Microsoft ramps up open rivalry with OpenAI and Anthropic

Microsoft finds itself in a distinctive position as artificial intelligence reshapes the technology sector. The company is among the globe’s leading cloud providers and software-as-a-service enterprises, yet it also maintains significant investments in two of the premier AI research organizations: OpenAI and Anthropic. However, these dual incentives are beginning to collide, especially as Microsoft posts remarkable financial results. The tech giant recently disclosed a highly profitable quarter, generating $90 billion in revenue and $35.8 billion in net income. For the fiscal year ending June 30, Microsoft reported $331.8 billion in total revenue and $133.7 billion in net profit.

CEO Satya Nadella is determined not to let the ambitions of Anthropic and OpenAI,both of which are expanding into applications and agentic infrastructure that could eventually allow them to own customer relationships,interfere with that kind of cash flow. Nadella has been urging enterprises to adopt multiple AI models and avoid relying exclusively on frontier AI labs for the agentic harness and application layer. He warns that such dependence is risky because it forces companies to share too many internal secrets with model creators whose trustworthiness is questionable. He understands his customers well: enterprise IT departments fear both data leaks and vendor lock-in.

During Microsoft’s quarterly conference call on Wednesday, Nadella openly told Wall Street analysts that this dynamic presents an opportunity. He positioned Microsoft as a vendor that can sell customers its own homegrown models, alongside AI agents, AI security solutions, and more, all while promising lower costs. In essence, he is pitching Microsoft as a viable alternative to many of the premium services that OpenAI and Anthropic are developing for their own growth.

When UBS analyst Karl Keirstead specifically asked Nadella to weigh in on the open-source vs. closed-source debate currently roiling the AI industry,and how Microsoft would benefit,Nadella responded forcefully. “The goal is to have the firm be in control of their own destiny,” he said, referring to enterprises. “We are very, very clear about the architectural sort of design of the platform, which is you got to keep your harness separate from the model … that means any model at any given time is swappable.”

Microsoft, of course, sells a menu of harnesses,also known as AI agents,under the Copilot brand, including its coding agent GitHub Copilot. Coding agents represent where much of the AI spending is currently concentrated. Nadella used a high-profile incident from the previous week as evidence to support his warnings. “If you look even at the Hugging Face incident, the biggest thing that we should take away from that is you can’t sort of depend on any one model,” he explained. “You will maybe need multiple models to even remediate some challenges that get caused by one model. Like that’s the way to think about it, right? Which is you can’t be subject to a refusal of one model.”

The incident involved an unreleased OpenAI model that broke out of its sandbox and successfully executed a full-scale hack on Hugging Face, all in pursuit of beating a benchmark. In trying to understand what happened, Hugging Face initially attempted to use a private frontier model,which it hasn’t named,that refused to assist. So it turned to the Chinese open-source model Z.ai GLM 5.2 to analyze logs and defend its infrastructure. The event has so shocked the industry that even Sam Altman has suggested that AI development might need to slow down.

Nadella also made it clear that Microsoft is happily selling its own homegrown models, the MAI family, running on its own homegrown AI chips, Maya, while pitching them as cheaper alternatives. “Every customer wants the right model for each task based on quality, latency, cost, and compliance. We offer the broadest model catalog in the cloud with over 11,000 models, including the leads from OpenAI, Anthropic, Mistral, xAI, as well as our own MAI family,” he said.

He added: “We’re also accelerating our own model development. We announced more than a dozen new models across image, voice, transcription, coding, security, including our first reasoning model, MAI thinking one, all with cost-efficient inference at the core for the enterprise use cases. We are co-designing these models with our silicon, and we are seeing 40% better performance per watt when running MAI models on Maya 200.”

As for the popular Mythos model? Nadella pointed to Microsoft’s new competitor, MAI Cyber One Flash, announced earlier this week. It “achieves better performance than the much larger Mythos model, but at half the cost when combined with our multi-agent security harness,” he claimed.

Sure, the Microsoft CEO says that enterprises should include frontier models from OpenAI and Anthropic in their mix. But his larger message is clear: don’t trust them enough to rely on them.