Microsoft has spent billions trying to build an AI platform that can stand on its own. New figures suggest it still has a long way to go.
According to reporting from Windows Central, roughly 70% of Microsoft's AI business continues to rely on OpenAI's models. Despite heavy investment in proprietary AI systems and new partnerships, OpenAI remains the engine powering much of Microsoft's commercial AI offering.
That dependence highlights how difficult it is for even the world's largest technology companies to untangle themselves from the partnerships that helped launch the generative AI era.
"Microsoft wants AI independence. Its revenue still tells a different story."
A Partnership That Became a Dependency
Microsoft's relationship with OpenAI began as a strategic investment.
Today, it looks increasingly like a structural dependency.
Microsoft Copilot, Microsoft 365 AI features and large parts of Azure's generative AI platform continue to rely heavily on OpenAI's GPT models.
The company has launched its own models through its MAI division and expanded relationships with other AI developers, including Anthropic. But those efforts have yet to significantly reduce OpenAI's role across Microsoft's AI products.
That matters because OpenAI is no longer simply a partner.
It's increasingly becoming a competitor.
Why Microsoft Wants More Control
OpenAI has steadily expanded into enterprise software, consumer applications and other markets that overlap with Microsoft's core business.
Earlier this year, the two companies restructured their partnership, allowing OpenAI greater freedom to pursue new customers and infrastructure agreements while preserving Microsoft's access to its latest models.
The expectation was that Microsoft would gradually diversify its AI foundation.
The latest figures suggest that transition remains a work in progress.
The Business Risk Behind the Numbers
For a company of Microsoft's scale, relying so heavily on a single external AI provider creates an unusual strategic risk.
Large technology companies generally avoid single points of dependency, particularly in businesses expected to drive future growth.
If OpenAI changes pricing, shifts strategic priorities or experiences operational disruption, Microsoft's AI products could feel the effects almost immediately.
That exposure becomes even more significant as OpenAI develops products that compete more directly with Microsoft 365 Copilot and other enterprise offerings.
Microsoft's investment in proprietary AI models is intended to reduce that risk, but progress appears slower than many expected.
What Comes Next
Microsoft isn't standing still.
Its AI teams continue developing smaller, more efficient in-house models designed for specific business tasks, while executives have signalled a long-term strategy built around a mix of Microsoft-developed models, OpenAI technology and third-party providers.
Building that balance, however, won't happen overnight.
The AI market is evolving rapidly, and Microsoft is attempting to reduce its dependence on a company that remains central to its own commercial success.
What This Means for Miami
Many South Florida businesses already rely on Microsoft's AI ecosystem through Azure, Microsoft 365 and Copilot integrations.
As long as Microsoft's AI services remain closely tied to OpenAI, that dependency extends to local companies using those platforms for productivity, customer service and data analysis.
Startups building AI products on Azure should watch the relationship closely, as future changes to pricing, model availability or commercial terms could influence their own products and operating costs.
For Miami investors following enterprise AI, the story is also a reminder that even the industry's largest players remain dependent on critical technology partners. In the AI economy, supply-chain risk isn't limited to chips and data centres. It increasingly extends to the models themselves.
