As the iteration of technology collides with capital games in the wave of algorithms, China's AI entrepreneurship is also undergoing a profound transformation from fervor to rationality.
This technology revolution sparked by ChatGPT not only reshapes the boundaries of technology application but also drives the entrepreneurial ecosystem from "barbaric growth" to "meticulous cultivation." "Compared to the coal industry, the future of artificial intelligence (AI) mining and production will only increase and benefit future generations even more," said a coal boss during an interview with a reporter from China Business Journal. In the tide of the times, this coal boss is eager to know where the opportunities for involvement lie.
In a nondescript office building in Beijing's North Third Ring Road, a startup salon themed "Kunlun Nest" is crowded with people — most of them in their thirties and forties, carrying backpacks, discussing topics centered around AI and entrepreneurship — the opportunity that the coal boss is looking forward to may just be here.
The transformation of the entrepreneurial ecosystem
Since the birth of ChatGPT, in the past two years, AI technology has iterated at a "frog-jumping" speed, and the entrepreneurial ecosystem has gradually returned from early enthusiasm to rationality.
Some call the "Kunlun Nest" entrepreneurial salon the "barometer of AI entrepreneurship in China," while others jokingly refer to it as the "last bastion of idealists." From technical backbones with backgrounds in large companies, serial entrepreneurs, traditional industry players transitioning across fields, to freshly graduated university students, they come with their fervor and confusion about AI, trying to find their own answers in the waves of algorithms and capital.
For example, in the cultural industry, AI is reshaping the boundaries of creation. Writer and screenwriter Fu Yao has witnessed the efficiency revolution that AI brings to the film and television industry: a script adaptation that used to require 3 to 5 screenwriters and take half a year can now be generated in just 10 minutes by AI. The AI large model he participated in training can generate logically coherent murder plotlines by analyzing the narrative structures of classic series, yet it struggles in the creation of emotionally nuanced love dramas. This characteristic of "strong logic, weak emotion" highlights AI's unique positioning in content creation — assisting in framework construction rather than replacing human emotional insights.
Fu Yao emphasized that the essence of AI's "thinking" is the logical association of code to "token," rather than true creativity. Even so, the increase in efficiency is enough to overturn traditional perceptions. "In the past, preparing a 30-episode series took several years; in the future, it may be compressed to a few months. Costs can be significantly reduced, but the quality of art remains an unknown."
Compared to adapting scripts, AI generating video animations directly based on instructions is currently more convenient and effective. Han Lei, the founder of Intelligent Manufacturing Future, feels this deeply. His mecha sci-fi IP "Spirit Core World" had previously struggled to materialize due to high production costs, but now, with the help of AI animation technology, the production cost of a pilot episode has been significantly reduced.
In this context, more and more startup teams are beginning to focus on the application of AI in the entertainment industry. The reporter saw that products such as Hexinpai and Shenyandayi have already taken shape in the Zhongguancun Artificial Intelligence Large Model Industry Cluster located in Dongsheng Building. Taking Hexinpai as an example, this is a one-stop, low-threshold AI music workstation that can perform AI lyric writing, AI composition, AI arrangement, and song synthesis. The music works created by core users through this tool have already received over one hundred million plays across the internet.
In the downstream, entrepreneurs are also working hard to develop "AI+" application products, which include voice terminals, smart cars, visual products, robots, as well as smart education, smart healthcare, smart manufacturing, smart finance, and smart security; in the midstream, there are basic open-source frameworks and technology development platforms, as well as various algorithm models and general technologies; in the upstream, there are data services and hardware devices.
However, most of the startups focus on "AI+" application products. A relevant person in charge of the Dongsheng Technology Incubator told reporters that in 2024, the incubator added more than 100 AI startups, most of which are based on AI large models for application development.
