Author: Frank, PANews

At most technology conferences, the most common question is "who released what." However, at the mu Shanghai AI WEEK in May 2026, the frequently heard question from PANews shifted to a more practical topic: As AI makes it easier to create product prototypes, where has the real difficulty of entrepreneurship shifted?

What makes this event special is that it resembles more of a temporary developer space than a standard conference. There are few booths, little corporate presentations, and the topics of discussion are not fixed. A large number of overseas developers flew from Argentina, Silicon Valley, Japan, or Southeast Asia to Shanghai, just to connect with Chinese developers, model companies, investors, and the local ecosystem over the course of a month.

The event space was not set up like a traditional hotel venue, but rather a mixed space composed of open office areas, tiered cushions, beanbags, and temporary projection setups. Some participants sat at their desks coding, some gathered on the carpet and square cushions to listen to presentations, while others leaned in corners, continuing to work on their products on laptops. Colorful flags of mu Shanghai adorned the walls, and a world map with the question "Who am I? What shapes me?" was plastered with sticky notes and connection lines, resembling an identity network being collaboratively filled out by the participants.

After engaging with several organizers, project stakeholders, investors, and representatives from model companies at the event, PANews discovered that AI entrepreneurship is entering a new phase. If "who can access models faster to create products" was the first phase of AI entrepreneurship, then the second phase is "who can find real scenarios, acquire users, build communities, and survive over a sufficiently long period." Models are like water, electricity, and gas, so what is truly scarce now is not the ability to connect to utilities, but finding who needs water the most.

A Deep Social Experiment for Global Developers

What distinguishes mu Shanghai is primarily reflected in its organizational form. Sun, the founder of the event, mentioned in an interview with PANews that mu did not initially start in China, but rather spread through pop-up cities and entrepreneurial communities in Thailand, Argentina, Africa, Japan, and other places. Compared to traditional conferences lasting two or three days, it emphasizes a group of people entering the same city to co-create, communicate, live, and build relationships over about a month.

This format inherently imbues the event with a strong community attribute. According to Sun, about 2,000 people registered for mu Shanghai, and more than 800 were ultimately selected. The participant composition is quite diverse: approximately 20% from China, around 18% from Japan, South Korea, and other Asian regions, about 16% from Southeast Asia, and around 10%, 10%, and 11% from Latin America, the United States, and Europe, respectively, with Africa making up about 6%. In terms of industry background, around 40% of participants are AI practitioners, and those involved in Web3 account for about 20% to 30%, along with various groups from hardware, biotechnology, investment, and more.

Sun explains the attraction of this event format: "After leaving college, people rarely have that kind of deep relationships. It is also difficult to form such connections in work and big cities, so I think it is very valuable." In his view, mu tries to replicate not the traffic moment of traditional conferences but a relationship density that is closer to university, community, and co-living.

The on-site atmosphere indeed reflects this state. The main stage is not always at the center of the space; the subtitle screens next to the projection screen, the temporarily set up display racks, and computers scattered around collectively form the backdrop of the event. During a session about user experience, the audience did not sit neatly in chairs but rather dispersed among low cushions, the floor, and open workspaces. The speaker shares up front while the audience listens, taking notes, responding to messages, or continuing to work on their projects. This somewhat loose state more closely resembles the real operational mode of a developer community.

The significance of these numbers lies not in the scale of the event itself but in the way it showcases an organizational logic different from traditional exhibitions. Traditional conferences often connect brands with users and companies with clients, while mu Shanghai seems to connect developer cultures from both China and abroad. The event featured model roundtables, hackathons, co-creation activities, language learning sessions, community sharing, and discussions that were spontaneously organized. Feng Wen, the product leader at MiniMax, mentioned during the interaction that the atmosphere here is not just about "sharing AI on stage," but also includes cultural exchanges, co-creation among developers, and community participation.

The influx of numerous Web3 practitioners has also made these connections more complex. The Web3 industry has sedimented not only on-chain assets and speculative narratives over the past few years but also a set of methods for community mobilization, global cooperation, social media dissemination, and developer organization. As AI entrepreneurship shifts from competing on model usage to competing on user reach, these methods have regained their value.

