The biggest partnership in AI just quietly dissolved.
On April 27, Bloomberg reported that Microsoft and OpenAI have ended the exclusive and revenue-sharing agreement that had defined their relationship since 2019. Under that deal, Microsoft was the only cloud provider allowed to run OpenAI models at scale, and OpenAI shared a cut of its revenue in exchange for Microsoft's billions in infrastructure investment.
That arrangement is now over.
Both companies confirmed the change. Microsoft says it will retain preferential access to OpenAI models as a non-exclusive partner. OpenAI, for its part, is now free to work with other cloud providers - including Google, Amazon Web Services, and anyone else with enough compute.
What It Actually Means
This was the deal that made Azure the default home for anything built on GPT-4, ChatGPT Enterprise, and OpenAI's API. If your business pays for any of those tools, you're currently on Microsoft infrastructure whether you know it or not.
The end of exclusivity does not mean OpenAI is walking away from Microsoft. Both sides emphasized the relationship continues. But it does mean:
- OpenAI can now optimize for the cheapest or fastest infrastructure rather than being locked to Azure
- Other cloud providers can bid for OpenAI workloads competitively
- The revenue-sharing arrangement that gave Microsoft a percentage of OpenAI's enterprise sales is reportedly ending
Think of it like a restaurant that was contractually required to use one food supplier exclusively. The contract expired. Now they can shop around. Prices may fall.
Why This Matters to Small Business AI Budgets
The AI tool pricing you see today was built in a world where OpenAI had limited negotiating leverage with infrastructure providers. That is no longer the case.
Competition between cloud providers for OpenAI workloads is likely to drive infrastructure costs down. Whether that savings gets passed to end users - the business owners paying $20 to $200 per month per seat for AI tools - depends on the market.
But the directional signal is clear: the era of "one partnership controls everything" is ending. More providers means more competition. More competition means pressure on pricing.
There is also a secondary effect. OpenAI is now reportedly in conversations to use Google's TPU infrastructure for some workloads. Google's Gen 8 TPUs launched this month and are widely considered more cost-efficient for inference than Microsoft's Azure GPU clusters at scale. If OpenAI shifts meaningful workload to Google's infrastructure, the savings in compute costs are potentially significant.
What Hasn't Changed (Yet)
Your existing tools are not going anywhere. If your business uses:
- ChatGPT Team or Enterprise - no change to your subscription or access
- Microsoft Copilot - Microsoft still has preferred access to OpenAI models and the product is unaffected
- Azure OpenAI Service - still available, still supported
- Any third-party tool built on OpenAI's API - no disruption
The operational layer is stable. The business layer is in motion.
The One Question to Watch
The old deal reportedly gave Microsoft a percentage of OpenAI's revenue as compensation for its infrastructure investment. That arrangement ending means OpenAI keeps more of what it earns.
Whether that translates to lower prices for end users, faster model updates, or simply higher margins for OpenAI is the open question. The competitive dynamics now exist for prices to come down. They do not guarantee it.
For any small business that is currently choosing between AI tools based on cost, the next 12 months are worth watching closely. The infrastructure monopoly that kept prices where they were no longer exists.
Sources: Bloomberg (April 27, 2026), Hacker News discussion thread #47921248, OpenAI and Microsoft public statements.