Zhou Shangjin, the head of Kunlun Nest AIGC Open Laboratory and an internet veteran with 20 years of experience, believes that the competition threshold for large models is too high for small companies and even medium-sized listed companies to manage. On one hand, training large models is expensive; on the other hand, the talent requirements are high. Therefore, there are very few players in the large model space worldwide, and most people can only develop applications on top of large models.
"Ten years ago, everyone wanted to be Lei Jun; now, earning 200 yuan a day using AI might be the 'success' for ordinary people." said an AI entrepreneur.
Open up new blue oceans in segmented scenarios.
The changes in the entrepreneurial ecosystem have made capital more cautious, and the increased requirements for technological thresholds have made market opportunities more segmented.
Mo Xiaoyi is a serial entrepreneur who is applying AI to the financial sector to assist banks in managing non-performing assets. "Currently, the scale of any non-performing asset package in banks is quite large. Previously, non-performing asset packages relied on manual analysis of debt composition, default situations, and forming evidence chains, which required auditors to review each case individually. However, with the help of AI for batch processing, efficiency can be improved, at least forming a complete data structure and data chain, ultimately establishing a full set of processes."
Wang Yongwang previously worked as a developer at a "metaverse" gaming company. Noticing the rapid development of AI, Wang Yongwang decided to join the wave of AI entrepreneurship. He is currently working on applying AI to the catering industry, establishing a local knowledge base for each restaurant, interacting with consumers, chatting with consumers, and helping consumers plan meal sets, etc.
Another entrepreneur, Gao Shan (a pseudonym), previously worked in software development. In this wave of AI, he is targeting the silver-haired economy sector. "Helping the elderly create a memoir, and then based on the information in the memoir, establishing a local knowledge base. The final product will take the form of a desktop robot that can tell their stories in the tone and manner of the elderly, interacting with their descendants," Gao Shan stated.
The exploration in the field of education is quite representative. Li Shanming breaks down the barriers to technology popularization with the "gamified AI class," allowing children to master AI command logic through programming games. In the past year, he has tapped into two trends at the intersection of preventing addiction and AI enlightenment: on one side is parents' anxiety about "mobile phone addiction," and on the other side is the educational revolution brought about by the popularization of AI technology.
Unlike Li Shanming's popular education, Ke Qiang (pseudonym) emphasizes that "humans are responsible for strategic thinking, while AI executes tactical details." What he is doing is helping primary and secondary schools build AI courses and write teaching materials, focusing on the cultivation of teenagers' thinking and logical abilities. Currently, the understanding of AI in the education sector is still insufficient, and those engaged in technology development do not have a complete understanding of education, which allows Ke Qiang to see market potential: "Now various schools and education committees are looking for people who understand both education and AI, and our team happens to have the capability to do this part of the work."
Although the niche areas targeted by the aforementioned grassroots entrepreneurs are different, there is a common point that their startup costs are not high. Wang Yongwang and Gao Shan are currently working alone, Li Shanming has also made it clear that he will not expand his team, and Ke Qiang's team consists of only four people.
Pang Pengdan, Deputy Secretary-General of the China Mobile Communications Federation's Computing Power Ecological Alliance, believes that among grassroots entrepreneurs, there is a special term called "one-person company," where one person can complete the entire business loop of the company. On one hand, AI reduces labor costs and improves work efficiency; on the other hand, AI can also help people quickly produce new products and then conduct online transactions, quickly generating revenue.
During a visit to several incubators in Beijing, reporters found that many startup companies are essentially small teams of two or three people.
Ke Qiang bluntly stated: "In the current market environment, the cost of starting a business cannot be high. You must use minimal social resources, minimal funds, and a very small team to complete the entrepreneurial process. This is the prerequisite to test whether each entrepreneur can get started."
"These practices indicate that the value of AI entrepreneurship lies not in disrupting tradition, but in empowering niche scenarios with technology to address real needs," commented a market observer.
The "last mile" of technology implementation.
