Microsoft Corp. (NASDAQ:MSFT) used its fiscal fourth-quarter earnings call to deliver upbeat updates on Azure, Copilot and AI spending. But one of the company’s most consequential messages for investors wasn’t about cloud growth or capital expenditures — it was about the future role of AI models themselves.
• Microsoft stock is charging ahead with explosive momentum. What’s behind MSFT gains?
Throughout the call, CEO Satya Nadella emphasized that Microsoft’s AI strategy is increasingly built around model choice rather than dependence on any single AI provider. Instead of betting exclusively on OpenAI, Microsoft is building a platform where customers can mix and match models based on cost, performance and use case.
Microsoft Is Building a Multi-Model AI Platform
While OpenAI remains one of Microsoft’s closest AI partners, Nadella said Azure now offers customers broad model choice.
“We offer the broadest model catalog in the cloud with over 11,000 models, including the latest from OpenAI, Anthropic, Mistral, xAI, as well as our own MAI family,” he said.
Microsoft said customer adoption is accelerating. Nadella said the company has seen “5x increase in the number of customers building with models from multiple providers.”
He cited Levi Strauss & Co. (NYSE:LEVI), which is using both OpenAI and Anthropic models through Microsoft’s Foundry platform while deploying more than 1,000 domain-specific AI agents.
‘Every Model Is Substitutable’
The clearest signal came when Nadella explained how Microsoft is designing its AI platform.
“We are building a new model system where the harness, context, memory, and action space are separate from any one model family.” He said the approach improves more than just efficiency. “It also has the added benefit of business continuity and resilience because every model is substitutable.”
The theme carried into the analyst Q&A. CFO Amy Hood said Microsoft is intentionally separating the software layer from the underlying model so enterprises can switch providers over time.
“You’ve got to keep your harness separate from the model… That means any given model at any given time is swappable,” Hood said.
Nadella reinforced the point moments later, arguing enterprises shouldn’t become dependent on a single AI provider. “The biggest thing that we should take away from that is you can’t depend on any one model,” he said.
Why It Matters For Investors
For years, Microsoft’s AI narrative has been closely tied to OpenAI. This earnings call suggested the company’s competitive advantage increasingly lies elsewhere: becoming the infrastructure layer that can host OpenAI, Anthropic, xAI, Mistral and Microsoft’s own MAI models side by side.
That strategy appears to be resonating. Azure revenue grew 43% year over year during the quarter, while Microsoft said customers are increasingly adopting multi-model AI architectures. Microsoft Cloud revenue reached $214.4 billion for the fiscal year, and commercial remaining performance obligations climbed to $678 billion, underscoring continued enterprise demand for the company’s AI platform.
Photo Courtesy: Tada Images on Shutterstock.com
Login to comment