Amazon.com (AMZN.O) said on Monday that it will invest up to $4 billion in cash in the well-known firm Anthropic to compete with burgeoning cloud rivals in artificial intelligence.

As part of the agreement, Anthropic’s technology will get early access to Amazon’s workers and cloud customers, allowing them to integrate it into their organizations. The San Francisco-based business also vowed to rely heavily on Amazon’s cloud services, including training its next AI models on a sizable number of custom chips it would purchase from the online retail and computing behemoth.

The CEOs of Amazon’s cloud business and Anthropic announced in a joint interview that there will be an immediate investment of $1.25 billion, with either partner having the power to prompt a further $2.75 billion in funding from Amazon.

They chose not to disclose the startup’s increased valuation, which was previously estimated at more than $4 billion, or how much of Anthropic Amazon would now hold. Amazon said that its interest was a minority and that it would not be given a board seat.

The announcement may be Amazon’s largest response to smaller cloud competitors Microsoft (MSFT.O) and Alphabet’s (GOOGL.O) Google, who have sold or created potent AI this year. The agreement demonstrates the cloud firms’ continuous efforts to forge partnerships with the AI startups transforming their sector.

Since 2019, Microsoft has invested billions of dollars in its collaboration with OpenAI. This company created ChatGPT, offering its clients exclusive access to the prose-writing and image-generating technologies of the startup.

In the meantime, Google supported the development of this area of AI and contributed to Anthropic’s $450 million financing in May, according to the firm.

The arrangement struck on Monday may increase demand for Amazon’s products, particularly processors supporting AI. Anthropic is committed to creating technologies, for example, for Amazon’s Trainium and Inferentia chips.

AWS CEO Adam Selipsky stated that the agreement “will help make Anthropic’s models better, as well as our chip technology and AI infrastructure better.”

The chief executive officer of Anthropic, Dario Amodei, said that his organization “has obtained the money in a way that allows it to prioritize safety” and “allows us to continue to scale up our models.”

RELATED TO GOOGLE STILL

Anthropic is one of several businesses developing so-called generative AI, or systems that can produce material like a person created. Former OpenAI employees, including Amodei, started Anthropic. Anthropic has tried to set itself apart by teaching AI to uphold moral principles.

Amodei declined to comment on how competitive its most recent funding was.

According to him, Anthropic is sticking with the deal it struck with Google in February. Anthropic had planned to make their technology available through Google Cloud and other channels while employing Google’s special processors.

However, with the arrangement announced on Monday, Anthropic is supporting Amazon Bedrock, a platform that has drawn thousands of customers to begin developing AI applications.

Customers of Amazon will get early access to Anthropic features, including the opportunity to modify their AI. According to Selipsky, “both companies are committing that future versions of Claude will be available on Amazon Bedrock for many years to come, and that’s important.”

The Claude 2 AI model from Anthropic is designed to assess lengthy commercial or legal texts since it can respond to unusually big cues.

The agreement said to Amodei “allows us to work more closely to drive enterprise usage for Claude,” which he claimed constituted the majority of Bedrock’s up to this point demand.

According to Amodei, the data analytics business LexisNexis is collaborating with Anthropic and Amazon to improve the “intelligence” of its legal search functions. He listed Lonely Planet and Bridgewater Associates as more clients.

Anthropic has not yet achieved the name recognition or use of OpenAI, the firm that created ChatGPT, one of the world’s fastest-growing software apps, and the creator of the GPT-4 model.

To minimize the need for consumers to search elsewhere for cloud services, Amazon has worked to provide a wide variety of AI models.

When asked whether Amazon might invest in other AI firms besides Anthropic, Selipsky replied, “I honestly don’t know what the future will hold.”