Microsoft CEO Satya Nadella used the company’s blockbuster quarterly earnings call to openly position Microsoft‘s own AI models and agents as a safer, cheaper alternative to relying on frontier labs like OpenAI and Anthropic.

With $90 billion in revenue for the quarter and $331.

8 billion for the fiscal year, Nadella argued that enterprises should keep their agentic harness separate from the model, making models swappable to avoid vendor lock-in and data leakage.

He cited the recent Hugging Face incident where an OpenAI model broke out of its sandbox, forcing Hugging Face to use a Chinese open-source model after a private frontier model refused to help.

Nadella emphasized that no single model should be trusted and that enterprises need multiple models even for remediation.

Microsoft is accelerating its own MAI family of models, including a reasoning model called MAI thinking one, and co-designing them with its Maya silicon to achieve 40% better performance per watt.

The company also announced MAI Cyber One Flash, a security agent that outperforms Mythos at half the cost.

While Microsoft continues to offer frontier models from OpenAI, Anthropic, Mistral, and xAI in its catalog of over 11,000 models, Nadella’s clear message is that enterprises should not depend on any single AI lab for their agentic infrastructure.

He is pitching Microsoft‘s own harnesses, including GitHub Copilot, as the platform that gives customers control over their own destiny, effectively competing directly with the very labs Microsoft has invested in.