AI news

Young people in the AI hackathon boom: startup chances, education choices and selection pressure

By 肖思佳 王欣逸1 viewsSource: 智能涌现
Young people and the AI hackathon boom

Intelligent Emergence interviews participants, organizers and investors about funding opportunities and dropout narratives, alongside project homogeneity, contract constraints and elusive hopes of upward mobility.

Editorial correction (November 13, 2026) Typographical inconsistencies in the names of the same individuals throughout the article have been corrected. Interviewees’ views and pseudonyms have been retained.

Article illustration

In early April 2026, downstairs from Manus’s Beijing office, Hong Xiao was surrounded by a group of hackathon participants. He had appeared as a surprise guest at a hackathon called Ultra Maker, catching many people off guard.

At the time, Manus was in its most sensitive period: its acquisition had still not been finalized and, according to the report, its founder Hong Xiao had been barred from leaving the country a week earlier. Yet he still made a low-key appearance in a black polo shirt. For Xiao, a hackathon in the past may have changed the course of his life, making it hard for him not to come and show his support.

In 2016, short of funds and meeting rejection everywhere, Hong Xiao was on the verge of giving up on entrepreneurship. A turning point came when he attended a hackathon near Beijing’s Yonghe Temple: after hearing Xiao’s presentation from the audience, Yuan Liu of ZhenFund invested CNY 1 million in him the very next day.

That story perfectly fits the original purpose of hackathons and the “hacker ethos”: hackers who challenge authority believe that a person’s value lies in their ability and creativity, rather than being judged solely by educational qualifications, status and titles.

Today, that narrative has been reactivated in the AI era and amplified at extraordinary speed.

After Claude Code’s release in May 2025, the barrier to programming fell to its lowest point in history. An ordinary person who previously could not code could now use AI to build a product in a few dozen hours. From that moment, hackathons were no longer just competitions for programmers. They also became an entry point for ordinary people to showcase themselves, find opportunities and even change their destinies.

Lower barriers to AI technology have brought hackathons to places across China, with the boom rapidly reaching a peak in 2026. According to data from the community AI Saitong, nearly 500 hackathons were held across China in the first half of 2026, an average of two or three a day.

In an era of constrained possibilities, hackathons offer a deeply tempting narrative: regardless of where you come from, your background or your resources, as long as you have ideas and creativity, you still have a chance to stand out here, start a business, become wealthy and turn into the next “Hong Xiao.”

The democratization of technology brought by AI has also led some young people to believe that learning and building products today need not involve going to university. They can create things directly and prove themselves through projects, rather than relying on the traditional route of selection by educational credentials. They are reshaping their life plans accordingly. Some devote most of their energy to hackathons; others take the more radical step of dropping out, even declaring that “universities are heading toward extinction.”

But have hackathons really given young people with dreams a more effective route to upward mobility? As hackathons become the entry point to this experiment in AI entrepreneurship, what can the young people flocking to them find here, and where will they be taken?

Young people decide to take a chance

After attending one hackathon after another over the past two years, Donglai Sun has spent less and less time on campus, gradually entering a state of “perpetually skipping classes.” The head of studies often summons him for “tea,” and each meeting turns into a verbal sparring match. But he believes he has more important things to do. “It’s hard to imagine someone with 10 companies on their hands still having to go back and retake a course.” He burst out laughing after saying this.

Born in 2001, this young man has already begun his second venture, having once achieved a “big result” through a hackathon. When we met Sun this summer, he had just squeezed meetings with four or five investors into a single day. Everything that subsequently happened to him can be traced back to that hackathon two years earlier.

In 2024, when Sun, then an incoming graduate student at the University of Science and Technology of China, noticed a hackathon called Adventure X, registration had already closed. Nevertheless, he emailed the organizers, pleading for one extra place. “I don’t need any expenses reimbursed. Just let me go.” He could not afford a hotel and at one point was even prepared to sleep in the lobby.

