Alibaba's Qwen Goes Viral Globally as AI Monetization Struggles Tear the Team Apart
Miles Bennett
Alibaba's open-source model Qwen hit 1 million daily downloads, yet AI product revenue covers less than 4% of total sales — a commercialization rift has cost the team its chief engineer and several core developers, turning monetization into an existential question for Alibaba's AI strategy.
How popular is Qwen — and how much money is it actually making?
Qwen is the world's most-downloaded open-source AI system. Hugging Face data shows roughly 1 million downloads per day as of January this year.
But traffic has not turned into revenue: Alibaba's AI-related product sales totaled about $1.3 billion in Q1 — less than 4% of total revenue.
This means → Qwen commands massive mindshare among developers, but Alibaba is barely earning from it. In plain terms = people are using it, just not paying for it.
$55 billion committed — where does the money go, and when does it come back?
Alibaba plans to invest over $55 billion in AI infrastructure by the end of next year.
Revenue of $1.3 billion against a $55 billion outlay puts the return ratio near 1:42 — a staggering gap.
This means → Alibaba is betting on future scale effects, not near-term profit. But if a viable monetization path fails to materialize, the investment itself becomes a source of pressure.
Why did the team fracture?
The New York Times reported that the Qwen team split over commercialization strategy. In March, chief engineer Lin Junyang resigned; several core engineers left around the same time.
The core disagreement: one camp wants to keep chasing frontier models in Silicon Valley's mold; the other believes technical leadership without commercial returns is unsustainable.
In plain terms = "build the best model" versus "make money first" — the two camps could not reconcile, and the key people walked.
Open-source to closed-source — what signal is Alibaba sending?
Alibaba's longstanding playbook: keep the most advanced models proprietary, release open-source versions alongside them. Now the company is accelerating toward paid, closed-source offerings — in April alone it launched three proprietary models in quick succession.
This reflects a management answer to the internal debate: commercialization priority is overtaking open-source expansion.
Yet open source is what built Qwen's global user base. Shifting to closed source means finding a new balance between user scale and paid conversion.
What unique pressures does Alibaba face on the AI track?
Brookings Institution fellow Kyle Chan notes that Alibaba has a clear-eyed view of its own position — it is part of China's national AI effort, and it carries the weight of not repeating past regulatory missteps.
Qwen's pricing sits well below the proprietary systems of U.S. rivals such as Anthropic and OpenAI; aggressive pricing is its current tool for capturing market share.
This means → Alibaba must compete with Silicon Valley on technology while navigating domestic policy constraints. The departure of key engineers turns the team's execution capacity itself into an open question.
Content is for reference only, not financial advice.