Microsoft’s OpenAI Bill Passes $100 Billion

Michael Wetter does not give many interviews. The Microsoft executive who handles the company’s biggest transactions is usually the one doing the negotiating, not the talking. So when he confirmed on May 11 that Microsoft has invested more than $100 billion in its partnership with OpenAI, the number landed with the weight of an official disclosure.

The figure covers the full ledger of the relationship: Microsoft’s original investment in OpenAI, the capital it has put into building infrastructure for the startup, and the cost of providing compute hosting for OpenAI’s models. Wetter’s confirmation, reported by Jiemian, gives the first authoritative public total for a partnership that has been the defining alliance of the AI boom — and one that has become more complicated with every passing year.

The scale is worth pausing over. A hundred billion dollars is more than the annual revenue of most countries’ largest companies, and it represents the single largest corporate bet on any AI startup in history. Microsoft made the wager in stages, starting with a $1 billion investment in 2019, expanding it to roughly $10 billion in 2023, and then pouring in capital and compute capacity at a pace that has accelerated as OpenAI’s needs have grown. Each tranche was justified at the time as essential to keeping Microsoft at the center of AI; the cumulative total shows how expensive that position has become.

The partnership’s structure has evolved along the way. In its original form, Microsoft held a significant stake in OpenAI and enjoyed a privileged position as the exclusive provider of the startup’s cloud computing through Azure. That arrangement loosened in 2025, when OpenAI struck deals with other cloud providers to secure additional capacity, and the two companies renegotiated their terms. The exclusivity is gone, but the money has kept flowing — and Microsoft has made clear it still expects to be OpenAI’s largest investor and infrastructure partner.

What Microsoft gets in return has shifted from control to economics. The company retains the right to use OpenAI’s models across its products, and it has built its own AI strategy on top of them, from GitHub Copilot to the AI features woven into Windows and Office. Its cloud business sells OpenAI’s models to customers around the world. As OpenAI’s usage has grown, so has the traffic running through Microsoft’s data centers — a virtuous cycle that executives have described as the core of the deal.

The tensions are real, though. OpenAI’s reported dissatisfaction with partners it believes underdeliver — including its public friction with Apple over ChatGPT integration — has made Microsoft watchful, according to people familiar with the relationship. The two companies have also competed directly in areas that once seemed settled: OpenAI’s own agent tools compete with Microsoft’s Copilot products, and OpenAI has explored building hardware that would reduce its dependence on any single cloud. Analysts said the relationship has become less an alliance of equals and more a negotiation conducted at scale.

The $100 billion figure also illuminates the broader economics of AI infrastructure. The capital being poured into compute — by Microsoft, by OpenAI’s other backers, by the hyperscalers building data centers on every continent — has become the defining feature of the industry’s current phase. Companies are spending enormous sums on machines that may not pay for themselves for years, on the conviction that AI demand will keep growing. Microsoft’s OpenAI bill is the clearest single number attached to that bet.

Shareholders have mostly tolerated the spending, in part because Microsoft’s cloud business continues to grow and in part because the alternative — letting OpenAI fall to a competitor — was considered worse. But the scrutiny has increased. Every quarter, analysts ask how much more capital the OpenAI relationship will absorb, and Wetter’s confirmation has given them a larger denominator than they had before. The question now is whether the investment starts to pay back at a similar scale.

For OpenAI, the disclosure cuts both ways. The company can point to the number as evidence of its importance — no startup in history has commanded a hundred billion dollars of committed capital from a single partner. It also signals dependence: OpenAI’s computing strategy rests on relationships with companies that expect returns. As OpenAI moves toward a public listing, expected by some investors as early as this year, the terms of its infrastructure agreements will be among the most scrutinized items in its filings.

Neither company is saying the partnership is ending. Wetter’s remarks, brief as they were, framed the relationship as an ongoing commitment rather than a completed transaction. But the number he confirmed will now follow both companies around: a hundred billion dollars invested, with the return still being counted. Investors in both companies will be doing the arithmetic for years to come.

Related Posts

  • September 6, 2026
  • 6 views
Anthropic Moves Its IPO Filing to Late September

The bankers and lawyers running Anthropic’s initial public offering had told investors to expect the company’s registration documents as soon as this week. The calendar has moved. Anthropic now plans…

  • September 6, 2026
  • 6 views
OpenAI Quietly Revises GPT-6 Astra Scores After Launch

When OpenAI released GPT-6 Astra on Sept. 3, the launch post carried the usual furniture of a modern model debut: coding results, speed comparisons and a figure for how often…