In fact, apart from the application of large model privatization deployment and AI education, hardware products are more favored by capital due to the lowering of the AI usage threshold. For example, AI integrated machines are experiencing a boom.
A report released by Peking University mentioned that the marginal cost in the AI era is significantly higher than in the internet era (requiring continuous computing power and token consumption), while the all-in-one machine can alleviate this pain point through "software-hardware synergy + localized deployment."
The reporter learned that the Kunlun Nest humanoid robot open laboratory has designed an all-in-one product named DeepBook. In appearance, it is a black square box, priced below ten thousand yuan, and has already sold over a hundred units. According to Dr. Guo Chengkai from the Kunlun Nest humanoid robot open laboratory, most of the current buyers are state-owned enterprises, universities, or some companies looking to avoid the risk of data leakage. By turning the company's relevant data into a knowledge base and feeding it into the all-in-one machine, and deploying the machine in the company's server room, it can be accessed directly and then interact with the company's knowledge base.
In addition, there are robotic arm products. Yang Kai from the Kunlun Nest humanoid robot open laboratory team told reporters: "Our robotic arms are also in a semi-commercial state. We have developed two generations of products and are about to update to the third generation, which can be sold externally. One of the member units of the laboratory, the 'Dexterous Hand,' sold out last year, priced at over 60,000 yuan, with two to three hundred sets sold."
However, Yang Kai admitted that it would take 5 to 10 years for embodied intelligent robots to develop into what is seen in movies. "The hardware will definitely lag behind the development of large models because the hardware needs to be based on the computing power provided by the large models."
"Currently, some technologies of robots are still very immature, and many are experimental. For example, the current humanoid robots have a load capacity of only 10-20 kilograms, which does not meet the needs in many scenarios. Additionally, for service robots, the strength to assist an elderly person to the bathroom or to lift an elderly person from a wheelchair is insufficient." Yang Kai stated that these physical issues cannot be resolved merely by software updates; the entire industry needs to iterate.
During the interview, Yang Kai casually showed the reporter a small white box, about the size of an AirPods case, with a metal button on top. Pressing it allows for conversation, similar to a portable version of "Doubao". "There are quite a few products like this on the market currently, but there is still a lot that can be done later, such as setting different roles, different personalities, and then binding different knowledge bases to different roles, making the conversations even more interesting."
At the capital level, grassroots entrepreneurs are experiencing "ice and fire". According to data released by IT Orange, the overall scale of primary market financing in Chinese's artificial intelligence industry has expanded from 30.07 billion yuan in 2015 to 105.251 billion yuan in 2024, achieving a 3.5-fold increase. Early-stage investments accounted for 42% of the overall financing scale in 2024.
However, the proportion of grassroots entrepreneurs receiving investment is extremely low — the market prefers teams with efficient laboratory transformation, top research backgrounds, or experience from large companies.
The reporter learned that large entrepreneurship incubators generally provide entrepreneurs with a complete full-cycle service. For example, Taili Station divides business growth into three key stages: the R&D stage, the market and growth stage, and the maturity stage. At different stages of business development, different support such as angel funds, industrial capital, and mergers/IPO will be provided.
The Xingdi AI Application Incubator in Chaoyang District, Beijing, has joined forces with seven well-known investment institutions, including Innoangel Fund, Plum Ventures, Lenovo Capital, Tsinghua Alumni Fund, Oasis Capital, Xiaomiao Langcheng, Yuanhe Yuandian, as well as more than 20 ecological resource partners, to provide capital market consulting and training for startups, helping them solve the early-stage funding difficulties.
Not all enterprises can successfully settle into the incubator. For instance, when selecting entrepreneurial teams, the Star Earth AI Application Incubator takes multiple factors into consideration, such as project management team, technological maturity and innovation, market prospects, business models and operational results, as well as social responsibility.
