31 AI companies listed on the STAR Market participated in the event collectively: The commercialization of AI has entered a crucial period for revenue realization.
Multiple companies in the artificial intelligence industry stated at their performance briefings that 2026 is regarded as the first year for the large - scale implementation of AI agents. AI security, embodied intelligence, and edge - side industrial vision have become the key areas for corporate layout.
The special session on the artificial intelligence industry of the collective performance briefing on the "15th Five - Year Plan" theme of the Science and Technology Innovation Board was held today (May 20th).
The companies participating in this event include 31 Shanghai - listed companies such as Tianzhun Technology, Bohui Technology, Starring Technology, Danghong Technology, Baoland, Longruan Technology, Xingtu Xinke, Kingsoft Office, and ArcSoft Technology.
Overall, the operating inflection point of the AI industry is gradually emerging. Multiple participating enterprises stated that 2026 is regarded as the first year for the large - scale implementation of AI agents. AI security, embodied intelligence, and edge - side industrial vision have become the key areas for corporate layout, and the business model is accelerating the transformation from one - time software and hardware delivery to subscription - based operation.
The AI industry is moving from "single - point pilot" to "full - scale penetration"
Since 2025, the overall operating quality of the AI industry has been continuously improving, and many enterprises have presented outstanding performance reports. In the first quarter of 2026, Qi'anxin achieved the best single - quarter performance in Q1 since its listing in multiple indicators; Hongquan Technology achieved a non - recurring net profit of 45.9139 million yuan in 2025, turning losses into profits; In the first quarter, CloudWalk Technology narrowed its losses and improved its operating cash flow. In 2025, its artificial intelligence solutions achieved a revenue of 415 million yuan, accounting for 82.95% of the total revenue.
The participating enterprises generally believe that the AI industry is moving from "single - point pilot" to "full - scale penetration", and many companies have made progress in AI commercialization.
Among them, Danghong Technology, centered on the BlackEye 2.0 multimodal large - scale model, has achieved full - link integration of the large - scale model with robot bodies, multi - sensors, and motion control modules, creating a full - stack multimodal solution for embodied intelligence scenarios. Tan Ya, the director, secretary of the board of directors, and deputy general manager of the company, said that the company has completed the intelligent upgrade and commercial implementation of relevant products, and will focus on expanding relevant businesses of embodied intelligent robots this year.
Zou Tao, the chairman of Kingsoft Office, said that the company's private deployment solution at the software level has mature delivery capabilities, which can support customers to fully run WPS AI - related functions in their own computing power environment. Currently, it has accumulated a certain number of benchmark customers in key industries such as government affairs, finance, and high - end manufacturing.
Edge - side AI is driving industrial vision to evolve from the traditional "image transmission" model to the local "perception - decision - making" closed - loop. In scenarios such as industrial quality inspection and production line guidance, it can reduce latency and simplify the system architecture, and has been put into practical application. Huang Yuanhao, the chairman and general manager of OPPO Megvii Technology Group Co., Ltd., said that the industrial - grade vision sensors such as Femto MegaI launched by the company are widely applicable to scenarios such as warehousing and logistics, loading and unloading on production lines, and industrial inspection.
While the industry as a whole is improving, the participating enterprises also generally mentioned the common challenges they are currently facing. Among them, supply - chain fluctuations, especially the rising prices of electronic materials, have become the main factors affecting the short - term profitability of enterprises.
In the first quarter of 2026, Hongquan Technology achieved an operating revenue of 179 million yuan, a year - on - year increase of 7.51%; however, its net profit attributable to the parent company was - 7.6695 million yuan. He Junqiang, the chairman of the company, said that the main reason for the loss was the significant price increase of electronic materials such as memory chips, which led to a decline in gross profit margin. Currently, the company has taken relevant countermeasures and hopes that the situation will gradually improve in subsequent quarters.
In addition to supply - chain issues, challenges in the commercial implementation of AI still exist. Zou Tao, the chairman of Kingsoft Office, said that the company's AI commercialization is at a critical stage of transitioning from "user - scale accumulation" to "gradual revenue realization", but there are still two major challenges in the current implementation: First, it takes a certain period to cultivate users' habits of using AI; Second, the work of transforming the massive unstructured data of enterprise customers into knowledge assets is relatively complex.
Chen Yanxing, the chairman and general manager of Luopute, also said that currently, for the AI office projects of government and enterprise customers, the transformation cycle from pilot to large - scale implementation varies greatly among customers in different industries and of different scales, and the cycle depends on the customers. Regarding the inflection point of profit contribution from AI investment, the company is currently in a stage where continuous AI investment and deepening of the commercialization model go hand in hand. The overall logic of the company's AI investment is to focus on the long - term while considering efficiency, and it will maintain a reasonable balance between capacity building and operating quality.
The business model transformation is accelerating, and AI security has become a strategic high - ground
For a long time, domestic AI enterprises have generally adopted a project - based business model, that is, custom - developing solutions according to customer needs and charging a one - time fee. At this performance briefing, many enterprises said that they will accelerate the transformation to productization and subscription - based models to create a sustainable revenue model.