From "How to Do" to "Who to Sell To": AI Entrepreneurship Enters Deeper Waters

The most evident feeling PANews had on site was that AI entrepreneurs are no longer excited for too long about "whether they can create a product." Multimodal models, code generation tools, agent frameworks, and automated workflows are rapidly lowering the barriers to product prototyping. A small tool that once required designers, engineers, and operators can now be created by just a few people over a few nights with the help of AI coding tools.

Updated data better illustrates this change in threshold. The AI Pulse survey conducted by JetBrains in January 2026 showed that 90% of professional developers are now regularly using at least one AI tool in their work, while 74% have adopted developer-specific AI tools. For entrepreneurs, "being able to make it" is becoming a more universal capability and no longer a natural barrier.

However, once a product has been created, the real questions begin. An entrepreneur named Nathan told PANews that he is working on a product that helps AI entrepreneurs find their entrepreneurial direction. Its logic is that AI can already expand the scope of information gathering and distill the judgment and taste of serial entrepreneurs into a set of judgment rules, which can then be handed to AI to discover business opportunity signals. But this product itself reveals a larger reality: when creating products becomes easier, the question of "what to create" has become the more scarce issue.

Nathan told PANews: "With the help of AI coding tools, creating something new is quick. The real key is whether this direction is worth pursuing." The product he is working on essentially turns the act of "finding direction" into a product itself. This small case, however, reflects a new shift in AI entrepreneurship: as execution ability is amplified by AI, judgment ability has become a scarce asset.

In the roundtable discussion on "Innovative Practices and Path Explorations in the AI Consumer Ecosystem" hosted by PANews, multiple guests also expressed similar views: while AI does make rapid prototyping, demo samples, and initial launches easier, the real difficulties of entrepreneurship have not disappeared. Acquiring customers, commercial realization, community stickiness, user education, and the connections between people still require teams to possess more complex capabilities.

In other words, AI lowers the threshold for development, not for entrepreneurship. The first threshold of product competition in the past was "can it be created," and now that threshold has been significantly lowered, the true filtering begins to shift to distribution, scenarios, and commercialization. One interviewee summarized it as follows: creating tools is not difficult now; the challenge is making products, IP, and value visible to more people.

This is also a common dilemma faced by many AI tools. The more tools available, the harder it is for users to make choices; the stronger the model, the more likely a single-point function will be swallowed up by the next model update. For entrepreneurs, a product that seems viable today may lose its presence in six months due to improved underlying model capabilities. Therefore, the real question is not "should we do AI" but whether a specific scenario can be found that a model cannot completely overshadow in the short term.

The use of AI is spreading rapidly, but there remains a gap between tool usage and stable value, separated by scenario, process, governance, and organizational capability.

Web3 People Surging into AI, Not Just Chasing Trends

From a narrative perspective, the influx of Web3 people into AI may seem like another trend migration. However, at mu Shanghai, there are more realistic reasons behind this migration.

On the one hand, the wealth effect, capital dividend, and technological dividend of the crypto industry are diminishing, prompting many practitioners to search for new technological directions; on the other hand, AI applications require capabilities that the Web3 industry is most familiar with, such as community involvement, global dissemination, developer relations, and social media distribution.

A seasoned Web3 practitioner candidly expressed on site that the crypto industry has been around for ten years, and the capital and cognitive dividends have largely ended, indicating that it is time to move toward new technological directions. He suggested that entrepreneurs gradually shift their careers, personal brands, and asset allocations toward AI instead of continuing to invest vast amounts of energy in cryptocurrency. This judgment may not represent all Web3 practitioners, but it does reflect the true mindset of some attendees.

His expression was straightforward: "I believe AI is worth long-term investment. Investment doesn't just mean using tools but gradually shifting careers, personal brands, and asset allocations towards AI." His personal choice is to transition into an AI blogger, using a sports camera to capture Vlogs of teams working on AI products onsite.

These judgments may not represent all Web3 practitioners but clearly illustrate the atmosphere of the event: AI is no longer just an optional track but is becoming a direction for some Web3 practitioners to reconfigure their time, assets, and professional identities.

AI-driven social media assistant XerpaAI set up a booth onsite, where staff stated in an interview, "We are a pure AI project, technically not closely related to Web3. However, from the user's perspective, we will definitely reach Web3 users. For example, the X AI assistant will serve some Web3 users with operational needs." This statement represents the current ambiguous relationship between AI applications and the Web3 community: products may not be Web3, but users, dissemination, and early demand often cannot bypass Web3.