Not only that, when incubators choose entrepreneurial teams, they share a similar perspective with capital institutions, where "people" are a key consideration. A representative from an investment institution told reporters: "The most important thing we look at when evaluating projects is the team. The best combination is to have academic research resources from universities, along with talents who understand the market and marketing."
Paradigm shift in entrepreneurial logic
"In the last wave of entrepreneurship, grassroots entrepreneurs could generally secure investment, ranging from several million to a few hundred thousand yuan, which was enough to support a startup team for research and development. However, at this stage, many investors are facing tight cash flow and are unwilling to invest based solely on one entrepreneur's dream. Therefore, it is generally difficult for grassroots entrepreneurs to obtain funding," Ke Qiang said.
Su Dian, the founder of Garage Cafe and Kunlun Nest, also mentioned that the current AI application layer is facing a dual dilemma: accelerated technological iteration has led to unclear investment directions, and the narrowing exit channels have intensified capital anxiety.
The aforementioned investment institution personnel told reporters that currently many investment institutions are investing in hard technology manufacturing, with a greater focus on upstream investments, such as semiconductor chips and materials for upstream chips. The end products are similar to embodied intelligent robots, leaning more towards certain aspects of the industrial chain related to hard technology.
"From large models to application implementation in AI, among the mainstream investment institutions in China, many are observing, and very few are actually taking action. The iteration of large models is very fast, and whether the investments made now will align with the true technological direction of the future is not very likely, so everyone is still in a wait-and-see mode. The situation is quite similar at the application level; AI is changing very quickly, making it difficult to judge which application can truly succeed." The aforementioned investment institution personnel stated that for investment institutions, it is also a process of learning and adaptation.
Liu Yang, a founding member of Baidu Ventures with 15 years of venture capital experience, shares the same view: "Today, AI has not clearly conveyed its future direction. Up to now, it is still unclear what can create tremendous commercial value, and the overall situation is still in exploration."
Unlike the logic of "burning money to gain scale" in the era of mobile internet, today's entrepreneurs need to confront cash flow pressures head-on. "Projects without self-sustainability will not survive beyond three months," said Han Lei.
In fact, Han Lei experienced a long period of career rise. In 2015, a mecha robot named NK01 made its debut at Hunan TV's New Year's Eve Gala, generating quite a buzz. This robot was designed by Han Lei's Intelligent Manufacturing Future. Today, it is still possible to find news online about this product's appearances at various music festivals and variety shows.
According to Han Lei, the leasing business of the NK01 prototype in 2015 had quite good revenue. "But due to a lack of systematic commercial awareness, the business model went off track, not focusing on the robot performance market, but instead choosing to engage in cultural tourism." In Han Lei's words, "A 'big piece of meat' was right in front of them, yet they did not cherish it properly."
It is worth noting that many entrepreneurs today are no longer obsessed with "disrupting giants," but are more focused on "ecological symbiosis." Liu Yang stated: "In the era of mobile internet entrepreneurship, people have witnessed too many stories of giants fighting against each other and entrepreneurs disrupting giants. In the age of AI entrepreneurship, more entrepreneurs will choose to jointly explore the future business ecology."
Zhou Shangjin Hang also mentioned: "Instead of complaining about big companies reaping the rewards, it's better to find the niche markets they overlook." For example, some teams focus on customizing private AI knowledge bases for enterprises or developing specialized tools using open-source models, thus avoiding direct competition with large companies while creating a differentiated advantage.
"The shift in entrepreneurial logic is not decline, but maturity." Su Lin summarized, "When the era of crazy expansion ends, surviving and creating real value is the main theme of the new cycle."
From the anxiety of coal bosses to the pragmatism of entrepreneurs, AI entrepreneurship is returning to its commercial essence. When the technological frenzy subsides, those projects that can solve specific problems and create real value may truly be the pioneers of the AI era.
The content is for reference only, not a solicitation or offer. No investment, tax, or legal advice provided. See Disclaimer for more risks disclosure.
