Microsoft and OpenAI have announced a significant revision to their partnership, which will now allow OpenAI to access compute capacity from competitors. The new agreement, announced on Tuesday, includes a right of first refusal (ROFR) for Microsoft on supplying new capacity, marking a shift away from the exclusive arrangement that previously bound OpenAI to Microsoft's Azure cloud infrastructure.

The revised partnership will enable OpenAI to build additional capacity, primarily for research and training of models, and marks a significant departure from the original agreement. Prior to this change, OpenAI was locked into using Microsoft's Azure cloud infrastructure exclusively for its computing needs. The move is seen as a response to concerns from OpenAI shareholders that Microsoft wasn't moving fast enough to supply OpenAI with computing power.

The news comes on the heels of a joint venture between Arm, Microsoft, Nvidia, Oracle, and OpenAI to build a system of data centers in the U.S. called Starbase. The partnership aims to provide the necessary compute power for OpenAI's ambitious projects, which require billions of dollars in compute. OpenAI had previously partnered with Oracle in June to secure additional compute capacity, with Microsoft's blessing.

The revised partnership also sheds light on the complex and sometimes strained relationship between Microsoft and OpenAI. The New York Times reported that the relationship has grown increasingly tense due to financial pressures at OpenAI, concerns about stability, and growing friction between employees at both companies. The hiring of Inflection CEO Mustafa Suleyman by Microsoft to lead its consumer AI efforts, along with most of Inflection's staff, in a $650 million deal, reportedly angered some OpenAI leadership, including CEO Sam Altman.

The deal between Microsoft and OpenAI also includes an unusual escape clause, which would allow OpenAI to close off Microsoft's access to some of its most powerful models developed after the creation of artificial general intelligence (AGI). AGI is reportedly defined as a system capable of generating more than $100 billion in profits. This provision was originally meant to keep such powerful AI from being commercialized, but OpenAI is now reportedly considering dropping this provision to secure more Microsoft funding.

The revised partnership has significant implications for the AI industry, as it allows OpenAI to access a wider range of compute capacity and potentially accelerate its research and development efforts. The move also underscores the complex and evolving nature of partnerships in the AI space, where companies are constantly seeking to balance their interests and secure their positions in the market.

As the AI landscape continues to evolve, the revised partnership between Microsoft and OpenAI will likely have far-reaching consequences for the industry. With the joint venture on Starbase and the revised partnership, OpenAI is well-positioned to make significant strides in AI research and development, and Microsoft is poised to maintain its position as a key player in the AI ecosystem.