To his surprise, quite a few people really were sleeping at the venue. With too many people registered, the hotels booked by the organizers were insufficient. Many participants pitched tents or slept on the floor at the venue, while others slept directly on chairs or in corridors. Discussions continued day and night in an atmosphere that was both fervent and chaotic.

The arrival of Adventure X, investor Yue Lin said, brought “hackathons, young people and the venture capital community—those three things—together so closely for the first time.”

It seemed to offer a miniature portrait of the AI boom: after years of quiet, the venture capital community had come back to life. Stories of wealth even more extreme than those of the mobile internet era floated everywhere, alongside an anxiety about “missing out.” These emotions all drew young people almost involuntarily into the action.

This hackathon did indeed bring Sun a little closer to wealth. At the event, he formed a four-person team whose AI medical-visit companion app won prizes in three tracks. Afterward, an investor from a fund brought along the person in charge for a meeting specifically about his project. It all happened so quickly.

Sun’s AI startup journey began before he had even had time to take it in. At that hackathon, he met Zhidong Pan, who would become his co-founder. In October 2024, the two founded Dreamoo, an AI app for dream-based socializing and dream recording. In 2026, through a hackathon organized by Xiaohongshu, Sun developed a silent mask called “Mozao.” It became a somewhat well-known project in the community and quickly secured several million yuan in funding from individual investors.

“Rather than sitting in a classroom, it’s better to get out there and take a chance early.” Over the past two years, more and more university students have come to think this way. In their eyes, hackathons have become an ideal place to look for opportunities.

Zhongxu Wang, 26, once had no interest in AI at all. A master’s graduate of China University of Mining and Technology, he had always planned to find work at a mine after graduation. Even if he did not get rich, he could live a stable life. “Within the industry, I’d count as a direct disciple of the school, at any rate.”

Reality, however, was harsh. As graduation approached, he interviewed with a leading company in the industry and asked for a salary of CNY 8,000. The company thought this was too high and hired another fresh graduate for CNY 6,000 instead. During the autumn recruitment season, he sent out around a hundred résumés and obtained 20 or 30 interviews. In the end, he received only one offer, and that was just an internship with no guarantee of conversion to a permanent role.

Joe, a university student majoring in marketing, was disappointed with university education as a whole. “The courses at school have fallen behind the times. Teachers use marketing cases from years ago, and the theories in the textbooks are still stuck in the 1950s.” Early this year, Joe began teaching himself vibe coding. Three months later, he was working as an outsourced contractor, customizing and developing agents for businesses. To him, the useful skills were those that could not be learned in the classroom but could bring him real experience and income.

The disconnect between university curricula and industry needs, along with the difficulty fresh graduates face in finding jobs, has already been a much-discussed social issue in recent years. AI has magnified these tensions once again.

As traditional routes through education and into employment become less secure, the alternative possibility promised by hackathons has begun to attract young people disappointed with the established path. “A hackathon is another form of university,” said Wenjie Chen, one of the founders of the community brand AttraX. “You find like-minded people there, much like classmates. You learn and create rapidly in a very short time, and ultimately achieve a result.”

As the only adult on the team, Wang volunteered to help organize the event.

He broke down the elements of running a hackathon: guests, corporate sponsors and a venue, each corresponding to connections, resources and opportunities. He did not know exactly what he might gain from it. But rather than continuing to compete for a job step by step along an overcrowded job-hunting path, he would rather throw himself into it and try. “There’s no harm in it anyway, right?”

They named their hackathon brand MoonStone. “Stone represents ordinary people, so that ordinary people can come [to hackathons] too.”

Students at elite universities and middle schools begin to turn away from university

“Everyone has a project, whether it is any good or not.” That is how Ruohan, a physics student at Columbia University, described Silicon Valley’s current situation to 36Kr. In Silicon Valley, everything points toward the flourishing activity brought by AI: AI billboards are everywhere, people’s attention is continually drawn to the next generation of models, and cafés around campus are packed with student entrepreneurs discussing AI.

