Alibaba to Charge Big Users of Next Open-Source Qwen AI

Alibaba Rethinks the Meaning of 'Free' AI

Alibaba is preparing to ask the largest commercial users of its next-generation Qwen model for a cut of the revenue they generate from it, according to people familiar with the matter cited by Reuters 1. The move would apply to the upcoming version of Qwen, one of the most closely watched releases in the open-source AI world, and marks a notable shift for a company that has built its AI reputation on giving away powerful models for free 1.

The plan does not necessarily mean Qwen would stop being open-source in the traditional sense — code and weights could still be published — but it signals that Alibaba wants a commercial return from the companies profiting most from its work 1. That distinction sits at the center of a broader argument now playing out across the industry: what does 'open' actually mean when applied to AI, and can it sustain a business the way open-source software once did?

Open Weights Are Not Open Source

A growing body of commentary argues that Chinese 'open' AI models, while impressing developers with their capability and low cost, are structured very differently from the open-source software movement that preceded them 4. Critics note that releasing model weights is not the same as releasing the full recipe — training data, methodology and infrastructure choices often remain proprietary — which makes it harder for outside firms to build durable, defensible businesses on top of these models 4. Investors, in particular, have grown wary of the economics, since freely available weights make it difficult to charge for the model itself, pushing monetization toward services, infrastructure or, as in Alibaba's case, revenue-sharing arrangements 14.

Hardware and Security Stakes

The debate extends beyond business models into infrastructure and security. AMD CEO Lisa Su recently defended the open-source AI approach even after a security breach at Hugging Face tied to OpenAI agents raised concerns about the safety of open ecosystems, while also using the moment to promote AMD's newest chips aimed at AI workloads 3. Separately, analysts at Bank of America pointed to a wave of new open-source releases — including China's Kimi K3 — as reinforcing their bullish outlook on memory-chip makers like Micron, arguing that proliferating open models drive demand for the hardware needed to run them 5. Meanwhile, security researchers have flagged a shift in open-source malware away from cryptomining toward tactics focused on gaining persistent access to systems, a trend now affecting UK supply chains and underscoring the risks that come with widely distributed, community-maintained code 2.

Why It Matters

Taken together, these developments suggest the open-source AI landscape is entering a more commercially pragmatic — and contested — phase. Alibaba's revenue-sharing plan, if confirmed, would test whether large AI labs can keep models broadly accessible while still capturing value from their biggest beneficiaries, a balance the software industry spent decades refining and one the AI industry is only beginning to navigate 14.