This article is part of our series that explores the business of artificial intelligence
Microsoft’s Build conference was an amazing display of new features and products powered by different artificial intelligence technologies, including powerful large language models (LLM). The conference came a mere two weeks after Google convinced at I/O 2023 that it had caught up with Microsoft in the race to dominate the growing market for generative AI.
But as I watched the fascinating product demos on day 1 of MSBuild, one thing that stood out was the role that OpenAI is playing in Microsoft’s growing AI power. OpenAI is (still) an independent AI research lab and company. But its products were heavily featured in Build, in some cases in ways that almost felt as if Microsoft owned them.
ChatGPT at the forefront
Microsoft CEO Satya Nadella began his keynote address with a background image of a timeline of milestones in the history of computation. The timeline started with Vannevar Bush’s “As We May Think” essay and ended with the launch of ChatGPT, OpenAI’s flagship LLM.
“Last November, we got an upgrade. We went from the bicycle to the steam engine with the launch of ChatGPT,” Nadella said.
Nadella continued by saying that Build will feature 50-plus new announcements. But which one did he decide to showcase first?
“I want to highlight five of them. The first is we are bringing search grounding in Bing to ChatGPT,” he said and adding that ChatGPT is the fastest-growing consumer app in history.
Again in his speech, as he talked about the Copilot Stack (amazing technology, by the way), Nadella mentioned ChatGPT as if it was a Microsoft product:
“We’re also very excited to launch the Copilot Stack. After we built all these Copilots with one common architectural stack, we want to make that available so that everyone can build their own copilot for their applications. We will have everything from the AI infrastructure to the foundation models to the AI orchestration all the way up to your copilot and its extensibility. We will have common extensibility across all these surfaces, whether it’s ChatGPT, Bing Chat, Microsoft 365 Copilot, all Microsoft Copilots, and of course, your own copilots.”
And when Yusuf Mehdi, Microsoft’s Corporate Vice President and Consumer Chief Marketing Officer, took the stage, his presentation was marked by the heavy presence of ChatGPT, ChatGPT plugins, and the integration of other OpenAI models in Microsoft products.
In comparison, Google’s generative AI strategy is also heavily influenced by ChatGPT. Its I/O conference was marked with instruction-following chatbots, the introduction of the PaLM API (a competitor to GPT-4), and the use of LLMs in different Google apps. But it did so without publicizing ChatGPT or OpenAI.
Microsoft and OpenAI’s non-profit/for-profit dilemma
It can be argued that ChatGPT was a watershed moment in the recent history of AI. It has had a profound impact on how companies view the role of generative AI in their business and products. And it has surely triggered an AI arms race between tech giants. But it is an OpenAI product hosted on the OpenAI domain. This is why I find it a bit unsettling that it is being featured so heavily in Build and talked about as if it is a Microsoft product.
On the one hand, this shows how much ChatGPT and OpenAI have become vital to Microsoft’s AI strategy. On the other, it fortifies some of the criticism against OpenAI about transforming from an open research lab to a closed for-profit subsidiary of Microsoft.
Elon Musk, who co-founded OpenAI, has repeatedly warned about Microsoft’s growing sway over OpenAI. In a recent interview with CNBC, Musk said, “I worry that Microsoft may be in more control than the team at OpenAI realizes. As part of the Microsoft investment, they have the rights to all of the software, all of the model weights, and everything that is needed to run the inference system… At any point, Microsoft can cut off OpenAI.”
Musk is not the only person warning about the relationship between Microsoft and OpenAI. In a recent Senate hearing on AI governance, Gary Marcus said, “OpenAI’s original mission statement proclaimed our goal is to advance AI in the way that most is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return. Seven years later, they’re largely beholden to Microsoft, embroiled in part an epic battle of search engines that routinely make things up.”
Both companies deny this. In an interview with NBC, Nadella said, “OpenAI is very grounded in their mission of being controlled by a non-profit board. We have a non-controlling interest in it. We have a great commercial partnership in it. And quite honestly, I am very comfortable in partnering with a capped-profit company that has a mission of fundamentally pursuing this very powerful technology that is ultimately going to be controlled by a non-profit. In fact, the last time I checked, we are the only for-profit company that is comfortable with a non-profit company and a board controlling technology, and I’d welcome others to do that as well.”
But the partnership with the non-profit has been a huge boon for Microsoft, and its AI services are on their way to becoming a multi-billion-dollar business.
The undeniable reality is that Microsoft and OpenAI have become so entangled that it is almost inconceivable how one could succeed without the other—with Microsoft having the upper hand at the end of the day. Imagine what the Microsoft Build conference would look like if you took out all of the OpenAI products and graphics from it. And imagine where OpenAI would be without Microsoft’s financial backing. On the one hand, Microsoft relied heavily on OpenAI’s talent and research to create state-of-the-art AI systems that it later integrated into its products. On the other hand, OpenAI relies heavily on Microsoft to fund its research and provide it with channels to monetize its technology. Had it not been for Microsoft’s huge distribution channels, OpenAI would be another commercial AI lab competing for API customers.