OpenAI transition from Microsoft to Amazon turns aggressive after subtle

Dennis Dresser, head of revenue at OpenAI, said on Tuesday that the AI ​​company Amazon It had nothing to do with the startup’s announcement the day before that it had restructured its relationship with. microsoft Second time in half a year.

In an interview after OpenAI and Amazon’s announcement, Dresser told CNBC that “there is no relationship between the two.”

Analysts aren’t so sure either.

A lot has happened since late October, when OpenAI completed a recapitalization and gave Microsoft a 27% stake in the commercial arm of the artificial intelligence company. As part of that deal, OpenAI agreed to purchase $250 billion in additional Azure services. The revenue-sharing agreement will remain in place until an independent panel verifies that OpenAI has reached artificial general intelligence (AGI) status.

One of the big developments since then is that OpenAI is moving closer to Amazon, Microsoft’s biggest rival in cloud infrastructure.

OpenAI unveiled a $38 billion deal with Amazon Web Services in November. And in late February, Amazon announced it would invest $50 billion in OpenAI and use 2 gigawatts worth of custom AWS Trainium chips to train its AI models.

Amazon and OpenAI also agreed at the time to have Amazon’s engineering team co-develop “customized models” to power its consumer products, increasing OpenAI’s spending commitment to AWS by $100 billion.

“That was the big thing that was happening,” Rishi Jallia, an analyst at RBC Capital Markets who recommends buying Microsoft stock, said in an interview.

This week’s one-two punch is the clearest sign yet that dramatic changes are afoot in Microsoft and OpenAI’s decade-long relationship.

It started in 2016 when OpenAI began large-scale experiments on Azure. Three years later, Microsoft invested an initial $1 billion in OpenAI, which grew to $13 billion in several subsequent rounds.

However, in 2024, Microsoft began referring to OpenAI as a competitor in its financial disclosures, and early last year the software giant lost its designation as OpenAI’s exclusive cloud provider. In an internal memo earlier this month, Dresser wrote that OpenAI’s partnership with Microsoft is “foundational to our success,” but it “also limits our ability to respond to businesses where they are.”

Against this backdrop, the latest agreement between the two companies is “very fluid and, to our knowledge, could change again within six months,” UBS analysts said in a note on Monday.

Additional elements of the agreement include the termination of Microsoft’s exclusive license to OpenAI’s intellectual property and the termination of Microsoft’s revenue share payments to OpenAI. Additionally, Microsoft will no longer be the only cloud provider for API products built in collaboration with third parties.

“While some changes seem inevitable, Microsoft appears to have made more concessions than gains,” wrote UBS analysts who rate Microsoft a buy.

Amazon CEO Andy Jassy said in a post on X that Monday’s announcement was “very interesting,” adding that more details would be shared on Tuesday.

Hours later, his company announced a service that uses OpenAI models to build AI agents.

For years, developers interested in these models had to go through Microsoft’s Azure cloud or work directly with OpenAI. Going forward, companies that have made significant investments in AWS will be able to more easily adopt this model while leveraging large spending plans.

Dresser, speaking at an Amazon event, said the review of OpenAI’s agreement with Microsoft was not inspired by the expanded collaboration with Amazon.

“Microsoft is our original partner,” she said. “They’ve been a great partner for us. They’ll be a great partner as we move forward. What we’re focused on is making sure that when we meet our customers where they are, we have access to the environment they’re working in. And we want to make sure we provide the best model in the best environment for our customers to succeed.”

The Financial Times reported that Microsoft is considering legal action regarding OpenAI’s plans with Amazon, and Microsoft told the newspaper it is “confident that OpenAI understands and respects the importance of meeting (its) legal obligations.” Microsoft has not commented beyond Monday’s announcement.

Microsoft is also moving toward diversification away from OpenAI.

Microsoft announced in September that it was starting to use Anthropic’s AI models to answer some queries in the 365 Copilot assistant for commercial customers. Two months later, Microsoft agreed to invest up to $5 billion in Anthropic, and Anthropic committed to buying $30 billion in Azure compute capacity.

Capitalizing on the growing popularity of Anthropic’s Claude Code, Microsoft teamed up with Anthropic in March to announce a product called Copilot Cowork.

One of the downsides to the surge in demand for Claude was that it became less reliable. The company reported experiencing partial or major outages in 37 of the past 90 days. Amazon, Anthropic’s early partner and investor, took notice.

AWS vice president Anthony Liguori said his team building the Bedrock service for working with AI models switched to OpenAI’s Codex as its primary development platform after relying on Claude Code and Amazon’s own Kiro tool.

The reality for all the key parties involved is that they need each other.

With very limited capacity, OpenAI and Anthropic must work with all major cloud vendors to secure as much compute as possible. Microsoft and Amazon also need easy access to all major models to serve their large customer bases.

So while Microsoft and OpenAI may be drifting apart, Jaluria is quick to point out that “Microsoft still needs OpenAI, and OpenAI still needs Microsoft.”

WATCH: Private investors don’t believe OpenAI is as valuable as it pretends to be, says CFR’s Sebastian Mallaby