The Entrepreneurial Landscape Amidst AI Frenzy
As the iteration of technology collides with capital games in the wave of algorithms, China's AI entrepreneurship is also undergoing a profound transformation from fervor to rationality.
This technology revolution sparked by ChatGPT not only reshapes the boundaries of technology application but also drives the entrepreneurial ecosystem from "barbaric growth" to "meticulous cultivation." "Compared to the coal industry, the future of artificial intelligence (AI) mining and production will only increase and benefit future generations even more," said a coal boss during an interview with a reporter from China Business Journal. In the tide of the times, this coal boss is eager to know where the opportunities for involvement lie.
In a nondescript office building in Beijing's North Third Ring Road, a startup salon themed "Kunlun Nest" is crowded with people — most of them in their thirties and forties, carrying backpacks, discussing topics centered around AI and entrepreneurship — the opportunity that the coal boss is looking forward to may just be here.
The transformation of the entrepreneurial ecosystem
Since the birth of ChatGPT, in the past two years, AI technology has iterated at a "frog-jumping" speed, and the entrepreneurial ecosystem has gradually returned from early enthusiasm to rationality.
Some call the "Kunlun Nest" entrepreneurial salon the "barometer of AI entrepreneurship in China," while others jokingly refer to it as the "last bastion of idealists." From technical backbones with backgrounds in large companies, serial entrepreneurs, traditional industry players transitioning across fields, to freshly graduated university students, they come with their fervor and confusion about AI, trying to find their own answers in the waves of algorithms and capital.
For example, in the cultural industry, AI is reshaping the boundaries of creation. Writer and screenwriter Fu Yao has witnessed the efficiency revolution that AI brings to the film and television industry: a script adaptation that used to require 3 to 5 screenwriters and take half a year can now be generated in just 10 minutes by AI. The AI large model he participated in training can generate logically coherent murder plotlines by analyzing the narrative structures of classic series, yet it struggles in the creation of emotionally nuanced love dramas. This characteristic of "strong logic, weak emotion" highlights AI's unique positioning in content creation — assisting in framework construction rather than replacing human emotional insights.
Fu Yao emphasized that the essence of AI's "thinking" is the logical association of code to "token," rather than true creativity. Even so, the increase in efficiency is enough to overturn traditional perceptions. "In the past, preparing a 30-episode series took several years; in the future, it may be compressed to a few months. Costs can be significantly reduced, but the quality of art remains an unknown."
Compared to adapting scripts, AI generating video animations directly based on instructions is currently more convenient and effective. Han Lei, the founder of Intelligent Manufacturing Future, feels this deeply. His mecha sci-fi IP "Spirit Core World" had previously struggled to materialize due to high production costs, but now, with the help of AI animation technology, the production cost of a pilot episode has been significantly reduced.
In this context, more and more startup teams are beginning to focus on the application of AI in the entertainment industry. The reporter saw that products such as Hexinpai and Shenyandayi have already taken shape in the Zhongguancun Artificial Intelligence Large Model Industry Cluster located in Dongsheng Building. Taking Hexinpai as an example, this is a one-stop, low-threshold AI music workstation that can perform AI lyric writing, AI composition, AI arrangement, and song synthesis. The music works created by core users through this tool have already received over one hundred million plays across the internet.
In the downstream, entrepreneurs are also working hard to develop "AI+" application products, which include voice terminals, smart cars, visual products, robots, as well as smart education, smart healthcare, smart manufacturing, smart finance, and smart security; in the midstream, there are basic open-source frameworks and technology development platforms, as well as various algorithm models and general technologies; in the upstream, there are data services and hardware devices.
However, most of the startups focus on "AI+" application products. A relevant person in charge of the Dongsheng Technology Incubator told reporters that in 2024, the incubator added more than 100 AI startups, most of which are based on AI large models for application development.
