Before November 17, the prevailing view in Silicon Valley was that Microsoft's collaboration with OpenAI had achieved enviable success. The investment boosts Microsoft's cloud computing business, gives it access to OpenAI's best technology, breathes new life into its search engine Bing, and helps streamline its research efforts in artificial intelligence. Because Microsoft holds less than 50% of OpenAI's shares, it has avoided the antitrust scrutiny that has followed it since the 1990s.
But one of the downsides of outsourcing key technology to a startup (even though many consider it a de facto subsidiary of Microsoft) is that it can self-destruct without being ventilated ahead of time. Microsoft found out about the ouster of OpenAI CEO Sam Altman just minutes before the public, catching company executives off guard and sending its stock price plunging. Soon, Microsoft CEO Satya Nadella and his staff helped engineer a dramatic counterattack that reinstated Altman's job and ensured the ouster of the board members least aligned with Microsoft's interests. “He played a personnel game from afar,” said Sheila Gulati, a longtime manager at Microsoft who is now president of venture capital firm Tola Capital.
Through this "coup" and the subsequent reversal, it can be seen that Microsoft's leading position in the field of artificial intelligence relative to its main competitors is still uncertain. OpenAI's revamped board is planning to launch an investigation into Altman, which could reignite the controversy surrounding him. Meanwhile, Google, Facebook, Anthropic and other competitors appear to be catching up.
But this underestimates Microsoft's leverage in the event of a dispute. As a major investor in OpenAI, Microsoft has the right to resell OpenAI's technology to its own enterprise customers and has a broad license to use the startup's artificial intelligence models. Microsoft's Azure cloud computing unit also built supercomputers specifically for OpenAI to train its models. In other words, Microsoft is well-positioned to quickly launch a reliable OpenAI clone if something goes wrong.
That would allow Microsoft to do more or less business as usual: aggressively adding artificial intelligence assistants, which it calls "copilots," to all of its software products. These AI assistants don’t come cheap. For example, any business that wants an AI assistant for Word and Excel will have to pay an additional $30 per user per month, which is roughly twice what the average business customer pays for Microsoft's Office software suite. Meanwhile, free and open source AI assistants are everywhere. Microsoft is banking on customers being willing to pay for the productivity gains that Co-Pilot brings and the convenience of having it embedded into such a wide range of software.
"It's going to be present in all your experiences," said Rajesh Jha, executive vice president for the office software suite, Windows and search product teams. He also said that Microsoft wants to "become a co-pilot company." In a way, the name of this product is apt. Microsoft is betting its future on an uncertain technology, although it's unclear whether control of the technology lies with Nadella or OpenAI's board of directors.
In summary, while OpenAI may look like a de facto subsidiary of Microsoft, the tech giant has much less control over it than investors typically get. Microsoft's lead in AI relative to its main rivals remains uncertain.