Domestic AI Industry Welcomes Turning Point of Accelerated Commercialization; Large Model Enterprises Pursuing "Value Realization" Chased by Capital Market

Recently, China's AI industry has been experiencing a comprehensive resonance extending from macro policies to micro financial data. Weighty documents from the State Council made it clear that artificial intelligence legislation is accelerating, the National Development and Reform Commission (NDRC) is planning to issue supporting documents for "AI+", ByteDance has increased its AI capital expenditure, and both Zhipu and MiniMax have been included in the Hang Seng Tech Index.

This intensive series of events releases a clear signal that the domestic AI industry has moved past the catch-up period of technical capabilities and has officially stepped into an accelerated interval of large-scale commercialization and implementation—marking the arrival of the commercial application era.

Macro Policies Release Signals, Pressing the Accelerate Button for AI Industrialization

On May 22, the NDRC released the strongest policy signal at its monthly press conference, stating that it is currently planning to introduce supporting documents to accelerate the implementation of "Artificial Intelligence+". This initiative will further increase resource guarantees, continuously promote central and state-owned enterprises to open up high-value application scenarios, build artificial intelligence benchmark applications, and guide the integration of AI into production, operations, management, and other key stages.

Earlier on May 8, the Cyberspace Administration of China (CAC), the NDRC, and the Ministry of Industry and Information Technology (MIIT) jointly issued the nation's first specific policy for intelligent agents, titled the "Implementation Opinions on the Standardized Application and Innovative Development of Intelligent Agents," covering 19 application scenarios including scientific research and industrial development. From the specific policy on intelligent agents to the "AI+" implementation supporting documents, policies are advancing layer by layer. The AI industry has fully upgraded from point-like technical exploration into a national main line at the strategic economic development level.

These favorable policies rapidly received positive feedback in the secondary market. Following the NDRC press conference on May 22, Hong Kong-listed AI concept stocks shot up in a straight line. Zhipu (2513.HK) once surged over 30%, hitting another historical high, while MINIMAX (0100.HK), Mininglamp Technology (2718.HK), and Xunce (3317.HK) all rallied strongly.

In the capital market, investors have clearly demonstrated confidence in the opening of high-value scenarios within the policy perimeter, as well as the "proving grounds" provided by large model companies to batch-verify commercialization logic.

Kaiyuan Securities stated that AI innovation continues unabated across models, computing power, and applications. On the model front: model innovation has entered a stage of orderly and high-quality development, AI model capabilities are consistently improving, and commercialization will be a critical theme for large model companies in 2026. On the computing power front: global CSP capital expenditure growth is robust, and the AI chip businesses of NVIDIA (NVDA) and Broadcom (AVGO) continue to see high growth, proving that demand for AI computing power remains strong. On the application front: demand for VibeCoding and Agents is booming.

Large Model Enterprises Intensively Deliver Commercial Results

Recently, due to their outstanding performance in the capital market, large model companies have been elevated to a higher capital stage.

After the market close on May 22, the results of the Hang Seng Index Advisory Committee's quarterly review were officially announced. Both Zhipu and MiniMax were included in the Hang Seng Tech Index, with the constituent changes officially taking effect on June 8.

Bloomberg Intelligence estimates that Zhipu could attract HKD 51 billion to HKD 92 billion in Southbound capital inflows, while MiniMax could attract up to HKD 47 billion. On May 12, China Merchants Securities initiated coverage on both companies, issuing "Overweight" ratings with target prices of HKD 1,282 and HKD 982, respectively.

On the same day the Hang Seng Index quarterly review results were officially announced, Zhipu launched its GLM-5.1 High-Speed version API for enterprise clients, achieving an output speed of 400 tokens/s and breaking the global speed ceiling for APIs among large model vendors. Its high-frequency, high-quality technical iterations are directly converting model capabilities into commercial revenue—the ARR (Annual Recurring Revenue) of its API platform has soared to USD 250 million, representing a 6.4-fold increase year-to-date.

Among the recent performances of large model-related companies, SenseTime-W (0020.HK) has built its presence on the trinity synergistic development of "SenseCore AI Infrastructure—Large Models—Applications." SenseTime's 2025 financial results showed a total revenue of RMB 5.01 billion, a year-on-year increase of 32.9%. Revenue from its generative AI business reached RMB 3.6 billion, up 51% year-on-year, with its proportion of the group's total revenue rising to 72%.

Guotai Haitong Securities pointed out that SenseTime's recently open-sourced SenseNova U1 Lite native multimodal model achieved the same performance as closed-source models using only 8B parameters, marking the official opening of the commercialization window for domestic AI. Generative AI is no longer a track exclusive to tech giants; incremental growth can also be found at the application level, such as SenseTime’s "Raccoon Family" and Vimi short-drama tools.

