The world’s largest software maker has pumped billions into the startup since 2019. Most of that came in a reported $10 billion deal that secured Microsoft 49% of OpenAI’s future profit, subject to certain financial milestones.
It’s tough to keep a huge business partnership like this out of what can be intense scrutiny from antitrust regulators. Lina Khan‘s FTC has shown clear interest in keeping large tech companies from swallowing new potential rivals. Microsoft tried anyway.
Intentionally avoiding scrutiny
One of Microsoft’s top concerns with OpenAI has been and continues to be antitrust, according to a person familiar with the startup and its leadership. Both companies have been dedicated to “keeping a wall between us, very intentionally,” this person said.
Many OpenAI announcements were not shared with Microsoft, and when they were, the two companies would ensure they did not announce anything in tandem or too close together.
“It was all intentional to avoid antitrust scrutiny,” the person added.
This was done even though it created more operational “friction” in the partnership, which “happened a lot,” according to this person, who asked not to be identified discussing sensitive matters.
Friction is a polite term for what was created by the shock November ouster of Altman. Microsoft got little warning of the board’s intention to fire Altman over accusations of a lack of candor.
Leaders like Microsoft CEO Satya Nadella and his CTO Kevin Scott didn’t communicate much with OpenAI’s independent board, but they were regularly in touch with Altman and other OpenAI senior executives.
“Leadership spent much more time in the presence of Kevin and Satya than the board,” the person added. “Much more.”
A longtime Microsoft insider said the company has been paranoid about antitrust issues since the days when it was known as the “Evil Empire” and fought a major antitrust lawsuit over its bundling of Internet Explorer with the Windows operating system. Under Nadella, Microsoft has had a largely effective strategy to avoid antitrust scrutiny. It even beat out regulators to buy Activision-Blizzard recently.
Now, with the dramatic return of Altman, Microsoft’s role in negotiating his reinstatement and its new position as an “observer” of a mostly new board, the company is finding itself under the antitrust microscope after all.
‘De facto control’
The UK’s Competition and Markets Authority said Friday it is taking a preliminary look at the implications of Microsoft’s investment in OpenAI, including whether the former has “material influence” or “de facto control” over the latter. The FTC is also looking at the tie-up more closely, in the wake of the Altman fiasco, according to Bloomberg.
An OpenAI spokesperson did not respond to a request for comment. Microsoft President Brad Smith said the partnership with OpenAI “has fostered more AI innovation and competition” and its new observer status on the board is “the only thing” that’s changed between the two companies. Chief Communications Officer Frank Shaw said Microsoft “does not own any portion” of OpenAI.
Microsoft leaders got only a few minutes warning before the announcement of Altman’s firing, insiders said at the time, and company leaders have since repeated that they were blindsided by the decision. The longtime Microsoft insider said that bodes well for Microsoft in any possible future antitrust case.
“The fact that Sam got fired and Microsoft didn’t know anything about that decision speaks to whether Microsoft has the ability to control OpenAI,” the person said.
Of course, then Nadella played a major role in Altman’s return. Altman and Nadella stayed in touch after Altman was forced out, with president and co-founder Greg Brockman quitting in solidarity and most of OpenAI’s staff threatening to do the same – however little they wanted to. Nadella agreed to give Altman and Brockman their own research arm at Microsoft, if he couldn’t negotiate their return to OpenAI.
Back to basics
With Altman returned as CEO of OpenAI now, Microsoft leaders have gone back to basics, working hard to publicly distance the company from its ties to the startup and its operations.
At Microsoft’s annual shareholder meeting on Thursday, Nadella said that, while its OpenAI partnership has given the company a “leading frontier model,” it’s working on separate internal AI projects, like a language model called Phi. It also offers open-source models on Azure, including Mistral, Llama, and Cohere, all of which compete with OpenAI’s models and tools, he noted.
Microsoft Chief Financial Officer Amy Hood jumped in during the meeting to emphasize the company has AI partners beyond OpenAI, as well. “We’re working with lots of other companies and partners in the AI space,” she said.
One interpretation of this is that Microsoft wants to calm investor concern following the chaos over Altman’s ouster. Another interpretation is that Microsoft is keen to show antitrust regulators that OpenAI is an independent company, and not controlled by the software giant.
That level of control will be one element of any antitrust review of its deal with OpenAI, according to a person familiar with government actions.
If the company is found to have exerted real control over OpenAI, or shared competitive information with the startup, before notifying the FTC, it could be fined for what’s known as “gun jumping,” one of the easier FTC rules to enforce, the person familiar with government actions said. Such fines can run upwards of $50,000 per day of violation.
“Which could explain why they’re trying to avoid the appearance of coordination now,” this person said.
Large deals like Microsoft’s investment in OpenAI often need to be reported to federal regulators, and that came sometimes trigger an investigation.
Microsoft has opted not to report the deal to regulators, partly because OpenAI was a relatively small business until ChatGPT took the world by storm a year ago. Nowadays, Microsoft insists that it doesn’t actually own part of OpenAI at all. Instead, the software giant says it only has rights to future income from the startup.
Even some of the lawyers at the firm representing OpenAI are confused by what exactly Microsoft’s legal relationship is to the startup, a person familiar with the matter told BI.
Do you work for OpenAI or Microsoft, or are you someone with a tip or insight to share?
Contact Kali Hays at email@example.com or on secure messaging app Signal at +1-949-280-0267. Reach out using a non-work device.
Contact Ashley Stewart at firstname.lastname@example.org, or send a secure message from a non-work device via Signal +1-425-344-8242.
Contact Darius Rafieyan using a non-work device on Signal +1-714-651-1367, or email email@example.com.