Amazon on Friday announced it would invest an additional $4 billion in Anthropic, the artificial intelligence startup founded by ex-OpenAI research executives.
The new funding brings the tech giant’s total investment to $8 billion, though Amazon will retain its position as a minority investor, according to Anthropic, the San Francisco-based company behind the Claude chatbot and AI model.
Amazon Web Services will also become Anthropic’s “primary cloud and training partner,” according to a blog post. From now on, Anthropic will use AWS Trainium and Inferentia chips to train and deploy its largest AI models.
Anthropic is the company behind Claude — one of the chatbots that, like OpenAI’s ChatGPT and Google’s Gemini, has exploded in popularity. Startups like Anthropic and OpenAI, alongside tech giants such as Google, Amazon, Microsoft and Meta, are all part of a generative AI arms race to ensure they don’t fall behind in a market predicted to top $1 trillion in revenue within a decade. Some, like Microsoft and Amazon, are backing generative AI startups with hefty investments as well as working on in-house generative AI.
The partnership announced Friday will also allow AWS customers “early access” to an Anthropic feature: the ability for an AWS customer to do fine-tuning with their own data on Anthropic’s Claude. It’s a unique benefit for AWS customers, according to a company blog post.
In March, Amazon’s $2.75 billion investment in Anthropic was the company’s largest outside investment in its three-decade history. The companies announced an initial $1.25 billion investment in September 2023.
Amazon does not have a seat on Anthropic’s board.
News of Amazon’s additional investment comes one month after Anthropic announced a significant milestone for the company: AI agents that can use a computer to complete complex tasks like a human would.
Anthropic’s new Computer Use capability, part of its two newest AI models, allows its tech to interpret what’s on a computer screen, select buttons, enter text, navigate websites and execute tasks through any software and real-time internet browsing.
The tool can “use computers in basically the same way that we do,” Jared Kaplan, Anthropic’s chief science officer, told CNBC in an interview last month, adding it can do tasks with “tens or even hundreds of steps.”
Amazon had early access to the tool, Anthropic told CNBC at the time, and early customers and beta testers included Asana, Canva and Notion. The company had been working on the tool since early this year, according to Kaplan.
In September, Anthropic rolled out Claude Enterprise, its biggest new product since its chatbot’s debut, designed for businesses looking to integrate Anthropic’s AI. In June, the company debuted its more powerful AI model, Claude 3.5 Sonnet, and in May, it rolled out its “Team” plan for smaller businesses.
Last year, Google committed to invest $2 billion in Anthropic, after previously confirming it had taken a 10% stake in the startup alongside a large cloud contract between the two companies.
Dell Technologies CEO Michael Dell said Tuesday that while demand for computing power is “tremendous,” the production of artificial intelligence data centers will eventually top out.
“I’m sure at some point there’ll be too many of these things built, but we don’t see any signs of that,” Dell said on “Closing Bell: Overtime.”
The hardware maker’s server networking business grew 58% last year and was up 69% last quarter, Dell said. As large language models have evolved to more multimodal and multi-agent systems, the demand for AI processing power and capacity has continued to be strong.
Read more CNBC tech news
Dell’s AI servers are powered by Nvidia‘s Blackwell Ultra chips. The company then sells its devices to customers like cloud service provider CoreWeave and xAI, Elon Musk’s startup.
Dell shares rose over 3% Tuesday after increasing its expected long-term revenue and profit growth in an analyst meeting.
The computer maker raised its expected annual revenue growth to 7% to 9%, up from its previous target of 3% to 4%, with diluted earnings per share now expected to be 15% higher, up from its previous 8% target.
The company reported strong second-quarter earnings in August, and said it planned to ship $20 billion worth of AI servers in fiscal 2026. That is double what it sold last year.
The Motion Picture Association on Monday urged OpenAI to “take immediate and decisive action” against its new video creation model Sora 2, which is being used to produce content that it says is infringing on copyrighted media.
Following the Sora app’s rollout last week, users have been swarming the platform with AI-generated clips featuring characters from popular shows and brands.
“Since Sora 2’s release, videos that infringe our members’ films, shows, and characters have proliferated on OpenAI’s service and across social media,” MPA CEO Charles Rivkin said in a statement.
OpenAI CEO Sam Altman clarified in a blog post that the company will give rightsholders “more granular control” over how their characters are used.
But Rivkin said that OpenAI “must acknowledge it remains their responsibility – not rightsholders’ – to prevent infringement on the Sora 2 service,” and that “well-established copyright law safeguards the rights of creators and applies here.”
OpenAI did not respond to a request for comment.
Concerns erupted immediately after Sora videos were created last week featuring everything from James Bond playing poker with Altman to body cam footage of cartoon character Mario evading the police.
Although OpenAI previously held an opt-out system, which placed the burden on studios to request that characters not appear on Sora, Altman’s follow-up blog post said the platform was changing to an opt-in model, suggesting that Sora would not allow the usage of copyrighted characters without permission.
However, Altman noted that the company may not be able to prevent all IP from being misused.
“There may be some edge cases of generations that get through that shouldn’t, and getting our stack to work well will take some iteration,” Altman wrote.
Copyright concerns have emerged as a major issue during the generative AI boom.
Disney and Universal sued AI image creator Midjourney in June, alleging that the company used and distributed AI-generated characters from their films and disregarded requests to stop. Disney also sent a cease-and-desist letter to AI startup Character.AI in September, warning the company to stop using its copyrighted characters without authorization.
Thoma Bravo co-founder Orlando Bravo said that valuations for artificial intelligence companies are “at a bubble,” comparing it to the dotcom era.
But one key difference in the market now, he said, is that large companies with “healthy balance sheets” are financing AI businesses.
Bravo’s private equity firm boasts more than $181 billion in assets under management as of June, and focuses on buying and selling enterprise tech companies, with a significant chunk of its portfolio invested in cybersecurity.
Bravo told CNBC’s “Squawk on the Street” on Tuesday that investors can’t value a $50 million annual recurring revenue company at $10 billion.
“That company is going to have to produce a billion dollars in free cash flow to double an investor’s money, ultimately,” he said. “Even if the product is right, even if the market’s right, that’s a tall order, managerially.”
Read more CNBC tech news
OpenAI recently finalized a secondary share sale that would value the ChatGPT-maker at $500 billion. The company is projected to make $13 billion in revenue for 2025.
Nvidia recently said it would invest up to $100 billion in OpenAI, in part, to help the ChatGPT maker lease its chips and build out supercomputing facilities in the coming years.
Other public companies have soared on AI promises, with Palantir’s market cap climbing to $437 billion, putting it among the 20 most valuable publicly traded companies in the U.S., and AppLovin now worth $213 billion.
Even early-stage valuations are massive in AI, with Thinking Machines Lab notching a $12 billion valuation on a $2 billion seed round.
Despite the inflated numbers, Bravo emphasized that there’s a “big difference” between the dotcom collapse and the current landscape of AI.
“Now you have some really big companies and some big balance sheets and healthy balance sheets financing this activity, which is different than what happened roughly 25 years ago,” he said.