By Asa Fitch
It was once the hottest relationship in AI. As tech's latest craze started to blossom in 2023, Microsoft invested $10 billion in OpenAI, the startup that had captured the world's collective imagination with ChatGPT. In exchange, Microsoft got rights to OpenAI's technology and a revenue-sharing deal, among other terms.
More than three years in, though, circumstances have changed. And on Monday, the companies redrew that partnership agreement for the second time in about six months. The latest tweak freed OpenAI to sell AI tools outside of Microsoft's Azure cloud-computing service and capped how much revenue it must share with Microsoft through 2030.
It was, in one way, the latest indication that AI's corporate "it" couple is headed to splitsville. And the separation is in many ways logical for both sides.
For Microsoft, being closely tied to OpenAI makes less sense when AI models from OpenAI's competitors are just as good if not better in many contexts. Microsoft has already begun integrating Anthropic's models -- which excel in coding and corporate tasks -- into some of its AI tools. Microsoft is also developing its own AI models, rather than leaning on OpenAI.
OpenAI's incentives to step away are even stronger. Freeing itself from exclusivity with Microsoft is critical if it is to supercharge revenue growth and compete with the likes of Anthropic. Maximizing financial strength is a priority for OpenAI ahead of a planned IPO as soon as this year. That is an especially urgent task given the company has recently been missing its own targets for user and revenue growth.
Microsoft and OpenAI have been adamant that they remain extremely close, even as they have started to go their separate ways. OpenAI's models are still deeply embedded in many of Microsoft's AI offerings, and Microsoft remains a major investor in OpenAI.
Those factors will likely keep the companies together for some time more. But their diverging business interests suggest a fuller breakup is eventually in the cards.
This is an edition of the WSJ AI & Business newsletter, a weekly digest to help you make sense of AI's impact on business with news, insights and data from our global team of technology journalists. If you're not subscribed, sign up here.
China Takes Aim At $2.5 Billion Meta AI Deal
China ordered the unwinding of Meta Platforms' $2.5 billion acquisition of the AI startup Manus on Monday. That move could chill U.S. investment in Chinese-founded companies and inject further tension into the battle between the countries for technological supremacy. Meta has already integrated Manus's AI agent software into its products, but is preparing to back out of the deal. Beijing appears intent on limiting U.S. access to technology developed in China after U.S. measures including export controls severely hampered China's efforts to develop cutting-edge AI models.
The Number
Damages Elon Musk is seeking from OpenAI in a trial that began on Monday with jury selection in an Oakland courtroom. Musk alleges he was tricked into financially supporting OpenAI when it was a nonprofit, before it was effectively turned into a for-profit company seeking an IPO.
What the Humans Are Saying
AI in Charts
Layoffs have become the latest tech-industry fad. More tech-worker layoffs were announced in March than any other month in at least two years -- and more are likely coming.
Many tech companies believe the time is right to reduce their head count because AI is becoming more capable by the day. They are also making a simple calculation that to afford ever-growing spending on AI computing power, they need to cut employee costs.
Whether the story ends well for the tech companies is hard to say, but layoffs do come with costs. Jettisoned employees may launch startups that compete with them. And some of the layoffs may have more to do with correcting earlier overhiring than seizing the AI moment.
AI in the Wild
A rising tide of opposition to data centers is sweeping across the U.S. That animus may be most intense in the St. Louis suburb of Festus, where residents have voted out city council members who supported data-center construction. Surveys increasingly indicate that majorities of both Democrats and Republicans oppose AI data centers in their cities, posing challenges for tech companies eager to expand their computing footprints.
Other Highlights From the Week in AI
-- Opposition to data center construction may be mounting, but Maine's
governor vetoed a bill to temporarily ban their construction.
-- Google said it would invest up to $40 billion in Anthropic, building on
an already tight relationship ahead of Anthropic's potential IPO.
-- Meta signed a deal with Amazon to use Amazon's Graviton chips for AI work
-- a pact the companies said was worth billions of dollars.
About Us
WSJ AI & Business is a weekly look at AI's transformation of the business world. This newsletter was curated and edited by Dan Gallagher and Asa Fitch. Reach them at dan.gallagher@wsj.com and asa.fitch@wsj.com (if you're reading this in your inbox, you can just hit reply). Got a tip for us? Here's how to submit.
(END) Dow Jones Newswires
April 28, 2026 12:00 ET (16:00 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments