The first year of large-scale implementation of AI applications, AINNOVATION creates a new model for stable and sustainable development

In 2025, favorable policies, technological breakthroughs, and the industry's demand for AI are driving the AI + industrial internet to become a trillion-level opportunity.

On the policy front, the government work report in 2025 first proposed the "Artificial Intelligence +" initiative, and "large models" were included in the national top-level design for the first time. This is also the eighth consecutive year that the government work report has mentioned "industrial internet," accelerating the digital transformation of manufacturing and supporting the widespread application of large models.

So far, multiple provinces and cities have launched implementation plans and subsidy policies for the intelligent transformation of industries. For example, Jiangsu Province invests 1.2 billion annually in special funds for the "smart transformation and digital upgrade" of industrial enterprises, and Zhongshan City in Guangdong provides incentives of up to 5 million for enterprises recognized as digital intelligent demonstration factories. These favorable policies are expected to promote enterprises to accelerate their intelligent transformation.

On the technological front, by 2025, several tech giants and top research institutions will have simultaneously launched new generations of large models, covering general intelligence, vertical fields, multimodal interaction, and ethical safety. The millisecond-level latency of 5G makes cloud-coordinated AI possible, while the popularity of edge computing and the decrease in computing costs will also catalyze the landing of AI applications in multiple fields.

In this context, 2025 will be a key year for AI and industrial large models, marking the transition of AI technology from the focused research and development stage of "All in AI" to the widespread application stage of "AI in All."

As the intelligent transformation of the industrial sector deepens and the explosive growth of industrial large model applications occurs, AINNOVATION, as a rapidly developing enterprise-level AI solution provider in China and a leading "AI + manufacturing" solution provider, will also enter a period of dividend realization, ushering in a "Davis Double Play" in performance and valuation.

Establishment of the "One Model, One Body, Two Wings" Strategy: Comprehensive Empowerment of Industrial Manufacturing through Software and Hardware

As a rare tech company focused on the "AI + manufacturing" field, AINNOVATION has currently formed a unique business model and framework.

In simple terms, the company has established a technical strategy of "One Model, One Body, Two Wings," which is based on its self-developed AInnoGC industrial large model and AI intelligent agent development platform, creating a dual-core business system that integrates both software and hardware aspects of "industrial robots + industrial software."

As the core technical support of the company, the AInnoGC industrial large model achieved multiple performance breakthroughs in 2024, significantly improving computational efficiency, model accuracy, and stability.

According to Zhitong Finance APP, the new version of the AInno-75B large model has a parameter scale of 75 billion, which is a fivefold increase compared to the previous AInno-15B, and it has added multimodal processing capabilities, supporting input of text, images, videos, EEG (electroencephalogram), and specific modalities in industrial scenarios.

In terms of capability indicators, AINNOVATION's AInnoGC industrial large model has achieved the highest level of 4+ in the trusted AI industrial evaluation by the China Academy of Information and Communications Technology; it is listed among the 15 "Advanced Application Cases of China's AI Large Models" published by MIT Technology Review in December 2023; and in the first Chinese-native industrial large model evaluation benchmark SC-Industry released by SuperCLUEC, AInno-15B ranks just behind GPT-4 Turbo and GPT-4, topping the list in the domestic industrial large model field In addition, the AInno -75B industrial large model has passed the dual filing of generative artificial intelligence services and domestic deep synthesis service algorithms by the National Internet Information Office. It has also received the recognition of "General Artificial Intelligence and Industrial Integration AI Large Model Potential Service Provider" from the China Industrial Internet Research Institute, indicating that the company's large model has a solid technical foundation and application value, and will accelerate the advancement of commercialization practices in the future.

The "AI agent" is a mini-program that can accept natural language commands, interact with scenarios, and has a preliminary thinking chain. If the basic large model is the "brain" composed of a vast knowledge base, then the AI agent is the interactive layer between the brain and the real world, capable of fully utilizing the knowledge reserves of the large model through perception, retrieval, analysis, reasoning, planning, decision-making, and execution steps, collaborating and accompanying like human employees.

Based on the AInnoGC industrial large model technology platform, AINNOVATION has built an AI Agent development platform aimed at the industrial sector, achieving a significant shift from passive response to active execution in AI. It is understood that AINNOVATION's AI agent platform supports visual drag-and-drop rapid construction of generative AI applications, shortening the development cycle by more than 60%, significantly reducing the development costs of AI applications, and widely covering multiple industrial scenarios such as equipment operation and maintenance, process optimization, and quality inspection.

At the 2025 ICT Industry Trend Annual Conference, Academician Wu Hequan of the Chinese Academy of Engineering stated that the upcoming year 2025 will be the inaugural year for AI agents and AI terminals; OpenAI CEO Sam Altman predicted that the first batch of AI agents will "enter" the labor market in 2025, effectively changing the output model of companies.

Built on the solid foundation of the AInnoGC industrial large model and the AI Agent development platform, AINNOVATION actively explores more AI commercialization application examples, continuously empowering the digital transformation of the manufacturing industry in both software and hardware aspects, promoting the "two wings" of industrial robots and industrial software to soar.

In 2024, AINNOVATION will increase its research and development of industrial embodied intelligent robots, constructing a fully autonomous and controllable complete industrial embodied intelligent technology stack. Embodied intelligence refers to integrating artificial intelligence into physical entities such as robots, endowing them with the ability to perceive, learn from, and dynamically interact with the environment. Embodied intelligent robots are an important carrier for the implementation of large models.

It is understood that the upgraded ChatRobot from AINNOVATION adopts a multimodal, end-to-end VLA (Visual-Language-Action) strategy model, possessing a closed-loop capability of "perception-understanding-planning-decision-making," significantly enhancing the task generalization and interaction friendliness of robot operations.

