With global regulators examining Microsoft Corp’s $13 billion investment in OpenAI, the software giant has a simple argument it hopes will resonate with antitrust officials: It doesn’t own a traditional stake in the buzzy startup so can’t be said to control it.
When Microsoft negotiated an additional $10 billion investment in OpenAI in January, it opted for an unusual arrangement, people familiar with the matter said at the time.Rather than buy a chunk of the cutting-edgeartificial intelligence lab, it cut a deal to receive almost half of OpenAI’s financial returns until the investment is repaid up to a pre-determined cap, one of the people said. The unorthodox structure was concocted because OpenAI is a capped for-profit company housed inside a non-profit organization.
It’s not clear regulators see a distinction, however. On Friday the UK Competition and Markets Authority said it was gathering information from stakeholders to determine whether the collaboration between the two firms threatens competition in the UK, home of Google’s AI research lab Deepmind. The US Federal Trade Commission is also examining the nature of Microsoft’s investment in OpenAI and whether it may violate antitrust laws, according to a person familiar with the matter.
The inquiries are preliminary and the agency hasn’t opened a formal investigation, according to the person, who asked not to be named discussing a confidential matter.
Microsoft didn’t report the transaction to the agency because the investment in OpenAI doesn’t amount to control of the company under US law, the person said. OpenAI is a non-profit and acquisitions of non-corporate entities aren’t reported under US merger law, regardless of value. Agency officials are analyzing the situation and assessing what its options are.
“While details of our agreement remain confidential, it is important to note that Microsoft does not own any portion of OpenAI and is simply entitled to a share of profit distributions,” a Microsoft spokesperson said in a statement. Earlier Friday, Microsoft president Brad Smith said “the only thing that has changed is that Microsoft will now have a non-voting observer on OpenAI’s board.” He described its relationship with OpenAI as “very different” from Google’s outright acquisition of DeepMind in the UK.
“Our partnership with Microsoft empowers us to pursue our research and develop safe and beneficial AI tools for everyone, while remaining independent and operating competitively. Their non-voting board observer does not provide them with governing authority or control over OpenAI’s operations,” said an OpenAI spokesperson in a statement.
From the beginning, Microsoft and OpenAI took pains to telegraph the two companies’ independence. Microsoft hoped to reassure investors and customers that it’s not overly reliant on one partner. OpenAI didn’t want employees, customers and other investors thinking it was merely an outpost of Redmond, Washington-based Microsoft. That careful positioning was upended last month with the firing of OpenAI chief executive officer Sam Altman and the startup’s near implosion.
The Altman imbroglio demonstrated both Microsoft’s lack of control and its influence. Microsoft received just minutes notice that the OpenAI board planned to announce Altman’s ouster, and its executives were not consulted in the decision. Still Microsoft CEO Satya Nadella played a key role, along with other investors, in forcing the board to reverse its decision. At one point Microsoft said it would hire Altman and his OpenAI colleagues to form a new Microsoft AI unit.
Once Altman was restored as CEO, Microsoft executives debated the wisdom of taking a seat on the OpenAI board, people familiar with the matter said at the time. On the one hand, executives feared that a board seat or observer slot might draw the attention of regulators. On the other hand, Microsoft wanted to keep a closer eye on its partner and protect its investment—an imperative that carried the day, despite the risks.
Ultimately, Microsoft could face a world of regulatory headaches. Regulators in Europe are also paying attention, according to a spokesperson for the European Commission. In order for a transaction to be notifiable to the Commission under the EU Merger Regulation, it has to involve a change of control on a lasting basis. While this transaction has not been formally notified, the Commission had been following the situation even before the management turmoil, the spokesperson said.
Last month, Germany’s competition authority said it wasn’t subjecting Microsoft’s OpenAI investment to a merger review. But the regulator said they would hold off only because OpenAI didn’t have substantial business in Germany. After reviewing the transaction and talking the companies, the regulator found the investment would give Microsoft a “material competitive influence” over the AI company that might warrant scrutiny in the future if OpenAI increases its activities in Germany.
The partnership raises competition issues if Microsoft cuts back on its own AI research and development or if the investment keeps OpenAI from partnering with the tech giant’s rivals, said Bloomberg Intelligence antitrust analyst Jennifer Rie. Antitrust enforcers may also have concerns about Microsoft’s board observer since it would give Microsoft additional information on OpenAI’s plans even if it doesn’t have rights to influence the decisions.