The industry model focuses on specific fields, scenarios, and problems, and can combine the unique advantages of the enterprise to cultivate and empower the industry, thus forming differentiated advantages. Only by moving towards segmented fields can there be more opportunities. Taking into account factors such as industry specialization, continuous iteration, and comprehensive costs, industry models are more likely to achieve commercial value implementation. It is not an exaggeration to use the term 'Hundred Model Wars' to describe the current popularity of large artificial intelligence models. Data shows that there are over 80 publicly released large models in China. Do we need so many large models? How long will this craze last? Will it bring foam? We need to think more coldly about this. Domestic enterprises are entering the big model market one after another because they are optimistic about the commercial value behind the big model. Although ridiculed by many as "serious nonsense," ChatGPT has opened up new space for the industry's understanding and imagination of artificial intelligence after showcasing its intelligence in simulating human thinking in chat conversations, creative programming, and other aspects. Some companies view big models as the core of artificial intelligence, believing that they are likely disruptive innovative technologies leading the fourth industrial revolution, which will drive changes in various industries. Being able to seize such a super trend is not only a symbol of corporate strength, but also easier to gain the favor of capital. Because of this, in just a few months, the number of large artificial intelligence models in China has shown explosive growth. A large model is not necessarily more beneficial. The universal big model represented by ChatGPT has a high technical threshold, and the costs of development, training, and operation are enormous. The cost of training can easily reach millions of dollars, making it known as the "luxury game for a few people". Only large leading enterprises or leading enterprises have the corresponding R&D and investment capabilities. Other enterprises blindly follow the trend and rush forward, which not only makes it difficult to succeed, but also brings investment waste, increases energy, computing power and other consumption, and produces low-quality products. In short, launching multiple universal large models is neither realistic nor has as much market demand, and it will also bring about homogeneous competition. The industry is now paying more attention to industry large models, which have significantly lower thresholds than general large models and require relatively lower investment costs. However, it should also be noted that writing poetry and painting is not all about big models. Industry big models focus on specific fields, specific scenarios, and solve specific problems, and can combine the unique advantages of the enterprise to cultivate and empower the industry, thus forming differentiated advantages. Only by moving towards segmented fields can there be more opportunities. Taking into account factors such as industry specialization, continuous iteration, and comprehensive costs, industry models are more likely to achieve commercial value implementation. Currently, some industry models have played a role in various fields such as finance, manufacturing, pharmaceutical research and development, coal mining, and railways. Currently, the development of large models is still in its early stages and still faces issues such as technological shortcomings and privacy security. In the actual use of some large models, it has been found that the generated content has problems such as poor quality, redundant answers, and even risks outputting harmful content. To this end, seven departments, including the Cyberspace Administration of China, jointly announced the Interim Measures for the Management of Generative AI Services a few days ago, which has made certain requirements and specifications for generative AI products and service providers, but there is still a lack of