As a key large model player, Mininglamp Technology made frequent moves in May. At the beginning of the month, it announced a "dual open-source" strategy, simultaneously opening two foundational layers of its edge Agent ecosystem. Its strategic logic marks an evolution in the deployment of Agentic AI from "calling large models in the cloud" toward "edge-cloud collaboration," where Cider and Mano-P are designed to serve as the underlying "infrastructure utilities" for edge-side inference and operations. Additionally, the company released its first consumer-grade AI Native hardware, Octic—an AI recording hardware device in the form of a magnetic card—in the middle of the month. From releasing Mano-P in April, to launching the Octo platform, and finally delivering hardware, Mininglamp's "Model→Platform→Hardware" product matrix officially completed its closed loop in May.

TF Securities initiated coverage on Mininglamp Technology on May 20 with a "Buy" rating, stating that the company turned its Non-GAAP net profit positive for the first time in 2025. It expects revenue to grow from approximately RMB 1.426 billion to RMB 3.481 billion by 2028, with a compound annual growth rate exceeding 35%. In May, Mininglamp Technology delivered a strong performance in the capital market, with its cumulative gain exceeding 20% as of the close on May 22.

Xunce Technology is currently securing its position as a vertical token data hub, validating the second layer of space for token commercialization. Deutsche Bank significantly raised its target price for the company from HKD 85 to HKD 351. The core logic is that the company's token invocation ARR exploded by 300% quarter-on-quarter in April, with its scenario token pricing reaching USD 10 to USD 100 per million tokens, over ten times higher than Anthropic. Guotai Haitong maintained an "Overweight" rating, believing that enterprise-level AI data governance demand is experiencing explosive growth, and token-based payments will drive subsequent earnings growth.

China Merchants Securities judged in its research report that the Agent commercialization flywheel in the Chinese large model industry is unstoppable. China’s daily token consumption skyrocketed from 100 billion at the beginning of 2024 to 140 trillion in March 2026, growing more than a thousand-fold within two years. The core driving force is the surge in invocation volume brought by Agent applications. In the view of CITIC Securities, accelerated model commercialization is the strongest catalyst for the current AI industry. It recommends focusing closely on the dual drive on the application side generated by the iteration of AI model capabilities and the enhancement of token consumption.

Meanwhile, J.P. Morgan pointed out in a May survey report that China's AI "monetization moment" has arrived, and the value of the workflow layer is beginning to surpass that of the foundation model itself.

Hong Kong-listed Large Model Concept Stocks:

KNOWLEDGE ATLAS (02513.HK)

Knowledge Atlas Technology Joint Stock Company Limited (02513.HK; commercially known as Zhipu AI) is a leading player in China's independent large language model (LLM) sector (recognized for its capabilities and market position by Frost & Sullivan). The company has released its next-generation flagship model, GLM-5, achieving open-source State-of-the-Art (SOTA) performance in coding and agentic capabilities. It has also open-sourced the multimodal image generation model GLM-Image in collaboration with Huawei. Focusing on novel model architecture design, generalized reinforcement learning paradigms, and autonomous model evolution, its business layout closely aligns with the trend of enterprise-level AI productivity transformation.

MINIMAX-W (00100.HK)

A top-tier player dedicated to omni-modal Artificial General Intelligence (AGI). The company continuously iterates its video generation model matrix and has partnered with mainstream domestic and overseas chip manufacturers and inference platforms to complete underlying adaptations. The total monthly active users (MAU) of its B2C and B2B LLM applications, along with multimodal API call volumes, maintain rapid growth, further broadening its computing power ecological layout.

MININGLAMP-W (02718.HK)

As an LLM concept stock in the Hong Kong market, Mininglamp Technology is dubbed the "First Agentic AI Stock in Hong Kong." It has released DeepMiner, a multi-agent collaborative data analysis platform for commercial data analytics, and the open-source edge GUI (Graphical User Interface) model series, Mano. Deeply engaged in the enterprise service and data intelligence tracks, the company adopts a multi-agent architecture focusing on commercial data analysis. By utilizing structured factual data to mitigate AI hallucination risks, it provides decision-making agent pipelines for the marketing and mass consumption sectors. Additionally, it has launched AI Native hardware to expand the carriers for its agents.

SENSETIME-W (00020.HK)

Accelerating its transition towards Model-as-a-Service (MaaS) and generative AI services. The company has released and open-sourced the SenseNova U1 series, a native unified understanding and generation model, and the lightweight multimodal model SenseNova 6.7 Flash-Lite. Leveraging the computing reserves of its AI infrastructure (AIDC), it translates foundational computing power into omni-modal generation pipelines tailored for vertical industries such as automotive and healthcare.

Risk Warning: The above content is a compilation of recent public research reports and market information and does not constitute any investment advice. The industrial implementation of large models is affected by multiple factors such as technological evolution, market competition, and policy changes, and contains uncertainties. Investment involves risks, and caution should be exercised when entering the market.

NewTimeSpace Disclaimer: All content herein is the original work of NewTimeSpace. Any reproduction, reprinting, or use of this content in any other manner must clearly indicate the source as "NewTimeSpace". NewTimeSpace and its authorized third-party information providers strive to ensure the accuracy and reliability of the data, but do not guarantee the absolute correctness thereof. This content is for reference only and does not constitute any investment advice. All transaction risks shall be borne by the user.