Shenzhen Bureau of Industry and Information Technology: HQVT (01392) "Zhiyuan Origin Large Model" is proposed to be included in the official list of Shenzhen! It can enjoy the benefits of the model voucher policy.

Recently, the Shenzhen Municipal Bureau of Industry and Information Technology (MIIT) announced the proposed list of new model service institutions to be added to the registry for 2026. HQVT (01392) has successfully been included thanks to its independently developed "ZhiYuan Origin Model." As the department responsible for coordinating the development of the artificial intelligence (AI) industry in Shenzhen, the registry list issued by the MIIT serves as a significant market indicator. This inclusion of HQVT is expected to leverage Shenzhen's AI support policies to further expand its market space among local industrial clients. It is understood that Shenzhen has launched a series of policy tools including "model coupons, training vouchers, and corpus vouchers." Partner enterprises of registered institutions can apply for model coupons to offset part of the procurement costs of large models, thereby lowering the threshold for implementing AI technology. This initiative is intended to promote the empowerment of the real economy by AI technology, and more than ten institutions have already been incorporated into this service registry. The inclusion has certain mandatory prerequisites; the large model under evaluation must complete registration with the National Internet Information Office for generative AI. HQVT's "ZhiYuan Origin Model" has already completed this registration, establishing the basic conditions for compliant industry model service provision. The ZhiYuan Origin Model focuses on multi-spectral early risk warning, distinguishing itself from general generative large models. Its core capability lies in integrating multidimensional spectral perception data from ultraviolet, infrared, and visible light to achieve scene understanding and risk assessment. Backed by a long-accumulated risk sample database and industrial knowledge, it can identify latent risks such as arc flashes, overheating, and fire hazards difficult to perceive by the naked eye in IDC data centers and energy facilities. The model supports end-cloud collaboration, privatized deployment, and custom adaptations for various industrial environments. As a rare physical AI commercial entity listed on the Hong Kong stock market, HQVTs large model services have already become its first growth curve, with revenue from large model services reaching 320 million yuan in the first half of 2026, a staggering year-on-year increase of 382.5%, accounting for 78.8% of total revenue. With this proposed inclusion, eligible local enterprises in Shenzhen could purchase its multi-spectral AI services and benefit from policy subsidies, which will help reduce customer procurement costs, expand orders in more industrial scenarios, and accelerate the commercial penetration of vertical large models. It is worth mentioning that the recently released "2026 China Physical AI Industry Panorama Report" by the Leopard Research Institute included HQVT in the top 10 list of physical AI industry application practices. The evaluation criteria included the completeness of the technology stack, engineering delivery, and commercialization performance, validating the company's industry position in the multi-spectral physical perception arena. Rooted in Shenzhen, HQVT is recognized as a national-level specialized and innovative "Little Giant" enterprise. Previously, the ZhiYuan Origin Model was included in the first batch of large models in the AI industry in Guangdong Province. Its proposed inclusion in Shenzhens model service institution registry represents a dual recognition from local authorities of its multi-spectral AI technological capabilities and its ability to commercialize those capabilities. In the future, HQVT will continue to rely on the policy ecosystem to iteratively develop its end-cloud collaborative multi-spectral AI solutions, aiding the safe and intelligent upgrades of local manufacturing, data centers, and energy enterprises, while also exporting replicable risk warning solutions to other industries.