Zhou Shangjin, the head of Kunlun Nest AIGC Open Laboratory and an internet veteran with 20 years of experience, believes that the competition threshold for large models is too high for small companies and even medium-sized listed companies to manage. On one hand, training large models is expensive; on the other hand, the talent requirements are high. Therefore, there are very few players in the large model space worldwide, and most people can only develop applications on top of large models.
"Ten years ago, everyone wanted to be Lei Jun; now, earning 200 yuan a day using AI might be the 'success' for ordinary people." said an AI entrepreneur.
Open up new blue oceans in segmented scenarios.
The changes in the entrepreneurial ecosystem have made capital more cautious, and the increased requirements for technological thresholds have made market opportunities more segmented.
Mo Xiaoyi is a serial entrepreneur who is applying AI to the financial sector to assist banks in managing non-performing assets. "Currently, the scale of any non-performing asset package in banks is quite large. Previously, non-performing asset packages relied on manual analysis of debt composition, default situations, and forming evidence chains, which required auditors to review each case individually. However, with the help of AI for batch processing, efficiency can be improved, at least forming a complete data structure and data chain, ultimately establishing a full set of processes."
Wang Yongwang previously worked as a developer at a "metaverse" gaming company. Noticing the rapid development of AI, Wang Yongwang decided to join the wave of AI entrepreneurship. He is currently working on applying AI to the catering industry, establishing a local knowledge base for each restaurant, interacting with consumers, chatting with consumers, and helping consumers plan meal sets, etc.
Another entrepreneur, Gao Shan (a pseudonym), previously worked in software development. In this wave of AI, he is targeting the silver-haired economy sector. "Helping the elderly create a memoir, and then based on the information in the memoir, establishing a local knowledge base. The final product will take the form of a desktop robot that can tell their stories in the tone and manner of the elderly, interacting with their descendants," Gao Shan stated.
The exploration in the field of education is quite representative. Li Shanming breaks down the barriers to technology popularization with the "gamified AI class," allowing children to master AI command logic through programming games. In the past year, he has tapped into two trends at the intersection of preventing addiction and AI enlightenment: on one side is parents' anxiety about "mobile phone addiction," and on the other side is the educational revolution brought about by the popularization of AI technology.
Unlike Li Shanming's popular education, Ke Qiang (pseudonym) emphasizes that "humans are responsible for strategic thinking, while AI executes tactical details." What he is doing is helping primary and secondary schools build AI courses and write teaching materials, focusing on the cultivation of teenagers' thinking and logical abilities. Currently, the understanding of AI in the education sector is still insufficient, and those engaged in technology development do not have a complete understanding of education, which allows Ke Qiang to see market potential: "Now various schools and education committees are looking for people who understand both education and AI, and our team happens to have the capability to do this part of the work."
Although the niche areas targeted by the aforementioned grassroots entrepreneurs are different, there is a common point that their startup costs are not high. Wang Yongwang and Gao Shan are currently working alone, Li Shanming has also made it clear that he will not expand his team, and Ke Qiang's team consists of only four people.
Pang Pengdan, Deputy Secretary-General of the China Mobile Communications Federation's Computing Power Ecological Alliance, believes that among grassroots entrepreneurs, there is a special term called "one-person company," where one person can complete the entire business loop of the company. On one hand, AI reduces labor costs and improves work efficiency; on the other hand, AI can also help people quickly produce new products and then conduct online transactions, quickly generating revenue.
During a visit to several incubators in Beijing, reporters found that many startup companies are essentially small teams of two or three people.
Ke Qiang bluntly stated: "In the current market environment, the cost of starting a business cannot be high. You must use minimal social resources, minimal funds, and a very small team to complete the entrepreneurial process. This is the prerequisite to test whether each entrepreneur can get started."
"These practices indicate that the value of AI entrepreneurship lies not in disrupting tradition, but in empowering niche scenarios with technology to address real needs," commented a market observer.
The "last mile" of technology implementation.
