Microsoft has reportedly decided to stop paying royalties to OpenAI for using its models in first-party products like Copilot and Bing.
This move is primarily a response to immense financial pressure. Microsoft is investing billions—a staggering $37.5 billion in capital expenditures in a single recent quarter—to build the infrastructure needed for AI. To make this investment worthwhile, the company must improve the profitability of its AI products. Cutting down on royalty payments is a direct way to reduce the cost of goods sold (COGS) and improve margins as it scales services like Copilot to millions of users.
Several key factors created the perfect conditions for this strategic shift. First, the relationship between the two companies has grown more complex. OpenAI's recent partnership with Amazon's AWS, a major Microsoft competitor, created channel conflict. This friction gave Microsoft a strong incentive to reassess its financial arrangements. Second, Microsoft is no longer solely dependent on OpenAI's technology. By developing its own powerful MAI models, Microsoft has gained significant bargaining power and the ability to substitute OpenAI's models in some of its products. Third, the underlying cost of AI is falling. Advances like NVIDIA's Blackwell platform are drastically reducing inference costs, making it cheaper to run AI models. This technological shift gives Microsoft more room to negotiate or eliminate royalty fees without hurting product viability.
It's important to understand that this is not a complete separation. The Microsoft-OpenAI partnership is a two-way street. While Microsoft may stop paying OpenAI for use in its own products, OpenAI continues to have a revenue-sharing agreement where it pays a substantial portion of its revenue (reportedly around 20%) back to Microsoft for using its Azure cloud infrastructure. This change appears to be an asymmetric adjustment—Microsoft is trimming its expenses while its revenue stream from OpenAI remains secure.
In essence, this decision reflects a maturing of the AI landscape and the complex relationship between two of its biggest players. Microsoft is making a calculated move to control costs, leverage its own technological advancements, and strengthen its competitive position in the rapidly evolving AI market.
- Capex (Capital Expenditures): Money a company spends to buy, maintain, or upgrade physical assets like servers and data centers.
- COGS (Cost of Goods Sold): The direct costs of producing the goods or services a company sells. For Microsoft's AI, this includes server costs and model royalties.
- Inference: The process of using a trained AI model to make predictions or generate outputs based on new data.