Table of Contents
Alibaba is preparing a new commercial approach for its upcoming qwen 3.8 max AI model, according to a Reuters report citing two people familiar with the company’s plans.
The company is reportedly considering requiring large commercial users to share a portion of the revenue they generate from products and services built using the model. The change could take effect as early as next week, although Alibaba has not publicly confirmed the final terms.
The reported move would represent a shift from Alibaba’s previous approach, under which most customers could run its open-weight AI models in their own data centres without paying licensing fees.
Model Weights Expected to Remain Freely Downloadable
The proposed arrangement would not appear to eliminate free access to the qwen 3.8 max model weights.
Instead, the reported revenue-sharing requirement would focus on larger companies using the model as the foundation for substantial commercial services.
This could allow developers and researchers to continue downloading and experimenting with the model without a conventional licensing charge, while major commercial deployments could require separate agreements with Alibaba.
However, the final commercial threshold, licensing conditions and revenue-sharing percentage have not been announced publicly.
Alibaba’s Strategy Mirrors Moonshot’s Kimi K3
Alibaba’s reported approach resembles a commercial model already being used by Chinese AI startup Moonshot.
Under the reported terms for Kimi K3, companies that provide the model as a commercial service and generate more than $20 million in annual sales must negotiate a revenue-sharing agreement.
One person familiar with Moonshot’s arrangements said the company can require partners to share up to 30% of the relevant revenue.
Alibaba is reportedly considering a similar revenue-sharing mechanism for qwen 3.8 max, but the exact percentage remains under discussion.
Chinese IT services company Chinasoft International disclosed that it had entered a revenue-sharing agreement with Moonshot in a regulatory filing last month, although it did not reveal the percentage involved.
A Freemium Model for Commercial AI
The proposed strategy resembles a freemium model in which basic access remains available while larger commercial users may have additional financial obligations.
Paddy Srinivasan, CEO of DigitalOcean, described the approach as a tested open-source “freemium” model. DigitalOcean is among the U.S. companies offering Kimi K3 and other Chinese AI models.
Srinivasan confirmed that DigitalOcean has a commercial agreement with Moonshot but declined to provide details about its terms.
For Alibaba, a similar structure could allow the company to encourage widespread adoption of qwen 3.8 max while creating a potential revenue stream from companies that turn the model into large-scale commercial products.
Why Alibaba May Be Changing Its Approach
The reported plans come as Chinese AI companies increasingly compete with major Western developers.
Chinese firms have released open-weight models that have attracted attention for their capabilities and pricing, putting pressure on developers of closed AI systems from companies such as OpenAI, Anthropic and Google.
That competition has also created pressure on AI companies to find sustainable business models.
Rather than charging every developer for access to model weights, revenue sharing could give companies an alternative way to benefit financially when their models become the foundation of successful commercial services.
The approach could also become part of a broader debate over how open-weight AI should be licensed when models are freely downloadable but commercial deployments generate significant revenue.
Geopolitical Tensions Continue
The development comes amid continuing tensions between the United States and China over artificial intelligence and technology.
The White House has accused Moonshot of stealing technology from Anthropic. Chinese officials have rejected those allegations as unfounded.
These claims remain part of the wider geopolitical debate surrounding Chinese AI companies and should not be treated as established findings.
At the same time, open-source AI development is continuing in the United States.
Thinking Machines Lab, founded by former OpenAI chief technology officer Mira Murati, released its first open-source model last month. Fireworks AI CEO Lin Qiao has also argued that there is no fundamental barrier to the development of powerful open-source models in the United States.
Alibaba Has Not Finalised the Qwen 3.8 Max Terms
Despite the reported plans, Alibaba has not publicly confirmed the final commercial licensing structure for qwen 3.8 max.
The company has not disclosed the percentage of revenue it might seek from large commercial users, nor has it publicly confirmed all of the conditions that would determine which businesses would need to negotiate an agreement.
The reported timing should therefore be treated as a planned change rather than a finalized policy.
What the Move Could Mean for AI Developers
If Alibaba implements the reported strategy, qwen 3.8 max could demonstrate a different way of monetising open-weight AI.
Smaller developers and researchers could continue to benefit from freely downloadable model weights, while companies generating substantial revenue from commercial deployments could potentially share part of that income with the model provider.
Moonshot’s Kimi K3 already provides an example of this approach, and Alibaba’s reported plans suggest other AI companies may be exploring similar models.
The final licensing terms will determine whether the strategy encourages broader adoption or creates additional considerations for companies planning to build commercial products around qwen 3.8 max.
Conclusion
Alibaba’s reported plans for qwen 3.8 max point to an evolving business model for open-weight artificial intelligence.
The model weights are expected to remain freely downloadable, but large companies generating significant commercial revenue from deployments may be required to negotiate revenue-sharing arrangements.
The strategy follows a similar approach used by Moonshot and reflects the wider challenge facing AI companies: making powerful models widely accessible while developing sustainable ways to generate revenue.
Because Alibaba has not yet publicly confirmed the final terms, the scope and impact of the proposed qwen 3.8 max licensing model remain to be seen.