Is the OpenAI and Microsoft partnership at risk?

Two reports today highlight potential trouble brewing with Microsoft’s $13 billion deal with OpenAI. The partnership, which forged a deep strategic and technological alignment between one of the largest and oldest technology companies in the world and one of the youngest, most closely watched AI companies, is complex and unusual.

The 2023 deal brought huge computing resources to OpenAI through Microsoft’s vast Azure cloud infrastructure, and OpenAI licensed its large language models for use in Microsoft’s wide range of products. Azure was also locked in as the exclusive cloud provider for OpenAI’s services.

That all sounds fairly straightforward, with each side getting a benefit from the other, but here’s where it gets complicated. OpenAI’s recent chaos on its executive team, which included the departure of key founding members, was partly fueled by CEO Sam Altman’s intention to restructure the nonprofit into a for-profit company with a smaller, less powerful nonprofit attached to it.

That complicates the calculation of how much equity Microsoft is getting in the deal. The Wall Street Journal reported on this sticking point today, noting that both parties have hired large investment banks (and bitter rivals in Goldman Sachs and Morgan Stanley, by the way) to advise them on the negotiations.

Now valued at $157 billion after a recent round of high-profile fundraising, OpenAI’s financials revealed enormous expenses (and losses) that its business is incurring as it grows. The New York Times reported that the pressure these losses are placing on OpenAI has pushed it to renegotiate the terms of the deal, seeking lower costs for the use of Microsoft’s computing resources.

Adding to that, both companies appear to be making some moves to hedge their bets on each other, with Microsoft seeking alternative pools of talent and OpenAI diversifying its infrastructure providers.