Happy Oyster by Alibaba

Happy Oyster did not appear from nowhere. It is the product of a specific team, a deliberate corporate strategy, and a fast-moving competitive landscape in Chinese AI. Understanding who built it and why helps explain what the model is designed to do and where it is headed.

The ATH Innovation Division

Happy Oyster was developed by the Innovation Business Unit under Alibaba's newly formed Alibaba Token Hub (ATH) business group. The unit is also known as the Future Life Lab of Taotian Group (Alibaba's e-commerce entity).

The team is led by Zhang Di, who brings significant AI video experience. Before joining Alibaba, Zhang Di served as Vice President of Kuaishou (one of China's largest short-video platforms) and was head of Kling AI technology, one of the leading AI video generation systems. This background directly informs the kind of products the ATH division builds: AI tools for visual content at scale.

The organizational structure matters because it shows Alibaba is not treating this as a research project. The ATH division sits within a business group tasked with commercializing AI technology, and Alibaba has publicly stated that cloud computing and AI are core growth engines with a revenue target of $100 billion annually within five years.

The Happy Horse connection

Happy Oyster is the 3D sequel to a 2D breakthrough. On April 10, 2026, just six days before Happy Oyster launched, the same ATH Innovation Division released Happy Horse, a 2D video generation model.

Happy Horse arrived with immediate impact. It ranked first on the Artificial Analysis text-to-video leaderboard with 1389 Elo points, beating the second-place Dreamina Seedance 2.0 by nearly 115 points. The model was initially released under a pseudonymous identity before Alibaba publicly confirmed ownership, creating significant buzz in the AI community.

Happy Horse uses a Transfusion architecture that integrates discrete text modeling with continuous visual signal generation in a single framework. It generates both high-quality video and synchronized sound effects from text prompts. API access was planned for April 30, 2026.

Happy Oyster builds on the same multimodal foundation but extends the output from 2D video to interactive 3D world simulation. The progression from Happy Horse to Happy Oyster shows a deliberate product roadmap: first establish leadership in 2D video generation, then use that technology base to enter the emerging 3D world model space.

Alibaba's broader AI strategy

Happy Oyster and Happy Horse are part of a larger pattern. Alibaba has been accelerating AI product releases and adjusting its internal organizational structure to strengthen commercialization capabilities. The company's AI portfolio now spans:

• Qwen: Alibaba's flagship large language model and chatbot
• Happy Horse: 2D AI video generation (top-ranked)
• Happy Oyster: 3D world simulation (newly launched)
• Cloud AI services: Enterprise AI infrastructure through Alibaba Cloud

Alibaba recently led a $293 million funding round in a Chinese AI startup, further signaling its commitment to the AI ecosystem. The company is competing directly with Tencent (which released HY-World 2.0 the day before Happy Oyster), ByteDance, Baidu, and international players like Google and OpenAI.

What this means for Happy Oyster's future

The corporate backing and team credentials suggest that Happy Oyster is not a one-off experiment. With a dedicated business unit, experienced leadership, and explicit revenue targets, Alibaba has both the resources and the motivation to develop Happy Oyster into a commercially viable product.

The current limited early access phase is likely a controlled rollout before broader availability, similar to how Happy Horse API access was staged. For the latest on availability, see Try Happy Oyster. For technical details on the model itself, check What Is Happy Oyster?.