Did Alibaba just kneecap its powerful Qwen AI team? Key figures depart in wake of latest open source release

2026-03-03 · Source: VentureBeat · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Intermediate, medium

Summary

Alibaba's Qwen AI team, known for its prolific open-source generative models, is experiencing significant leadership changes following the release of its Qwen3.5 small model series. Junyang "Justin" Lin, the technical architect who led Qwen to over 600 million downloads, along with staff research scientist Binyuan Hui and intern Kaixin Li, announced their departures from the company. This exodus, occurring just 24 hours after the Qwen3.5 release, raises concerns about the future direction of Alibaba Cloud's AI strategy, particularly its commitment to open source. The Qwen3.5 models, ranging from 0.8B to 9B parameters, feature a Gated DeltaNet hybrid architecture, enabling a 9B-parameter model to rival larger systems while maintaining a 262,000-token context window and efficient operation on standard devices. Alibaba has consolidated its AI efforts into the "Qwen C-end Business Group," with Hao Zhou, formerly of Google DeepMind, reportedly taking over leadership, signaling a shift towards monetization and product-centric development.

Key takeaway

For CTOs and VPs of Engineering relying on Qwen's open-source models, the leadership changes at Alibaba suggest a pivot away from a research-first, open-source strategy towards aggressive monetization. You should evaluate the potential for future flagship models, like the rumored Qwen3.5-Max, to be locked behind proprietary APIs. Consider downloading and preserving current Apache 2.0-licensed Qwen models now if continued open access is critical to your operational strategy.

Key insights

Key Qwen AI team members' departure signals Alibaba's shift from open-source research to monetization and product-driven strategy.

Principles

• Algorithm-hardware co-design bypasses compute constraints.
• Open-weight models offer transparency and performance.

Method

The Qwen3.5 small model series employs a Gated DeltaNet hybrid architecture with a 3:1 ratio of linear to full attention, enabling high intelligence density and efficient operation on edge devices.

In practice

• Qwen3.5 models run natively on laptops and smartphones.
• Models can serve as "all-in-one AI workers."

Topics

• Qwen AI Models
• Open-Source AI
• AI Team Departures
• AI Monetization
• Gated DeltaNet Architecture

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Researcher, Machine Learning Engineer, AI Product Manager

Related on AIssential

• Something is afoot in the land of Qwen — Alibaba's Qwen team is experiencing significant upheaval, with lead researcher Junyang Lin and several other core members, including Binyuan Hui and Bowen Yu, announcing their resignations on March 4th, 2026.
• Open-Source AI Gains Ground as Rising Costs Push Shift to Smaller Models — Open-weight AI models offer substantial cost savings and near-parity performance compared to closed models, yet remain significantly underutilized.
• Is Alibaba’s $53B AI Bet in Trouble? Qwen Architects Exits — Key architect departures can significantly impact investor confidence and project direction, especially in large AI initiatives.
• Alibaba just open-sourced a model that rivals GPT-5.2 — Alibaba's Qwen 3.5, with 17 billion active parameters, rivals top closed-source models while being open-weight.
• Alibaba launches Qwen 3.5 AI model amid intensifying China chatbot race — Alibaba's Qwen 3.5 model offers significant cost reduction and workload capacity for agentic AI tasks.
• Last Week in AI #333 - ChatGPT Ads, Zhipu+Huawei, Drama at Thinking Machines — The AI industry is rapidly evolving with new monetization strategies, intense talent competition, and efforts to reduce hardware dependency.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by VentureBeat.