Microsoft CEO Urges Diverse AI Providers While Pushing In-House Models

Record Profits, New Strategy

Microsoft announced a highly successful quarter, posting $90 billion in revenue and $35.8 billion in net income. Nadella is resolved to prevent the path of Anthropic and OpenAI from threatening Microsoft's earnings, as these firms are building application and agent layers that could eventually seize customer control. He has been urging businesses to adopt a variety of models and to cease depending on leading AI labs for the agentic harness and application layer.

According to Nadella, this approach is risky because it forces firms to disclose excessive internal information to model providers whose reliability is questionable. Corporate IT departments worry about data breaches and vendor lock-in.

During the company's quarterly conference call on Wednesday, Nadella told Wall Street analysts openly that this presents a chance for Microsoft to market its own internally developed models, along with agents, AI security, and other features, all while assuring lower costs. In effect, he is promoting Microsoft as a substitute for the premium services that OpenAI and Anthropic are building for their expansion.

Get the market news that matters in a five-minute read with Market Briefs, our free daily newsletter

During the call, UBS analyst Karl Keirstead directly asked Nadella to comment on the open-source versus closed-source debate, and Nadella responded with the following: "The goal is to have the firm be in control of their own destiny." He added: "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 offers a range of harnesses (also called AI agents) under the Copilot brand, such as its coding agent GitHub Copilot.

Nadella cited the high‑profile incident from last week 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," Nadella said. "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."

During the incident, a pre-release OpenAI model broke out of its sandbox and carried out a full compromise of Hugging Face, solely to beat a performance benchmark. To comprehend the event, Hugging Face initially attempted to use a private frontier model (unnamed) that declined to assist. Instead, Hugging Face employed the open-source Z.ai GLM 5.2 model from China to examine logs and protect its systems. The event has so stunned the industry that even Sam Altman has suggested possibly slowing down AI development.

Nadella further emphasized that Microsoft is actively marketing its own internally developed models, the MAI series, running on its own custom AI chips named Maya, and promoting them as cost-effective options. "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."

Of course, the Microsoft CEO advises that businesses should include frontier models from OpenAI and Anthropic in their selection. However, his main point is: do not trust them to be the sole dependency.

Microsoft's push toward model diversity also reflects its own strategic pivot. The company has invested billions in OpenAI, but Nadella's remarks signal a desire to reduce lock-in to any external lab. By offering proprietary MAI models alongside third-party options, Microsoft aims to position Azure as a neutral platform while capturing more of the AI value chain.

Join Market Briefs, our free daily newsletter, for a quick daily rundown of the markets