SenseTime Group Inc. is a Chinese group specializing in artificial intelligence, with its initial expertise focused on computer vision: object recognition, video analysis and image processing. It markets software, models and computing capabilities for various sectors, from automotive and retail to urban services. Its website, sensetime.com, presents a portfolio now expanded to include generative AI, which is taking a central role in its strategy.
Academic origins followed by industrial expansion
Founded in 2014, SenseTime has its roots in the work of researchers at the Chinese University of Hong Kong, notably Tang Xiao’ou, a computer vision specialist and co-founder of the group. The company grew on the back of advances in deep learning and Chinese demand for automated image analysis tools. It gradually established a presence in several major cities across the country and in international markets.
Its listing on the Hong Kong Stock Exchange in 2021 marked an important milestone in this journey. This expansion has, however, been accompanied by controversy. In 2019, US authorities placed SenseTime on their trade restrictions list, citing its alleged involvement in surveillance and human rights abuses in Xinjiang. The group disputes these allegations. These measures, together with US investment restrictions, weigh on its business environment and international reputation.
From visual applications to generative models
SenseTime’s offering is built on several technological layers. Its SenseCore infrastructure combines computing resources and development tools to train and deploy AI models. The aim is to pool these resources to support a variety of applications, rather than develop each solution independently. The group offers technologies for scene analysis, medical imaging, digital content and interactions between users and machines, among other applications.
In the automotive sector, its SenseAuto platform covers uses related to driver assistance and smart cabins. For businesses and local authorities, its technologies can help make use of images or videos, raising issues of data protection and oversight of usage that are particularly sensitive when people are involved.
Introduced in 2023, the SenseNova family reflects its shift toward generative AI. It includes language processing capabilities and multimodal features, used notably in the SenseChat assistant. SenseTime is thus seeking to turn its research into services that can be integrated into business processes.
What next?
SenseTime’s trajectory will depend on its ability to convert these investments into recurring revenue while keeping computing costs under control. Access to advanced semiconductors, amid US export controls, is a strategic constraint. In the face of competition from major Chinese technology groups and generative model specialists, the company will also need to demonstrate the practical value of its solutions. Data governance and transparency about how its technologies are used will remain crucial to its credibility outside China.