Last May, Yilin, a Columbia junior, and Ruohan, a sophomore, chose to take leave from university and devote themselves to an AI startup. Unable to understand the obscure explanations in their physics classes, they had begun using AI to study, only to find that its explanations were clearer and easier to follow. So the two developed the AI learning app Hyperknow and received a US$1 million investment from ZhenFund.

“If you force a genius to finish primary school, middle school and high school, doing practice problems and taking exams every day, their genius may be worn away.” In August 2026, at the MoonStone hackathon in Beijing’s Zhongguancun, Yilin, an invited guest, shared his ideas with the university and middle-school students before him. He believes AI can compress what originally required 15 years of education into three, thereby “changing the global education system.”

Although this idea sounds highly radical, we heard it frequently in conversations with hackathon participants. “Students don’t need to spend so many years studying academic subjects. The public education system has long been disconnected from society.” Ming Chen, a graduate student in Tsinghua’s physics department, offered a similar assessment. What surprised us was that large numbers of students at elite universities were also disappointed with traditional education.

In 2025, he had still been intent on a career in academic research. But AI’s rapid development shook his faith in academia: in his view, AI’s ability to write papers had already reached, or even surpassed, the level of doctoral students. “The human elite have absolutely no idea how to react to AI.”

This August, Chen, who had already embarked on an AI venture, sat across from us, opened a PowerPoint presentation and passionately walked us through his product ideas and grand vision, page by page. Previously, at Xiaohongshu’s first hackathon, he had turned his idea into a product demo. Afterward, an investor from Meituan Longzhu contacted him proactively, and some other investment firms sent letters of intent.

Keenly aware that the window for starting a business is brief, he was considering whether he should drop out. “Life at school has become meaningless to me. If my studies and my startup conflict in the future, I will give up my studies.”

A year ago, the idea that “the people truly building the future don’t go to university” was already beginning to gain popularity among Silicon Valley’s younger generation. In June 2025, the well-known US outlet Business Inside [as written in the source] interviewed a Stanford student in an article titled “Inside Silicon Valley’s Anti-University Movement”: “The opportunity cost of going to university is too high. The tech industry changes in the blink of an eye these days.” Such thinking immediately received an enthusiastic response from Silicon Valley’s tech giants. In a March 2025 podcast, investors at the US venture capital firm A16z declared that this was “the best time in a decade for dropouts and recent graduates to start companies.”

Amid this vast wave of change, everything in front of these young people made them feel they could not afford to miss out. In the United States, “Over the past couple of years, everyone’s LinkedIn résumé was uniformly ‘interning at XX investment bank.’ This year, more and more classmates have changed their titles to ‘XX founder,’” Ruohan said.

As the ages of hackathon participants keep falling, narratives about education being useless and dropping out to start a business are spreading beyond university campuses to middle- and high-school students.

In 2025, Tudou attended his first hackathon at just 14 years old, although at the time he was only a spectator. He saw a participant write 5,000 lines of code with AI in a single night. Although he was not sure what the person had written, Tudou, who had always liked programming, thought “it was amazing.”

After that hackathon, Tudou had a goal and a dream. He began teaching himself programming through the internet and AI, sometimes studying for eight hours straight in a day. In January 2026, he dropped out of school. “I feel that going to school isn’t much use to me. It isn’t important anymore,” he said.

When we met Tudou, he had been out of school for eight months and was one of the co-initiators of the hackathon brand MoonStone. The other three middle- and high-school students on the team had all chosen to take leave from school. When discussing today’s high-pressure school life, they generally expressed feelings of suffocation and dissatisfaction. Team member Icy said: “There’s a small test every day and a major exam every week at school. Afterward, we’re required to write reflections, whether we did well or badly.”

Another 14-year-old boy at the event had been mocked as “competition-grade Jiahao” for sharing with classmates at school how to connect WSL to OpenClaw. He said he could find friends only at hackathons. In everyday life, he taught himself technical knowledge through the internet and AI, and believed school classes were no longer necessary: “Even Tsinghua’s courses are on Bilibili.”

In the eyes of these boys who “don’t like school,” a hackathon is like a utopia. Discipline and reproach are absent, replaced by recognition and acceptance.

