$50 billion. That's what Amazon committed to OpenAI last year in one of the largest cloud infrastructure deals in the industry's history. The problem: Microsoft had contractual grounds to object, creating a legal risk that hung over the entire arrangement.
According to TechCrunch's reporting, OpenAI secured the concessions from Microsoft needed to proceed with the AWS deal without liability. Under the restructured agreement, OpenAI can sell products on Amazon Web Services without triggering breach-of-contract claims. Microsoft receives additional revenue-share payments in return - a cut of the commercial activity it's allowing OpenAI to run on a competing cloud platform.
The Leverage Shift
The resolution reflects a fundamental change in who holds the stronger position in this relationship. When Microsoft made its initial investment, OpenAI depended on Azure compute and had no comparable alternative. Now OpenAI carries a $300 billion valuation, has commercial demand from multiple cloud providers, and is mid-conversion to a for-profit structure that will only accelerate that independence.
Microsoft's calculation appears to be that an independent, commercially successful OpenAI is worth more than a constrained one. If OpenAI's growth stalls because it can't do business on AWS, the value of Microsoft's $13 billion stake stalls with it. The revenue-share structure turns what could have been a damaging legal fight into a toll arrangement.
What AWS Customers Get
For teams building on AI, the outcome has a direct practical effect. ChatGPT and OpenAI's API will be commercially available on both AWS and Azure, with full support on both platforms. AWS customers who wanted GPT-4o and newer models no longer face pressure to move to Azure to access them.
The broader dynamic is worth understanding: cloud providers need AI companies' models to differentiate their platforms, and AI companies need cloud providers' infrastructure and customer reach. That mutual dependency gave OpenAI negotiating leverage - and it used it.