AWS Partners with Moonshot AI for Revenue Sharing

Amazon Web Services has entered into a revenue-sharing agreement with Moonshot AI, the developer of the Kimi K3 model. This deal, reported by 36kr.com, marks a significant departure from traditional licensing models where technology providers extract fees from hardware manufacturers without sharing downstream profits.

Historically, companies like Qualcomm and Microsoft collected substantial licensing fees from Chinese hardware makers, often leaving manufacturers with thin margins. The new arrangement reverses this dynamic, allowing a Chinese AI firm to retain a portion of the revenue generated when its models are used through US cloud platforms.

Cloud giants bypass closed-source rivals

AWS selected Kimi K3 after evaluating its performance, finding it comparable to leading closed-source models from competitors like Anthropic. By choosing a third-party model for its Bedrock service, AWS reduces its dependence on exclusive partnerships with US-based AI labs.

This move disrupts the bargaining power of closed-source providers, who previously captured a large share of the value chain. For cloud vendors, offering diverse model options helps attract enterprise customers who may prefer alternative architectures or cost structures.

Opening capabilities to industry partners

Moonshot AI is simultaneously launching the Moon Landing Plan, a partnership initiative with companies like Kingsoft Cloud and Chinasoft International. This program aims to integrate large language models into core enterprise software, creating a broader ecosystem of applications.

Unlike US competitors who often seek to control the entire stack, Moonshot AI focuses on the model layer while allowing partners to build applications. This approach distributes value more evenly across the industry, ensuring that software developers and infrastructure providers also benefit from AI adoption.

Shifting global profit distribution dynamics

The collaboration signals a structural change in how AI value is distributed. Instead of a single entity capturing most of the profit, the model now encourages a win-win scenario where hardware, cloud, and application layers all share in the growth.

This shift challenges the long-standing dominance of US tech firms in setting the terms of global technology trade. By leveraging competitive model performance, Chinese companies are now able to negotiate from a position of strength, altering the economic landscape of the AI industry.