Microsoft's OpenAI Fear and the $100B Partnership That Changed Everything

A lot of new information about the past of the most important partnership in modern technology has come out during the Musk v. Altman trial in Oakland. One of the most interesting is that Microsoft CEO Satya Nadella was worried about OpenAI taking over as early as April 2022, seven months before ChatGPT came out and made Sam Altman famous.

In April of that year, about three years after Microsoft gave OpenAI its first $1 billion check, Nadella wrote to execs, "I don't want to be IBM and OpenAI to be Microsoft." The reference to IBM was meant to be funny. In 1980, IBM decided to use Microsoft's operating system on its computers, but later saw the software company take over the industry and make a lot of money. Nadella was set on not playing that situation over from the other side.

What the Trial Revealed About Microsoft's Strategic Anxiety

This week, Nadella's testimony in court showed how Microsoft's AI policy has been shaped from the start by tensions within the company. "It was becoming even more core and important that we had real agency at every layer of the stack," he stated on Monday. The intellectual property rights that Microsoft negotiated, which gave it access to OpenAI's models and technology, were clearly meant to protect Microsoft in case OpenAI grew too big for Microsoft to control or use.

"Essentially, we were forgoing the opportunity on our side to develop it on our own, so it was very important to us to have IP rights," Nadella said.

It had changed from a strategic to a financial argument by July 2022: "Better to be an investor and not even take all this execution risk!" Nadella told the leaders in a message sent to them. That way of thinking—as a passive investor instead of an active developer—explains a lot about Microsoft's situation and its current strategic problem.

Michael Wetter, a business development executive at Microsoft, said in court that the company will have spent more than $100 billion on OpenAI by June 2026. This amount includes investments, infrastructure, and hosting costs. CTO Kevin Scott said that he and OpenAI worked for six months to build the first AI supercomputer using 10,000 GPUs. He was "very proud of what we've enabled OpenAI to go do."

The pride is real. The geopolitical situation is hard to understand.

The Exclusivity That Was Surrendered — and What It Cost

As OpenAI grew in size and power, the relationship has been renegotiated several times. Most recently, in April 2026 and made official on May 12, Microsoft and OpenAI put a $38 billion cap on all revenue-sharing payments and got rid of the rule that OpenAI had to route API calls through Azure in order to be exclusive. Users can now get "all of its products" from any provider, like Amazon, Google, or anyone else.

Microsoft still has a 27% stake in OpenAI and is still a provider of key infrastructure. But it isn't the only company that can offer OpenAI's technology to the market anymore.

The effect on business can be seen in the way competitors act. OpenAI's models and services are now used on so many systems—at the same time, on platforms from Microsoft, Google, Amazon, and Oracle—that the company is now worth $850 billion. Early on, Microsoft had exclusive access, which Nadella correctly saw as strategically important. Over time, Microsoft gave up that access as OpenAI needed more space to grow than Azure alone could provide.

At the end of 2025, about 45% of Microsoft's business performance obligations were still due. This was a two-way dependency. Azure makes money because OpenAI needs so much computing power. However, Microsoft's AI story has had a hard time setting it apart from a company that now serves as a technology provider for its most important AI partner.

Microsoft's Scattered AI Strategy

The MSFT price chart shows some things. Shares are down 16% so far this year, while cloud peers are selling higher. This difference shows that investors are more worried about Microsoft's AI products than its Azure infrastructure business.

As of January 2024, Nadella was talking about AI models within the company as "more of a commodity." Two months later, the company hired DeepMind co-founder Mustafa Suleyman to lead a new Microsoft AI unit that would be in charge of Bing and consumer goods. The unit's job would be to make models that could compete with OpenAI, Anthropic, and Google.

This was announced by Microsoft in August 2025. They were trying a homegrown AI model to make Copilot better. It changed the people in charge of Copilot in March 2026, hiring former Snap executive Jacob Andreou to help with consumer and business Copilot. Suleyman was left to work on developing new models.

As a result, the company is making a lot of bets at the same time: it is investing in OpenAI, making its own models, licensing Anthropic models for certain uses (an investment of $5 billion in OpenAI's competitor was announced last year), and putting xAI's models on Azure.

That range might be Microsoft's strong point—Azure as the open base for AI model access. What might also be its weakness is that no other AI product has become as popular with users as ChatGPT.

Technical Picture: $393 Is the Level to Watch

In the short term, the 50-day moving average at $398.68 can help MSFT stay above $403.95, which is below both the 20-day moving average at $418.20 and the 200-day moving average at $464.62. The RSI has dropped below 50, the stochastic signs are oversold, and the chance of a price rise in the next few weeks is thought to be less than 20%.

$393 is the important technical mark for the next week. If that level is broken for a long time, it takes away the last important moving average support and is likely to make selling pressure rise. A rise above $419 would be the first sign that the risk of falling in the near future has leveled off.

Microsoft is a real example of the paradox that is the current state of AI: the company that did more than any other to make the generative AI boom possible is now having a hard time getting its fair share of the value that was created.

Azure has become an essential AI infrastructure tool thanks to the $100 billion that was put into OpenAI. It has also made a partner that is now a customer, a competitor, and a business that gives Microsoft's main competitors money and skills. Nadella was right when he said that IBM would fail in 2022. Things haven't turned out exactly as he predicted, but the strategic trap he saw coming has partly materialized.

The next part will rest on whether Microsoft's model development and Copilot products can build a strong connection between AI and humans in both consumers and businesses like OpenAI has done with ChatGPT. The market has said so far that it hasn't.