During onsite discussions, representatives from model companies noted that the user bases of AI and Web3 are becoming increasingly difficult to separate entirely, as many heavy users of AI tools actually come from a Web3 background. Especially in scenarios like Hong Kong and Shanghai, AI and Web3 often share the same group of high-frequency attendees, early adopters, and community dissemination nodes. For them, there is no dismissal of whether community members are Web3 users, as long as the theme is AI, everyone’s goals are aligned.

From this perspective, Web3 entering AI is not just a "transition." What Web3 brings is not just on-chain technology itself, but a methodology for gathering global developers around a project, facilitating ongoing discussions and contributions of attention. For current AI applications, this capability may be harder to replicate than a short-term feature.

Hardware, Supply Chain, and the Chinese Foundation

Compared to anxieties about "whether AI software applications will be overshadowed by models," discussions on AI hardware, embodied intelligence, and China's supply chain seem to offer more certainty. Several respondents mentioned that as AI enters the real world in the future, hardware, robotics, embodied intelligence, and multi-sensory interaction will face greater opportunities. In the consumer-level AI roundtable hosted by PANews, Feng Wen, the product leader of MiniMax, also assessed that in the next three to five years, smart hardware, robotics, and embodied intelligence will reach significant turning points, with AI no longer just residing in software interfaces but also entering the physical world.

Outside the venue, the robotics field is also becoming a focus. An overseas robotics manufacturer, Figur, hosted a human-robot sorting competition on May 18 that sparked widespread discussion online; even though humans narrowly won in a 10-hour competition, it is clear that if the time frame were extended, robots would prevail. The Stanford HAI "2026 AI Index" indicates that the accuracy rate of AI agents in real computer task tests like OSWorld has increased from about 12% to 66.3%, and autonomous driving is also beginning to achieve large-scale deployment, with China's Apollo Go completing a total of 11 million fully driverless trips.

AI's entry into the real world through hardware, robotics, and edge-side deployments is no longer just a distant narrative.

This is precisely where the unique advantages of the Chinese ecosystem lie. Sun repeatedly mentioned in the interview that China has a nearly complete supply chain from hardware, AI, and life technology to infrastructure. For overseas entrepreneurs, if they want to develop AI hardware, whether it be raw materials, factories, engineers, or rapid prototyping capabilities, it is ultimately hard to bypass China. He also revealed that many entrepreneurs who traveled from abroad to China for this event aim to experience and closely observe China's complete industrial chain.

Sun stated: "As long as one is engaged in hardware, overseas teams will ultimately return to China to seek supply chains, raw materials, engineers, and prototyping capabilities." He believes that in the next five to ten years, more international talent will come to China seeking supply chains, raw materials, talent, and capital. For overseas entrepreneurs, China is not just a market but also a set of infrastructures that enable product realization.

A venture capital person expressed to PANews that their main goal for participating in this event was to look for more hardware-oriented, embodied intelligence, and world models rather than purely consumer-end applications. Their logic is that as the replication costs of software AI decrease, hardware, supply chains, and real-world interactions may paradoxically become barriers that are harder to directly overshadow by model updates.

However, the allure of the Chinese AI ecosystem for overseas developers comes not only from the supply chain. The emergence of domestic models like DeepSeek, Kimi, MiniMax, Zhiyu, and Qianwen has begun to prompt overseas developers to reevaluate the capabilities of Chinese models. Yet, there remain trust and deployment challenges for Chinese models going abroad. Feng Wen from MiniMax's open platform mentioned that Chinese models primarily rely on open source to gain attention and brand influence overseas, but many overseas developers still worry about data, compliance, and trust issues. Even if models are open source, most individuals may not have enough computing power to deploy them independently, resulting in a layer where American companies deploy Chinese open-source models and then offer them for use to overseas clients.

For overseas developers, the appeal of China's AI ecosystem is no longer solely from cost or market scale, but also from the continuously expanding supply of models, engineering capabilities, and industrial transformation capabilities.

