I just noticed a real story happening in the AI world. Lin Junyang, the man who led the Qwen team to the top of open-source large models globally, suddenly left. And the story behind this resignation is much deeper than it appears at first glance.

It all started last March. On March 3rd, there was a major strategic meeting at Alibaba focused on artificial intelligence — a clear sign that the group wanted to "focus entirely" on this field. On the same day, Lin Junyang's team released four small models of Qwen 3.5, and even Elon Musk commented, "Incredible intelligence density." By all measures, it was a great day.

But in the evening, something happened. According to multiple sources, Lin Junyang left the internal meeting due to disagreements and resigned immediately. The next morning, he posted a simple message on X: "me stepping down. bye my beloved qwen." — and thus, Qwen lost its leader.

The real question here: Was this a voluntary resignation? Sources close to the situation say no. One of the key contributors to the team commented directly: "I am truly heartbroken, I know that leaving was not your choice." This indicates there was pressure from above.

What actually happened? It seems there was a fundamental disagreement over how to develop large models. Lin Junyang believes that strength comes from integrated collaboration between teams. But Alibaba’s top management wants to divide the work into separate horizontal units — a line-production-like model.

Behind this, there is a deeper tension: the gap between technical vision and business goals. Qwen reached the top thanks to a strong open-source strategy — without immediate commercial pressure. But now, the group wants real commercial returns. This means priorities are shifting.

The real impact? In just three months, several key founders resigned — the tech lead, the subsequent training lead, even the code manager. This is not just the loss of one person, but the loss of an entire team that built Qwen from scratch.

By January 2026, Qwen had over 200,000 derivative models on Hugging Face with more than a billion downloads. It ranked third worldwide in contributions. Now, it will face serious challenges in maintaining this momentum.

The biggest question: Where will Lin Junyang go? The most likely scenario is that he will start his own project or join another star team in the large model field. His staying seems highly unlikely.

For Alibaba, this marks a real turning point. The phase that focused on building a global technical standard has ended, and now a new phase begins focused on commercial transformation. But the cost may be higher than the group expects.