Generative AI ushered in a singular moment or formed two trillion markets.
Our reporter Qin Xiao reports from Beijing.
Generative AI technology, represented by big model, has become the most important technological singularity to detonate digital nativity. Gartner predicts that by 2026, more than 80% enterprises will use API or model of generative AI, or deploy applications supporting generative AI in production environment, which will also bring great opportunities and challenges to industrial development.
The traditional technical system, organizational structure, business model and industrial form have been reconstructed rapidly, and the integration of data and reality has entered a deeper and broader new stage. Guo Wei, chairman and CEO of Digital China, told China Business News that this is a great opportunity and a challenge for us. Grasping it is an opportunity for development; If we can’t catch it, we will miss a major development period. In the next ten years, all enterprises will make full use of three primitives (cloud primitive, digital primitive and AI primitive) to subvert their business, construct their second and third growth curves, rewrite their business, and realize the leap-forward growth of enterprises in the digital age.
Two trillion-dollar markets
Since the beginning of this year, general artificial intelligence, represented by cognitive big model, has triggered a craze around the world. OpenAI, Microsoft, Google and other international companies are constantly overweight, and there is a "thousand-model war" in China, and many high-tech companies are competing to invest in research and development. According to the "Research Report on China Artificial Intelligence Large Model Map" published by the Ministry of Science and Technology, among the published large models in the world, China and the United States are far ahead, accounting for more than 80% of the global total.
At the same time, the big model is also regarded as a new engine to promote enterprise transformation. A few days ago, the report "The Economic Potential of Generative Artificial Intelligence: The Next Wave of Productivity" released by McKinsey shows that if the 63 kinds of generative AI analyzed are applied to all walks of life, it will bring 2.6 trillion to 4.4 trillion dollars to the global economy every year. Growth. This forecast has not counted all the applications of generative AI. If the applications that have not been studied are counted, the economic impact of generative AI may double.
Guo Wei said that this is a great change in human economic life and social life. The rise of today’s big model is actually to transform the small business model and digital module that were established manually in the past into the Transformer model through training. In the case that natural language is universally understood, the automatic generation process of data or knowledge is realized through generative AI, which accelerates the flywheel of data generation. In the past, new knowledge and data were generated by system data and alternative data processing, but today we automatically generate new data and knowledge through large models, making the generation of data assets become a perpetual motion machine driven by AI.
Specifically, all parties in the industry generally believe that the demand-side generative AI market is about to explode, and it may even form two trillion-dollar markets, C-end and B-end, in the future. At present, the demand of enterprises for generative AI is growing rapidly, and the general model will inevitably change the traditional AI market, and the improvement of productivity will also greatly promote industrial development.
Li Hui, technical director of 360 Intelligent Products Department, said that the big model market is an unprecedented and revolutionary market, which will be ten times the increment of the current Internet and will definitely reach the trillion level. Because with the development of 5G, the ecological content of the mobile Internet has exploded by leaps and bounds. In the field of AI generation, it will contribute to the outbreak, which will inevitably bring more explosive demand growth points to the market. At the same time, it will bring new changes to the traditional AI market, and it will be an alternative and updated process. With the further improvement of computing power in 2026, especially the further upgrading of domestic computing power, the influence and market scale of generative AI will be further expanded.
Zhang Lei, general manager of Chery Automobile Jinke Innovation Center, believes that it is possible to reach two trillion-dollar markets: first, To C-end, consumer-grade, and now it is in an explosive period. The second is the To B field. These two different fields are completely different in application scenarios, customer needs and customer industrial chain forms. It is predicted that these two fields will reach trillion-dollar markets in the future.
How to get to the "last mile"
Although the trillion-dollar market in the future is very attractive, in the process of the concrete landing of the large model, most manufacturers indicated that besides the technical landing perspective, the customer scenario needs are also the core of current concern, but in reality, there are still problems such as data acquisition and data governance. Although the Q&A interactive scene with rich corpus landed quickly, there are still great challenges in the actual enterprise-level commercial application, calling for ecological breakthrough. Or in this era of "Hundred Mode Wars" or even "Thousand Mode Wars", how to transform technology into real productive forces, break the game with the power of ecology, and accelerate the landing of generative AI at the enterprise level is the key.
Shi Jianhua, chief digital officer of Tasly Holding Group, admits that users don’t need big models, what they need is to solve problems. Now, some have made big models, some claim to be big models, and some have not. In the end, only a few giants will win, and there cannot be too many big model manufacturers. Everyone will think that AI is the future, the big model is the infrastructure, and what everyone has to do is to embrace and integrate. The most important thing for an enterprise is to find the matching user requirements, and to define them well is the key.
"Generative AI technology has a series of challenges to be solved in the enterprise scene. The last mile is not the last square kilometer. " Li Gang, vice president and CTO of Digital China, said, "In the closed-door communication with customers, customers said that some application developers said that they specialize in the last mile of large-scale model enterprises. Customers complain that it looks like the last kilometer, but it’s more than one kilometer to walk. Another customer said that you made a mistake with the company. He said that the distance between the big model and the enterprise application should not be measured in kilometers, but should be one square kilometer instead of one kilometer, that is, the landing of the big model enterprise needs more preparations. It is not a linear process, it is a comprehensive process. "
Li Gang emphasized that enterprises really need a relatively complete generative AI innovative application platform, which is capable of rapidly deploying computing power, preparing data, managing scenario applications and life cycle, and can also produce high-value and low-tech cost effects through continuous iteration, and can arrange and manage the knowledge, collaboration, architecture and roles related to the application of large models in enterprises, so as to truly fill a square kilometer between the so-called generative AI large model itself and enterprise scenario applications.
Based on this, Digital China released the platform of asking for learning in China, "White Paper on Application and Landing Technology of Generative AI Enterprises", and launched the "Ask for Learning and Co-create Plan".
It is worth noting that, unlike the AI big model, Shenzhou Wenxue is positioned as a one-stop enterprise-level big model integration platform, which helps enterprises to accelerate the innovation of generative AI, reduce the development threshold and landing cost of AI applications, and enable enterprises to combine their business needs with big model technology more quickly to realize the intelligence of business processes.
Li Gang introduced: "With the platform of learning in China as the core, we don’t make the basic big model, but make the integration and application development delivery platform of the big model, thus accelerating enterprise AI innovation; We become a service partner of big data, thus accelerating the upgrade of enterprise data governance; We do ecological ties, model markets, data marts and application stores, thus accelerating industrial innovation and ecological destruction. "
(Editor: Zhang Jingchao Proofreading: Yan Jingning)