Source: Venture State
In 2024, Phase 25 of Chuangye State Star Camp ——New AI Star Financing Acceleration Plan Launches From the start of recruitment to now, we have received applications and interactions for 400+projects. We launched the project around the most concerned issues of AI entrepreneurs [Insight into AI · Sailing for the Future] Column.
In this issue, Qiu Yuefeng, the director general manager of Turing Venture Capital, the tutor of the 25th episode of Starcamp, and his post investment enterprise, the founder of Fengma Technology, and Nan Siqiao, the outstanding alumnus of the 23rd episode of Starcamp BC23, were invited to bring their insights and thoughts to everyone, hoping that everyone who is full of curiosity and dreams about AI, technology, business, and venture capital can gain something.
Look for demand islands with common customers and smaller boundaries
Mr. Qiu Yuefeng focuses on technology driven investment direction, has more than 12 years of experience in Internet products, business analysis and early venture investment, has a deep understanding of industrial Internet and cutting-edge technology, has led the investment in dozens of science and technology projects in the early venture investment field, and has successfully listed projects.
Representative investment cases:
Wisdom Spectrum AI, Listen Intelligence, Visual Bit Robot, Star Movement Era, Huayi Quantum, Baker Micro, Oriental Space, Fengma Technology, Arctic Xiongxin, Wisdom Core Huaxi, Huayu Yuandian
1. In 2024, what were the questions you thought most about AI?
In the process of growing with dozens of AI start-ups over the years, our thinking and exploration on commercialization has never stopped. At the current stage of AI technology development, we expect that it has become more and more pragmatic, especially under the ToB scenario, so we will see that the transformation of sales leads is becoming more and more difficult, and customer needs are becoming more and more alike and systematic from the experience function point. This not only puts forward higher requirements for the effect breakthrough of new technologies, but also brings more engineering cost and efficiency problems. At present, the operation mode and cost structure of most AI enterprises cannot be sustained for a long time. It is particularly important to output appropriate commercial service capability based on the existing technical effect and acceptable engineering cost structure. This is not only the problem of customer expectation management, but also the ability to define and design products based on the effect of existing technology, that is, the vast and complex demand pool of customers is separated from a demand island with certain commonalities and smaller boundaries, so as to reduce the difficulty of implementation and have the opportunity to replicate and promote. This will also become the third important competitive barrier element besides data and customer resources.
2. Looking back to 2023, which node or event impressed you deeply?
Meta、 Google and other global Internet giants are constantly pushing forward the iterative updating of the big model framework, which makes people feel more impressive. This not only keeps the enthusiasm and expectation of the public and the industry for AI technology, but also helps to attract resources; More importantly, it provides AI start-ups, especially those focusing on application side scenarios, with more clear references and more powerful tools for technology development direction, so that start-ups can think about their positioning in the whole ecosystem earlier, focus on the business scenario itself, and optimize their product and service capabilities.
3. Please share your interesting stories with the founders of 1-2 AI invested enterprises.
In 19 years, we incubated and invested in Fengma Technology, an agricultural science and technology startup company. We and the founder often jokingly said that we had never thought that science and technology entrepreneurship turned into a farming company. Different from the previous smart agriculture companies, Fengma technology team went deep into the front line to polish AI technology tools with actual planting production practices, and create an overall planting solution. Therefore, we do not simply sell software and hardware solutions for data collection, because this will not solve the fundamental problems of traditional agriculture. This also led to a relatively slow initial business exploration, but it also prompted Fengma Technology to have thousands of acres of science and technology greenhouses that can be produced and sold by itself today. The internal production tools with AI technology as the core have greatly improved the production efficiency, and have also completed productization and started to market. This is also the only way for AI technology to develop business in segmentation scenarios, It is possible to realize a real technology enabling industry by polishing products in the demand scenario and combining vertical domain knowledge with artificial intelligence technology.
4. Which AI products do you think will highlight great value or lose value in the future?
In the past few years when the big model has become familiar to everyone, the attention of AI middle platform (middleware) has also increased. In essence, the importance of end customers has become more prominent. People are reaching the direct downstream market, and the shorter the path, the greater the value. Under this logic, AI products that not only integrate the underlying technology of the big model, but also have data or customer resource advantages in the segmentation scenarios, and have industry understanding, will be more easily accepted and have greater value. This is also the reason why in addition to the large model base, we have also made a layout for AI Agent business, and even more subdivided industry demand scenarios such as planting, heavy industry production lines, aviation air traffic control, etc.