In fact, apart from the application of large model privatization deployment and AI education, hardware products are more favored by capital due to the lowering of the AI usage threshold. For example, AI integrated machines are experiencing a boom.
A report released by Peking University mentioned that the marginal cost in the AI era is significantly higher than in the internet era (requiring continuous computing power and token consumption), while the all-in-one machine can alleviate this pain point through "software-hardware synergy + localized deployment."
The reporter learned that the Kunlun Nest humanoid robot open laboratory has designed an all-in-one product named DeepBook. In appearance, it is a black square box, priced below ten thousand yuan, and has already sold over a hundred units. According to Dr. Guo Chengkai from the Kunlun Nest humanoid robot open laboratory, most of the current buyers are state-owned enterprises, universities, or some companies looking to avoid the risk of data leakage. By turning the company's relevant data into a knowledge base and feeding it into the all-in-one machine, and deploying the machine in the company's server room, it can be accessed directly and then interact with the company's knowledge base.
In addition, there are robotic arm products. Yang Kai from the Kunlun Nest humanoid robot open laboratory team told reporters: "Our robotic arms are also in a semi-commercial state. We have developed two generations of products and are about to update to the third generation, which can be sold externally. One of the member units of the laboratory, the 'Dexterous Hand,' sold out last year, priced at over 60,000 yuan, with two to three hundred sets sold."
However, Yang Kai admitted that it would take 5 to 10 years for embodied intelligent robots to develop into what is seen in movies. "The hardware will definitely lag behind the development of large models because the hardware needs to be based on the computing power provided by the large models."
"Currently, some technologies of robots are still very immature, and many are experimental. For example, the current humanoid robots have a load capacity of only 10-20 kilograms, which does not meet the needs in many scenarios. Additionally, for service robots, the strength to assist an elderly person to the bathroom or to lift an elderly person from a wheelchair is insufficient." Yang Kai stated that these physical issues cannot be resolved merely by software updates; the entire industry needs to iterate.
During the interview, Yang Kai casually showed the reporter a small white box, about the size of an AirPods case, with a metal button on top. Pressing it allows for conversation, similar to a portable version of "Doubao". "There are quite a few products like this on the market currently, but there is still a lot that can be done later, such as setting different roles, different personalities, and then binding different knowledge bases to different roles, making the conversations even more interesting."
At the capital level, grassroots entrepreneurs are experiencing "ice and fire". According to data released by IT Orange, the overall scale of primary market financing in Chinese's artificial intelligence industry has expanded from 30.07 billion yuan in 2015 to 105.251 billion yuan in 2024, achieving a 3.5-fold increase. Early-stage investments accounted for 42% of the overall financing scale in 2024.
However, the proportion of grassroots entrepreneurs receiving investment is extremely low — the market prefers teams with efficient laboratory transformation, top research backgrounds, or experience from large companies.
The reporter learned that large entrepreneurship incubators generally provide entrepreneurs with a complete full-cycle service. For example, Taili Station divides business growth into three key stages: the R&D stage, the market and growth stage, and the maturity stage. At different stages of business development, different support such as angel funds, industrial capital, and mergers/IPO will be provided.
The Xingdi AI Application Incubator in Chaoyang District, Beijing, has joined forces with seven well-known investment institutions, including Innoangel Fund, Plum Ventures, Lenovo Capital, Tsinghua Alumni Fund, Oasis Capital, Xiaomiao Langcheng, Yuanhe Yuandian, as well as more than 20 ecological resource partners, to provide capital market consulting and training for startups, helping them solve the early-stage funding difficulties.
Not all enterprises can successfully settle into the incubator. For instance, when selecting entrepreneurial teams, the Star Earth AI Application Incubator takes multiple factors into consideration, such as project management team, technological maturity and innovation, market prospects, business models and operational results, as well as social responsibility.
