companies

SenseTime

Hong Kong-listed computer vision pioneer that expanded into large language and generative models under its SenseNova umbrella.

SenseTime was founded in Hong Kong in 2014 by the late Tang Xiao'ou, a computer vision researcher, and built its early business on facial recognition, image analysis, and autonomous driving perception software sold to governments and enterprises across China. It listed on the Hong Kong Stock Exchange in December 2021 and has since diversified into large-model AI, positioning generative AI as a new growth line alongside its established vision business as facial-recognition contracts came under greater scrutiny. Its SenseNova model family, launched in 2023 and rebuilt several times since, spans SenseChat for conversational AI, SenseMirage for text-to-image generation, and domain products like SenseChat-Medical and Code Raccoon for developers. In April 2026 the company open-sourced SenseNova U1, a unified model combining multimodal understanding and generation that it positioned as a technical answer to the fragmented, single-purpose model releases common elsewhere in the industry. SenseTime employs roughly 2,000 to 3,400 people depending on the reporting period, and continues to sell both to Chinese enterprises and, through a Cantonese-language version of SenseChat, directly to Hong Kong consumers.

Founded

2014

Headquarters

Hong Kong

Sector

lab