Alibaba Plans to Share in Revenue from Its Next AI Model

Vienna

The Chinese technology conglomerate Alibaba is planning a new billing model for its upcoming Qwen language model. Instead of fixed usage fees, large commercial customers will in the future pay Alibaba a share of the revenue they generate using the model. The model could be introduced as early as next week, but according to available information, the exact percentage of revenue to be shared is still the subject of ongoing negotiations.

From a Licensing Model to Revenue Sharing

Alibaba’s Qwen model series is one of the best-known open AI systems from China and is used by companies worldwide for their own applications. To date, Alibaba has primarily charged customers for usage based on computing power or through traditional licensing models. A shift to a revenue-sharing model would mean that Alibaba would directly participate in the financial success of applications built on Qwen.

This would set the company apart from the practices of many U.S. providers, such as OpenAI and Anthropic, which primarily bill based on usage volume. For Alibaba, this model could be particularly attractive to customers who integrate Qwen on a large scale into their own high-revenue products, such as in e-commerce, the financial sector, or cloud services.

Open-Weight Models Under Cost Pressure

Qwen is classified as a so-called open-weight model, whose weights are publicly available and can be customized by developers. Such models have gained significant prominence over the past two years because they give companies more control and reduce their dependence on individual cloud providers. At the same time, the costs of training and operating large models are rising steadily, prompting providers such as Alibaba to explore new revenue models.

A revenue-sharing model would allow Alibaba to benefit from smaller customers with rapidly growing applications without having to charge high upfront licensing fees. For startups and small-to-medium-sized businesses, this could make it easier to get started, but at the same time, it increases their long-term financial dependence on the provider.

Competition in the Chinese AI Market

In the Chinese AI market, Alibaba is in direct competition with companies such as Baidu, ByteDance, and DeepSeek, which are also developing their own language models and, in some cases, offering them for free or at low prices. A new pricing model could help Alibaba differentiate itself more effectively and generate higher revenue from large commercial customers without restricting broad, often free access for developers and researchers.

The planned launch comes at a time when many tech companies around the world are searching for viable business models for generative AI. Investments of billions in data centers and training infrastructure have often been out of clear proportion to the revenue generated so far, which is increasing the pressure to develop new pricing structures.

For European and Austrian companies that are increasingly integrating open-source Chinese AI models such as Qwen into their own products, a shift to revenue sharing would have concrete consequences. For example, companies in the domestic software or industrial sectors that rely on Qwen would have to calculate more precisely in the future what portion of their own revenue goes to the Chinese provider.

At the same time, this initiative highlights just how intense the global competition for AI business models is becoming beyond the well-known U.S. providers such as OpenAI, Google, and Microsoft—a factor that European regulators and companies are also likely to take into greater account when choosing their technology partners in the future.

Source: TechNode | Original Article