5. In the AI market, which business do you think will go faster on the C side or the B side?
The C-end and B-end demand explosion have different critical points for technology. The C-end requirements can accept more shallow technical effects, but the generalization of requirements is not easy to converge: the B-end requirements need to be deeply solved end-to-end problems, but the requirements are relatively focused. Therefore, in the relatively early stage of technology maturity, the B-end business will develop ahead of the C-end business, and the product engineering capability can make up for the defects of the new technology effect to a certain extent, but the overall speed is slow; When the new technology reaches a certain stage of maturity and the market homogenization competition intensifies, leading to price war, the C-end business scenario is likely to explode. Once it breaks out, the C end will quickly surpass the B end in development speed and scale. The AI market also follows this rule, and the progress of AI technology in different segmentation scenarios at the C end will also be very inconsistent.
6. Which elements do you think are most critical to the long-term survival of AI company?
The ability of the founding team to grasp market demand, product design and adapt to technological development is the key to the long-term survival of AI Company.
Agricultural AI model: interdisciplinary and unstructured empirical deconstruction
Nan Siqiao, founder of Fengma Technology and outstanding alumnus of Star Camp BC23
As a leader in the domestic smart agriculture industry, Fengma Technology has always focused on applying AI big data technology to crop planting and giving priority to facility agriculture, providing precision cultivation decisions, intelligent implementation plans and standardized planting systems for growers. Fengma Technology was selected as one of the 65 most socially influential start-ups in Fortune in 2023.
1. In 2024, what were the questions you thought most about AI?
We are an AI industry model, so we have always been pursuing how to better integrate with the industry; What we think most about in the agricultural field is how to combine botany, planting experience, etc. with AI technology to adapt to diversified needs and changing environment. Secondly, with the maturity of the big model, we are also trying to use the big model to improve efficiency in our R&D and production, and to complete more functions in product realization.
2. What new exploration does your company have in terms of products and commercialization?
We have been exploring and innovating. Our research on AI models in agriculture is to constantly deconstruct, integrate, train and iterate interdisciplinary and unstructured experience, which is actually a very complex system. But when it comes to products, we also need to think about how to use these complex technical capabilities more simply and directly, how to reduce economic costs and reduce the threshold of cognition. This year, we officially launched the "greenhouse butler" product on the planting side, which should be the first product with adaptive AI model in the field of facility agriculture planting in China, But farmers don't need to know how complicated the model is. They just need to know that this product can be used at a glance, and someone can help you decide how to plant, how to operate irrigation, fertilization and control the environment. It can save energy and worry to plant the land. It doesn't matter whether it is a planting expert. It can also increase production and reduce costs. The meaning of the name of Fengma Technology is "the password of Fengshou". What is the password of Fengshou? "Greenhouse Housekeeper" is the answer we gave to farmers in the product end this year. We think this is just a new starting point, and Fengma is still exploring on the AI agricultural road.
3. What type of AI products do you expect most from your own products?
If we put aside our own products, but they are related to our agricultural scenes, I have to mention the robot products that have always been concerned about, especially the embodied intelligence that has been discussed more than a year ago. In the agricultural scene, there is still a large part of labor that cannot be replaced, such as picking, I am very looking forward to the application of humanoid robots equipped with general artificial intelligence, which may really replace repetitive and simple manual labor in the agricultural field and liberate people from heavy manual labor.
4. What are the similarities and differences between the current AI entrepreneurial logic and the original Internet era?
In my opinion, technology entrepreneurship is essentially the same. It is the pursuit of how to improve efficiency in certain fields and scenarios, and how to make the results of this scenario have higher value. However, today's entrepreneurial teams have a clearer and more common understanding of AI entrepreneurial logic, which is also because we have experienced the era of Internet entrepreneurship and stood on the "shoulders of our predecessors". Our team has also attracted many talents who have ever started an Internet business. They have both succeeded and failed. But for me, this is the reason why Fengma has absorbed them. These will become the experience of Fengma's team. Let us have clear goals and avoid detours on the road of R&D and innovation.
About Venture State Star Camp
Founded in 2015, Chuangye Bang Star Camp is dedicated to providing acceleration plans for innovative companies. Through in-depth cooperation with top capital, industrial companies and industry experts, it helps entrepreneurs in the start-up - A round improve their underlying logic and help them become the top segment track.
The Star Camp has successfully held 24 sessions, Cumulative From 20000+ Selected from the registration items 1000+ Outstanding entrepreneurs have entered the incubation system and achieved excellent results in the past: the maximum amount of single student financing is up to 900 million The proportion of financing for roadshow projects reached 64.3% 。
Phase 25 of Starcamp focuses on AI track We will focus on cutting-edge technology trends, commercialization scenarios and industrial ecological development, help new and innovative enterprises in the upstream and downstream of AI, polish products, verify MVP and commercialization capabilities, efficient capital docking, and accelerate the financing process.
