its largest customers.
The likes of Amazon, Google, and Microsoft are all exploring and developing custom AI chips, however.
Amazon currently offers its Tranium and Inferentia chips for AI - for training and inferencing respectively - as well as its Graviton range of CPUs.
This would by no means be Amazon's first investment in Anthropic. The cloud giant invested $1.25 billion in the start-up in September 2023, and another $2.73 billion in March 2024.
The investment saw the UK's Competitions and Markets Authority (CMA) launch and investigation into the partnership in August, just a few months after its wider investigation of the AI sector and its key players. The investigation directed at Amazon's investment in Anthropic was dropped one month later.
AWS, Anthropic, and Palantir team up for defense AI
AWS and Anthropic are also working with Palantir Technologies to offer AI services to the US defense and intelligence agencies.
Anthropic's Claude 3 and 3.5 AI models will now be directly integrated with Palantir's platform and will use AWS for "secure cloud hosting."
The Claude models will reportedly be used to help improve intelligence analysis and "boost" operational efficiency by automatic complex data processing. Anthropic has also previously launched its Claude Models on AWS' GovCloud platform.
“Our partnership with Anthropic and AWS provides US defense and intelligence communities the toolchain they need to harness and deploy AI models securely, bringing the next generation of decision advantage to their most critical missions," said Shyam Sankar, CTO of Palantir.
“Palantir is proud to be the first industry partner to bring Claude models to classified environments. We’ve already seen firsthand the impact of these models with AIP in the commercial sector: for example, one leading American insurer automated a significant portion of their underwriting process with 78 AI agents powered by AIP and Claude, transforming a process that once took two weeks into one that could be done in three hours. We are now providing this same asymmetric AI advantage to the US government and its allies."
Palantir Technologies has similarly signed agreements with the likes of Microsoft and Oracle this year to provide cloud and AI services to US defense and intelligence agencies.
Palantir was launched in 2003 by Peter Thiel, and builds software to help organizations create and run artificial intelligence across private and public networks. An early investor of the company was the CIA through its venture capital branch, In-Q-Tel.
Documents provided by NSA whistleblower Edward Snowden suggested that Palantir played a role in creating the US government's international spy program, PRISM. Palantir denies any connection.
Earlier this year, Palantir agreed to a 'strategic partnership' with the Israel Defense Ministry to supply technology to support the war in Gaza.