Not only that, when incubators choose entrepreneurial teams, they share a similar perspective with capital institutions, where "people" are a key consideration. A representative from an investment institution told reporters: "The most important thing we look at when evaluating projects is the team. The best combination is to have academic research resources from universities, along with talents who understand the market and marketing."
Paradigm shift in entrepreneurial logic
"In the last wave of entrepreneurship, grassroots entrepreneurs could generally secure investment, ranging from several million to a few hundred thousand yuan, which was enough to support a startup team for research and development. However, at this stage, many investors are facing tight cash flow and are unwilling to invest based solely on one entrepreneur's dream. Therefore, it is generally difficult for grassroots entrepreneurs to obtain funding," Ke Qiang said.
Su Dian, the founder of Garage Cafe and Kunlun Nest, also mentioned that the current AI application layer is facing a dual dilemma: accelerated technological iteration has led to unclear investment directions, and the narrowing exit channels have intensified capital anxiety.
The aforementioned investment institution personnel told reporters that currently many investment institutions are investing in hard technology manufacturing, with a greater focus on upstream investments, such as semiconductor chips and materials for upstream chips. The end products are similar to embodied intelligent robots, leaning more towards certain aspects of the industrial chain related to hard technology.
"From large models to application implementation in AI, among the mainstream investment institutions in China, many are observing, and very few are actually taking action. The iteration of large models is very fast, and whether the investments made now will align with the true technological direction of the future is not very likely, so everyone is still in a wait-and-see mode. The situation is quite similar at the application level; AI is changing very quickly, making it difficult to judge which application can truly succeed." The aforementioned investment institution personnel stated that for investment institutions, it is also a process of learning and adaptation.
Liu Yang, a founding member of Baidu Ventures with 15 years of venture capital experience, shares the same view: "Today, AI has not clearly conveyed its future direction. Up to now, it is still unclear what can create tremendous commercial value, and the overall situation is still in exploration."
Unlike the logic of "burning money to gain scale" in the era of mobile internet, today's entrepreneurs need to confront cash flow pressures head-on. "Projects without self-sustainability will not survive beyond three months," said Han Lei.
In fact, Han Lei experienced a long period of career rise. In 2015, a mecha robot named NK01 made its debut at Hunan TV's New Year's Eve Gala, generating quite a buzz. This robot was designed by Han Lei's Intelligent Manufacturing Future. Today, it is still possible to find news online about this product's appearances at various music festivals and variety shows.
According to Han Lei, the leasing business of the NK01 prototype in 2015 had quite good revenue. "But due to a lack of systematic commercial awareness, the business model went off track, not focusing on the robot performance market, but instead choosing to engage in cultural tourism." In Han Lei's words, "A 'big piece of meat' was right in front of them, yet they did not cherish it properly."
It is worth noting that many entrepreneurs today are no longer obsessed with "disrupting giants," but are more focused on "ecological symbiosis." Liu Yang stated: "In the era of mobile internet entrepreneurship, people have witnessed too many stories of giants fighting against each other and entrepreneurs disrupting giants. In the age of AI entrepreneurship, more entrepreneurs will choose to jointly explore the future business ecology."
Zhou Shangjin Hang also mentioned: "Instead of complaining about big companies reaping the rewards, it's better to find the niche markets they overlook." For example, some teams focus on customizing private AI knowledge bases for enterprises or developing specialized tools using open-source models, thus avoiding direct competition with large companies while creating a differentiated advantage.
"The shift in entrepreneurial logic is not decline, but maturity." Su Lin summarized, "When the era of crazy expansion ends, surviving and creating real value is the main theme of the new cycle."
From the anxiety of coal bosses to the pragmatism of entrepreneurs, AI entrepreneurship is returning to its commercial essence. When the technological frenzy subsides, those projects that can solve specific problems and create real value may truly be the pioneers of the AI era.
(Source: China Business Network)
Source: Dongfang Caifu Net
Author: China Business Network