($) "Death" of Qwen and the “Hundred Claws War”

This post originally appeared in Interconnected.

“In China, this apprehension often manifests as a frantic rush to ‘learn the newest things’ so as not to fall behind.”

ICYMI: I went on the Kyle Chan’s High Capacity podcast and Jordan Schneider’s ChinaTalk to talk about my recently published long piece on the history of Chinese open source. If you haven’t read the piece, or prefer an audio digest of it, please give those two podcast episodes a listen!

---

Today’s post addresses two seemingly unrelated topics that are nonetheless (inter)connected by a common thread: open source.

One is the latest update on Alibaba’s AI strategy since the dramatic reorganization of its Qwen team, which I wrote about when the news first broke two weeks ago. The other is the sudden take off of OpenClaw, the open source agent framework – a seemingly inexplicable rise in China that actually has very good reasons behind the phenomenon. This phenomenon has quickly spawned into a so-called “Hundred Claws War” (百虾大战), with at least 30 Claw products released and counting.

Let’s dig it.

“Death” of Qwen As We Know It

When I initially analyzed the rather dramatic shake up of the Qwen team, with its leader Junyang Lin resigning publicly on X/Twitter, my thinking was that the company is shifting its AI strategy to be more commercially focused, but its open source commitment to Qwen will remain in tact for at least the next quarter or two, before more measures to close off future models in favor of revenue generation are introduced.

Looks like this shift is happening much faster than I anticipated.