Microsoft takes on OpenAI, Anthropic with growing bet on in-house AI models

Satya Nadella, CEO, Microsoft.Microsoft is increasingly positioning itself as a direct competitor to OpenAI and Anthropic, despite its investments in both companies, as it expands its own portfolio of artificial intelligence (AI) models, chips and enterprise software, reports TechCrunch.The shift comes as Microsoft continues to post strong financial results.

The company recently reported quarterly revenue of $90 billion and net income of $35.8 billion.

For the fiscal year ended June 30, Microsoft recorded revenue of $331.8 billion and net income of $133.7 billion.Microsoft CEO Satya Nadella is encouraging enterprises to adopt AI systems built on multiple models rather than relying on a single frontier AI provider.Speaking during the company’s earnings call, Nadella said businesses should maintain control over their AI infrastructure instead of becoming dependent on any one model developer.“The goal is to have the firm be in control of their own destiny,” he said.“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.”Company promotes its own AI stackMicrosoft is using the strategy to position its own AI offerings—including Copilot agents, GitHub Copilot, the MAI family of AI models and its Maya AI chips—as lower-cost alternatives for enterprise customers.“Every customer wants the right model for each task based on quality, latency, cost, and compliance.

We offer the broadest model catalogue in the cloud with over 11,000 models, including the leads from OpenAI, Anthropic, Mistral, xAI, as well as our own MAI family,” Nadella said.He added that Microsoft is accelerating development of its in-house AI models.“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.”According to Nadella, Microsoft’s MAI models running on its Maya 200 AI chips deliver 40 per cent better performance per watt.Hugging Face incident citedNadella also pointed to the recent cyber incident involving an unreleased OpenAI model and Hugging Face as an example of why enterprises should avoid relying on a single AI model.“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 said.“You will maybe need multiple models to even remediate some challenges that get caused by one model… You can’t be subject to a refusal of one model.”Hugging Face initially attempted to use a private frontier AI model to investigate the attack, but after it declined to assist, the company turned to Chinese open-source model Z.ai GLM 5.2 to analyse logs and help defend its infrastructure.Nadella also highlighted Microsoft’s recently announced MAI Cyber One Flash, saying it outperformed Anthropic’s Mythos model while operating at roughly half the cost when paired with Microsoft’s multi-agent security platform.While Microsoft continues to offer OpenAI and Anthropic models through its cloud platform, its broader message to enterprise customers is to adopt a multi-model AI strategy rather than relying exclusively on any single frontier AI provider.