It is understood that CloudWalk Technology is promoting the transformation to the "subscription + operation" model and deepening cooperation with Huawei Ascend in terms of underlying architecture adaptation, joint R & D, inference optimization, and industry scenario implementation, continuously improving the capabilities of domestic AI infrastructure and industry solutions.
Zhou Xi, the chairman and general manager of CloudWalk Technology, said that the company will actively promote the "second - curve" business. Based on the explosion of new technologies and new business forms, it will upgrade the company's intelligent base and agent lab, create a business model in the AI Agent era, promote the upgrade of the business model to subscription and operation services, and gradually shift from one - time software and hardware delivery to continuous operation services based on Tokens and agent applications, improving revenue quality and sustainability.
Chen Yanxing, the chairman of Luopute, said that in 2026, the core growth drivers of the company are the transformation of business productization and the in - depth implementation in multiple industry scenarios. Relying on its own AI technology platform, the company will vigorously promote the R & D and implementation of standardized products such as intelligent computing all - in - one machines and industry robots, accelerating the transformation from the project solution model to product - based operation.
Zou Tao, the chairman of Kingsoft Office, said that Kingsoft Office's AI commercialization is at a critical stage of transitioning from "user - scale accumulation" to "gradual revenue realization". The personal business shows a clear trend of increasing both volume and price driven by WPS AI, and the B - end WPS 365 maintains a high growth rate.
In April 2026, the AI model Claude Mythos released by Anthropic marked that cyberattacks have officially entered the "industrialized" production stage from the "manual" era. The limitations of the traditional rule - based protection model are prominent, and it is urgent to build a multi - layer defense system.
It is reported that at the beginning of this year, Qi'anxin specifically established an artificial intelligence subsidiary, bringing together three teams: one is the scientist model team; the second team consists of attack - defense experts for actual combat; the third team is composed of product engineers for actual measurement and implementation. The new company focuses on three major directions: first, large - scale models and agents in the vertical field of network security; second, security intelligence (AI for Security); third, intelligent security (Security for AI). Qi Xiangdong, the chairman of Qi'anxin, defined 2026 as the "singularity" year of AI security. He said that security has jumped from the previous "behind - the - scenes guarantee" to a "must - have foundation" that must be pre - set for the large - scale application of AI, and AI security has also become a strategic high - ground that "all forces must compete for".
Optimistic about the long - term development trend of AI, many enterprises have clear growth strategies
Although the current AI industry still faces many challenges, the participating enterprises are generally confident in the long - term development of the AI industry and have clearly defined their future development directions. Xu Yihua, the chairman of Tianzhun Technology, said that with the rapid development of AI, the entire industrial chain will inevitably have a large number of new demands for industrial equipment, especially in the fields of semiconductor chips, computing power infrastructure, intelligent terminals, and robots that Tianzhun focuses on. The company is full of confidence in its future development. Tianzhun drives the AI industrial chain with AI equipment and has widely implemented its solutions in the field of visual measurement and inspection, achieving a closed - loop of commercial value.
Zhu Guosheng, the chairman and general manager of Bangyan Technology, believes that 2026 is indeed generally regarded as the key first year for the large - scale implementation of AI agents in the industry, and relevant policies have also drawn a clear roadmap for the high - quality development of the industry. Currently, the market has formed a "dual - track competition" pattern where general platforms led by large - scale technology companies coexist with vertical manufacturers deeply cultivating niche scenarios. The company will subsequently focus on expanding the cultural and tourism, research and study directions through the IP operation model, and plans to expand to six major scenarios such as education and training in the medium term.
Ji Xinhua, the chairman of UCloud, said that from an industry perspective, cloud computing is an important infrastructure and computing power base for AI, and the overall industry prosperity is expected to continue to improve. Around AI, the company continues to explore and deeply cultivate six major fields, namely: basic models, industry models, AI agents, AI applications, AI hardware, and AI robots. The product services provided to customers include GPU computing power and its supporting CPU, storage, network and other resources, or high - power cabinet hosting services related to AI and their supporting services, or tools such as the intelligent computing scheduling platform and the parallel file storage product UPFS.
Liu Zhihong, the chairman of ProPlus Design Solutions, said that the company will closely follow the development trend of AI and layout the linkage of relevant new technologies such as large - scale AI models in the R & D of EDA products. Li Kai, the chairman of Borui Data, said that the company will seize the development opportunities of AI, improve the overall production efficiency through AI Coding, and achieve a steady improvement in operating quality and sustainable growth in benefits in the mature business fields.
Guo Zhongwu, the director and general manager of Bohui Technology, said that facing the "15th Five - Year Plan" period (2026 - 2030), the company will focus on the strategic main line of "AI + audio - visual", take media security, smart education, and intelligent display and control as the main pillars of its business, and take AI security as a new strategic growth pole to build a diversified growth pattern of "in - depth development of the main business + AI security + exploration of new tracks".
This article is from the WeChat official account "Science and Technology Innovation Board Daily", author: Chen Junqing, Dai Jiayi. Republished by 36Kr with authorization.