Alibaba Group’s latest AI video model has risen to the top of global benchmarks. This shows that Chinese companies are increasing their competitiveness in tools used for advertising, content creation and entertainment.
summary
- Alibaba Group’s HappyHorse-1.0 topped the global AI video benchmark shortly after its anonymous debut in artificial analysis.
- The model was later confirmed to be part of Alibaba’s ATH AI innovation unit, driving the rally in the Hong Kong-listed stock.
- The setbacks of rivals by OpenAI and ByteDance have created space for Alibaba to strengthen its position in the video generation space.
The model, called HappyHorse-1.0, first appeared on benchmarking platform Artificial Analysis on April 7, but its origins were not disclosed at the time. It quickly rose to the top of blind test rankings in both the text-to-video and image-to-video categories, garnering widespread attention across the industry.
The model’s developer later revealed that the model is part of Alibaba’s ATH AI innovation unit and is still under active development. The announcement was made through a newly created account on X, confirming the company’s involvement after several days of uncertainty.
Prior to the release, speculation had increased regarding the identity of the developer, with speculation ranging from Tencent to Alibaba or even an independent team.
Following this confirmation, Alibaba shares closed 2.12% higher in the Hong Kong market on Friday, shortly after Alibaba confirmed its involvement in the project. The stock had risen 6.75% earlier this week, led by a rebound in tech stocks, as tensions between the U.S. and Iran showed signs of easing and speculation about the model intensified.
The company has expanded its AI presence in China’s competitive market, building on Qwen’s large-scale language model and chatbot ecosystem. We have introduced video generation capabilities before, but none achieved the same level of attention and ranking performance in such a short period of time.
HappyHorse-1.0 could improve Alibaba’s position in video generation, an area where competitors have recently faced hurdles. One of its competitors, OpenAI, recently retired its Sora video platform in favor of coding tools, enterprise services, and AGI development due to high computing demands. Meanwhile, ByteDance has paused the rollout of its Seedance 2.0 model following copyright disputes with major studios and streaming platforms.
Chief Executive Officer Eddie Wu has put AI at the center of Alibaba’s long-term strategy, which includes investments in chips, cloud infrastructure and data centers. The company has a history of incorporating its models into e-commerce, advertising, and entertainment services, and a similar integration path could be considered for HappyHorse.
Parallel to the development of video generation, Alibaba is expanding its underlying infrastructure. The company is working with China Telecom on a new data center project in southern China as part of a national effort to boost domestic computing capacity.
The facility will be equipped with 10,000 Alibaba Zhenwu AI chips designed to handle both training and inference workloads. The system is built to support models with hundreds of billions of parameters, making it one of the most advanced clusters in operation today.
Alibaba says the chips will work as an integrated system, allowing the cluster to operate like a single supercomputer with about 4 microseconds of latency, increasing the efficiency of large-scale AI tasks.
In another development, Alibaba Cloud revealed that it has led a funding round of approximately $275 million in Chinese AI startup Shengshu Technology. Baidu Ventures and Luminous Ventures also participated in this round, following a 600 million yuan raise completed just two months ago.
Shengshu Technology develops video generation tools through its Vidu platform, competing with products from emerging startups such as ByteDance, Alibaba, Kuaishou, and PixVerse. This investment signals that competition in this space will continue, even as infrastructure and model development accelerate in parallel.
