Microsoft invests a reported $10 billion in OpenAI

Microsoft announced what it called “the third phase” of its partnership with OpenAI, describing it only as a “multiyear, multibillion dollar investment” without disclosing a figure. Bloomberg and other outlets reported the same day, citing people familiar with the deal, that the sum was $10 billion — a number that recurred everywhere in subsequent coverage but that neither company ever confirmed on the record.

The terms Microsoft did publish were substantial in themselves. Azure became OpenAI’s exclusive cloud provider, underpinning both OpenAI’s research and its API business. Microsoft said it would build out new AI-optimised supercomputing infrastructure for OpenAI’s training and deployed OpenAI’s models across its own products, while OpenAI retained the right to commercialise its technology independently. Satya Nadella framed the deal as extending a partnership “formed around a shared ambition to responsibly advance cutting-edge AI research,” and OpenAI said the funding would let it continue research that was “independent” while pursuing its stated mission.

The deal followed an initial $1 billion Microsoft investment in 2019 and a further round in 2021, and reportedly gave Microsoft a large share of OpenAI’s profits until it recouped its investment, after which its stake in the nonprofit-controlled entity would settle at a minority position — an unusual structure that later drew scrutiny as OpenAI moved toward a more conventional for-profit reorganisation.

The announcement landed a day before Microsoft’s quarterly earnings call and roughly two weeks before Microsoft put a GPT-4-class model inside Bing, making clear the investment was not passive capital but the financing for a specific product push. It set the template other tech companies followed over the next two years: large, opaque, multi-year capital commitments to a frontier lab in exchange for compute exclusivity and early model access, rather than a conventional equity stake with disclosed terms.