Chinese AI developers may shift to ‘paid weights’ commercial licensing: Goldman Sachs

The Shifting Landscape of AI Licensing

The landscape of artificial intelligence development in China is poised for a significant transformation, as suggested by Goldman Sachs. Historically, many Chinese AI developers have released their sophisticated models under permissive open-source licenses, such as the MIT License. This approach has allowed foreign platforms and developers to freely download, modify, and host the core 'weights'—the underlying parameters that define a model's intelligence—even for commercial purposes. This strategy has fostered widespread adoption and collaboration, but it has also limited the direct revenue streams for the original creators.

However, with Chinese AI models like Moonshot AI’s Kimi K3 and Zhipu AI’s GLM-5.2 now achieving performance levels remarkably close to leading US counterparts, the economic imperative to monetize these advancements is growing. The current open-source model, while promoting rapid dissemination, creates a revenue gap that developers are keen to address. This shift reflects a maturing industry where the focus is moving beyond just technological prowess to sustainable business models that can fund continued research and development.

Monetizing Open-Weight Models

Ronald Keung, a Hong Kong-based analyst at Goldman Sachs, highlights that Chinese model builders could substantially increase their revenue by requiring third-party providers to purchase commercial licenses. This would be necessary for platforms to serve these models—a process known as inference—on their own infrastructure. This 'paid weights' model would allow developers to capture value from the extensive use of their intellectual property, particularly as domestic adoption accelerates and global interest from small, medium, and even large enterprises grows.

The transition to commercial licensing for open-weight models represents a strategic move to convert widespread usage into tangible financial returns. Instead of relying solely on indirect benefits or ancillary services, developers would directly charge for the core technology. This model is common in other software industries and could provide a more stable and predictable revenue stream, enabling greater investment in future AI innovation and talent acquisition within China's competitive tech sector.

Implications for Global AI Competition

This potential shift has significant implications for the global AI ecosystem. For foreign platforms and developers currently leveraging Chinese open-source models for free, the introduction of licensing fees would introduce new costs and potentially influence their choice of foundational models. It could also spur greater investment in proprietary models or alternative open-source initiatives outside of China, depending on the pricing and terms of these new commercial licenses.

Furthermore, this move could intensify the competition between Chinese and US AI developers, not just on performance metrics but also on business models and market access. By establishing a clearer path to revenue, Chinese firms could gain more financial independence and resources to challenge global leaders more directly. It underscores a broader trend where nations are increasingly viewing AI as a strategic asset, seeking to control and monetize their technological advancements on the global stage.