In terms of AI industrial software, the company previously launched the generative-assisted industrial design application ChatCAD, which supports generating mechanical and electrical designs through language interaction. Complex industrial design models can be obtained through simple natural language input; currently, the company is further exploring the implementation path of image input to CAD technology (Image-to-CAD), which can automatically generate CAD models from static images, and the generated models are compatible with various mainstream CAD formats Considering that modern industrial design and manufacturing are closely related to CAD technology, this application will have a huge potential market and is expected to further enhance the company's brand awareness and market influence.

Further Focus on High-Margin Core Business, Significant Improvement in Financial Health

While deepening at the strategic level, the fundamentals of AINNOVATION are also impressive. Over the past 24 years of the company's financial reports, not only has revenue maintained steady growth, but several financial indicators have also shown significant improvement, with operational quality steadily increasing.

Specifically, the group's annual revenue reached 1.222 billion, with adjusted net losses narrowing by 23.9% year-on-year to 117 million; due to further focusing on high-margin, high-quality businesses, the gross margin increased by 1 percentage point to 34.6%; for the first time, the proportion of manufacturing revenue accounted for over 80% of the company's total revenue, continuously consolidating the differentiated market advantage of "AI + manufacturing."

Thanks to the company's important adjustments in operational philosophy, prioritizing operating cash flow, the net cash flow from operating activities significantly improved by 80.1% compared to the same period last year, with cash and cash equivalents at 1.205 billion at the end of the period; the aging structure has been significantly optimized, with trade receivables and notes net value decreasing by 28.8% year-on-year, and the proportion of accounts aged 3 months and above being less than 50%.

As of now, the company's financial health is at its best since its establishment, demonstrating the effectiveness of its refined management strategy.

By the end of December 2024, the company has cumulatively applied for 1,365 patents related to AI and industrial large models, including 1,118 invention patents; confirmed patents total 605, with 383 being invention patents. Strong R&D capabilities have also driven the total number of customers from 397 in the same period last year to 521.

Through continuous optimization and upgrading of business structure and resource allocation, AINNOVATION has enhanced its ability to cope with market uncertainty risks and provided ample support for ongoing investment in R&D innovation with abundant cash flow.

Focus on Vertical Fields, Accelerating Commercialization of "AI + Manufacturing"

Looking ahead, AINNOVATION is continuously expanding the application boundaries of AI in various aspects of the manufacturing industry, with dense commercial applications expected to accelerate the release of technological dividends and contribute more incremental performance.

By empowering industrial software with AI agents, AINNOVATION aims to address traditional pain points in industrial manufacturing in 2024, launching multiple applications in segmented fields, such as intelligent sales knowledge assistants for marketing, smart office assistants for management operations, intelligent formula process analysis and optimization for product R&D, and intelligent equipment operation and maintenance assistants for production manufacturing, fully empowering the industrial manufacturing supply chain;

In the field of industrial robotics, the company is actively exploring the application of logistics scenarios, significantly reducing operational blockage time and improving the response speed and accuracy of scheduling systems by introducing large model technology to optimize the operation and maintenance efficiency of on-site logistics robots.

For example, the company's subsidiary, SaiDiQiZhi, has independently developed the "iSmartBGSU Bridge Grab Unloading Machine Intelligent Operation System," which integrates high-precision positioning, intelligent perception, digital twin, deep learning, and other technologies to achieve intelligent operation of bridge grab unloading machines; For example, the iSmartLoco unmanned locomotive driving system developed by the company has been successfully deployed at a raw material terminal in southern China. Its core high-precision control system can intelligently match the locomotive's load (80-1500 tons), operating scenarios (indoor/outdoor), and weather conditions, maintaining positioning accuracy at the centimeter level (5-10 cm) even when towing loads exceeding 1,000 tons, creatively achieving a "fully automated driving closed loop."

In the company's traditional stronghold of industrial vision, it has made further technological breakthroughs by cleverly integrating AInnoGC large model technology with traditional MMOC small model technology. This has expanded application scenarios from a single quality inspection scene and video security to a full process that includes model production, model management, model performance assessment, quality inspection application construction, and intelligent analysis of quality inspection results, providing industry-leading clients with highly integrated and high-performance intelligent quality inspection software and hardware solutions.

Entering 2025, the AInnoGC industrial large model will undergo another technological upgrade, completing the adaptation of DeepSeek-R1 distilled AInno-75B technology. After this upgrade, the reasoning capability of AInno-75B has significantly improved while retaining its advantages in deep understanding of industrial knowledge and low-cost private deployment, enabling deployment applications in a wider range of industrial environments.

Summary

ZheShang Securities pointed out that the application of industrial large models in China is expanding from single links such as research and development, production, management, and services to multi-scenario coverage in vertical industries. In the future, it is expected to form product forms such as knowledge intelligence, business intelligence, embodied intelligence, and system intelligence, reshaping all elements of production and manufacturing, the entire industry chain, and the entire value chain, promoting the manufacturing industry to enter a new stage of digitalization, networking, and intelligence.

After years of continuous deep cultivation in AI + manufacturing, AINNOVATION has developed a wide range of application scenarios in the field of vertical large models. The application of its AInnoGC industrial large model has covered many fields, including automotive equipment, panel semiconductors, steel metallurgy, energy power, food and beverage & new materials, and intelligent manufacturing training, with sufficient momentum for continuous growth.

Driven by favorable policies and high market prosperity, the AI wave is pushing AINNOVATION into a new stage of high-quality development. Currently, the company is significantly undervalued in the secondary market, highlighting its long-term allocation value