Alibaba’s First Qwen Conference Just Put Affine on the Stage

Affine (SN120) has been invited by the Qwen team at Alibaba to speak at the first-ever global Qwen Conference, taking place May 26, 2026 at the Sands Expo & Convention Centre in Singapore.

The team will give a special talk on environment design and post-training techniques built directly on top of the Qwen ecosystem.

The recognition matters because it is one of the first formal acknowledgments by a major foundation model lab that meaningful agentic and post-training work is happening inside the Bittensor ecosystem.

Affine sits squarely at the center of that thesis. 

What the Conference Actually Covers

This is Qwen’s first-ever global summit, themed around the intersection of AI and real-world business transformation, with an agenda built around a flagship keynote, focused forums, and a 1,000 square meter immersive exhibition.

The day breaks into three core blocks:

a. The Keynote, led by Dr. Feifei Li, CTO of Alibaba Cloud and President of International Business, alongside global AI leaders covering the latest breakthroughs across Qwen models, MaaS services, and next-generation agent-native infrastructure.

b. The AI x Web3 Innovation Workshop, a focused forum on how AI is reconstructing digital asset trading and decentralized ecosystems.

c. The Qwen Exhibition, a 1,000m² interactive showcase covering foundation models, multilingual AI services, agent platforms, AI coding tools, and enterprise AI applications.

The structure signals that Alibaba is treating the convergence between frontier AI infrastructure and decentralized ecosystems as a core thread of the conference rather than an adjacent novelty, which is exactly the lane Affine has been building in.

Why Affine’s Recognition Actually Matters

Affine’s invitation reflects a real technical relationship rather than a ceremonial gesture, and the substance of what they will be presenting is worth tracking carefully:

a. Affine’s work is built directly on top of the Qwen ecosystem, with the team’s environment design and post-training techniques operating against Qwen’s open-source foundation models rather than on a parallel stack.

b. The post-training layer is exactly where Bittensor’s incentive structure has been outperforming, and Affine is one of the clearest examples of how mining competition produces real improvements on top of frontier open-source models.

c. Singapore as the venue puts Affine in front of Asian institutional capital and enterprise audiences that have so far been less exposed to Bittensor than their North American counterparts.

d. The session will publicly showcase what decentralized post-training actually looks like, which is a category most foundation model labs have not yet acknowledged as a credible alternative to in-house research.

What this lands on is the broader recognition that the most interesting work happening on top of frontier open-source models is increasingly being done through incentive-driven networks rather than centralized research labs.

Affine speaking in front of Alibaba’s leadership team is the first significant public demonstration of that thesis at the foundation model level. Also, the convergence between Bittensor and Qwen that this conference is bringing into focus will be worth watching closely through the rest of the year.