This means that the opportunities in the Chinese AI ecosystem are not linear. Model capabilities, hardware supply chains, government execution capacity, and developer communities need to operate together to truly attract overseas entrepreneurs. mu Shanghai plays a role in this process more like a connector for bringing overseas developers into China.

Large Model Companies Begin Competing for Developer Communities

If the competition among large model companies over the past year primarily manifested in parameters, rankings, and prices, then at mu Shanghai, the importance of the developer community has been brought to the forefront. Domestic large model companies need not only more API calls but also require developers to know them, trust them, and be willing to create applications around their models.

Feng Wen mentioned during the on-site interaction that they have conducted extensive work related to developers. Developer experience, event selection, guest participation, hackathons, judges, token sponsorship, and so on all need to be integrated into the ecological work of model companies.

"Developers are our users, which is why we pay great attention to developer experience and hope to help more developers understand what we are doing," Feng Wen stated. This statement can almost be seen as a footnote to the ecological strategy of domestic large model companies, where models are no longer just left on a platform waiting to be accessed but should actively penetrate the spaces where developers congregate.

This is not a choice unique to MiniMax. Participants on site revealed that Zhiyu has an Origin Academy in Beijing, holding activities nearly every week and getting close to resources from Tsinghua University and Peking University; the AIGC and AGI community is also continuously gathering talent through fixed spaces, hackathons, hotpot dinners, developer nights, and so on. These types of spaces are becoming an offline version of developer entrances.

Behind this is a larger change: model companies are no longer satisfied with "just releasing their models." They require documentation, trial platforms, case studies, video tutorials, as well as communities, hackathons, and developer activities to help users cross the initial hurdle. As agent capabilities improve, user education itself is also being restructured. In the past, developers needed to read documents, check error codes, and understand parameters themselves; now, agents can assist users in reading documents, searching for solutions, selecting models, and automatically correcting paths.

For model companies, true competition is not just about the price of model calls, but who can enter the developer's daily workflow earlier. For application entrepreneurs, the real opportunity is not merely about which model to integrate but whether they can find a group of early users willing to continue using, providing feedback, and even actively promoting the product.

Needed, Understood, and Retained

mu Shanghai did not provide a unified answer to AI entrepreneurship. Some see promise in hardware, others in social media growth assistants, some in identifying entrepreneurial opportunities, others discuss cultural export and spiritual consumption, and some view it as a gateway to understanding overseas developers and local partners.

Yet, these seemingly scattered clues constitute the most authentic state of AI entrepreneurship today. Model capabilities continue to advance, but application forms are still in search of stable scenarios; development thresholds have lowered, yet distribution and commercialization have become more critical; the enthusiasm for Web3 is cooling, but its community methodologies are being absorbed by AI; China's supply chain and model capabilities are gaining importance, but overseas developers still need a trustworthy entry point to understand China.

Sun mentioned in the interview that the long-term goal of mu Shanghai is not merely to host an event but to form a sustainable space where people from overseas and domestic can meet, collaborate, and create new things in the same place. In fact, mu has very few official employees, and much of the work is driven by contributors and collaborators. This organizational approach closely resembles that of Web3 and open-source communities—decentralized, contribution-focused, and relationship- network heavy—making it more attractive to those familiar with this culture.

Of course, this model still carries significant uncertainties. Whether the event can transform into a long-term space, whether community enthusiasm can be consolidated into real projects, whether overseas developers will remain in the Chinese ecosystem long-term, and whether large model companies can convert developer activities into stable usage rates, all remain to be observed. Communities can facilitate encounters but cannot replace business cycles; cities can provide scenarios but cannot guarantee product success.

Nevertheless, mu Shanghai at least clarifies one trend: AI entrepreneurship is shifting from "model worship" to "scenario competition," from "creating tools" to "being seen by users," and from single-point products to comprehensive competition involving communities, supply chains, and cross-border collaboration. For ordinary entrepreneurs, the opportunities brought by AI do not mean that everyone will easily become winners, but rather that more people will be exposed earlier to the same more intense selection process.

As products become increasingly easier to produce, what becomes truly scarce is the ability to understand users, enter scenarios, build trust, and maintain connections with people. AI will continue to lower the production costs of tools but will not automatically answer "why you." In this sense, creating a product is just the first step; being needed, understood, and retained is the more challenging second half of AI entrepreneurship.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。