Chinese AI Firms Monetize Open-Source Models Through APIs

Chinese artificial intelligence companies are building viable business models around their open-source large language models through API services and commercial licensing, according to Goldman Sachs analyst Ronald Keung.

Speaking on CNBC's Squawk Box Asia, Keung said Chinese LLMs have reached "a critical stage of global adoption," with growing API usage creating what he described as a feedback loop that both improves the underlying models and expands developer ecosystems around them.

The open-source monetization strategy

While many Chinese AI firms have released their models as open-source — a stark contrast to the proprietary approach of companies like OpenAI — Keung explained that this doesn't preclude revenue generation. The analyst pointed to three primary monetization channels: commercial deployments for enterprise customers, licensing agreements, and revenue-sharing arrangements with partners who build applications on top of the models.

The API-driven approach allows developers to integrate Chinese LLMs into their applications while generating usage-based revenue for the model providers. As more developers adopt these APIs, the increased usage generates data and feedback that can be used to refine the models, potentially making them more competitive and attracting additional users.

Why it matters

This monetization framework challenges the assumption that open-source AI models cannot support sustainable businesses. If Chinese firms successfully scale API revenue while maintaining open model weights, they could establish a viable alternative to the closed, high-margin model that has defined the U.S. AI industry. For enterprises evaluating AI vendors, this suggests that openness and commercial viability are not mutually exclusive — a consideration that could reshape procurement decisions and competitive dynamics in the global AI market.

Global adoption dynamics

The "critical stage" Keung referenced suggests Chinese AI providers are moving beyond domestic markets and gaining traction internationally. This adoption phase is particularly significant given ongoing geopolitical tensions and export restrictions that have shaped the AI technology landscape.

The analyst's comments indicate that despite offering free access to model weights, Chinese AI companies are finding pathways to commercialization that could support long-term development and competition with Western counterparts.

The details were first reported by CNBC during the network's Asia market coverage.