After dropping out, Tudou quickly became a well-known figure in hackathon circles through frequent participation. A media outlet even followed and filmed him for a week. Sometimes, as a “15-year-old entrepreneur,” he is invited to AI exchange events for middle- and high-school students to give a talk titled “You Don’t Have to Be a ‘Genius’ to Take the First Step.”

These signals from the outside world made the teenager increasingly certain that dropping out had been right. At school, his grades had been at the bottom, but the label “middle-school dropout” successfully attracted investors’ attention. This led him to believe, “If I start a business directly, my chances of success will be higher than if I go to school.”

As more such cases emerge, leaving school to start a business is gradually becoming a trend. Some students who are still in school have already put “high-school dropout” in their introductions. More people express envy of Tudou’s experience. His partner Sherry told us: “Among the underage participants who are enthusiastic about hackathons, 60–70% have already begun thinking that they might as well take leave from school early to start a business or look for work.”

Who is scouting the “young prodigies”?

Early this year, at a hackathon themed the “Zhongguancun Beiwei Lobster Competition,” Donglai Sun and Muran Jiang, a 14-year-old eighth-grader, chose the same creative brief. Jiang won the competition’s top award and CNY 200,000 in prize money.

Half a month later, Jiang and several other middle-school students won prizes again at Xiaohongshu’s hackathon. Overnight, Jiang became someone media outlets and investors competed to cover and meet. One investor told us that she had tried several times to arrange a meeting with him, without success: after he became famous, too many people wanted to meet him.

This left Sun anxious about his age for a time. “Although my [work] was more fully developed than his, he’s 10 years younger than me,” he said. “I later wondered: if I’d encountered entrepreneurship earlier, would there have been more advantages?” By comparison, Sun seemed to have become one of the disappointed players in this race to find young geniuses.

ZhenFund once divided founders into four types: young prodigies, seasoned veterans, operators (company executives) and technical specialists. As more and more middle- and high-school students enter the public eye through hackathons, young prodigies and investors alike seem to have entered the best of times.

Xin Wang, a 17-year-old who had dropped out two years earlier, showed us an “incubation group” on WeChat that had been set up around him. It included investors from 5Y Capital and Lisi, as well as fundraising advisers from financial-advisory (FA) firms. Everyone was waiting for a chance to place a bet.

Wang fits today’s investment firms’ image of a “young genius” remarkably well. He won a prize in an informatics competition in seventh grade, and after two months of teaching himself Olympiad mathematics, placed first in his city in a preliminary round. Unwilling to waste time on academic subjects at school, he dropped out of middle school and subsequently worked as an intern at AI startups and large technology companies: young enough, smart enough and unconventional enough.

The person who brought this group chat together for him was Kehan Dong, someone who builds networks exclusively for “young prodigies.”

Dong was a founding team member of MiraclePlus, formerly YC China. Over the past two years, he has repeatedly stated publicly that he focuses on discovering young builders aged 14–22, and has established N1. Its official description is a platform for discovering and supporting young creators. Put more plainly, it is a startup incubator that specifically invests in young people at an early stage.

According to public information, N1 currently has around 40 members, with an average age of just 19. Guangyu Chen, a name known to virtually everyone in AI circles, came out of N1. After leaving N1, this 17-year-old “young prodigy” joined Kimi as an intern and became one of the co-first authors of the key paper “Attention Residuals.” A year later, the report says, he announced that he was leaving Kimi to join a company founded by Jeff Dean, described in the article as Google’s former chief scientist.

In this decidedly atypical high-school student’s coming-of-age story, Dong played the role of talent scout. He discovered Chen in 2025 at a hackathon for middle- and high-school students held at Beijing’s Tanyue School, brought him into N1, taught him how to expand his connections in the community and helped him look for internships in Silicon Valley. After that, Chen realized he could never return to his former school life.

That is how Chen became Dong’s “signature achievement.” And through hackathons, Dong is reaching more and more teenagers.

“Basically every one of us has received an invitation from Kehan Dong’s team,” said Icy, who was not yet 16. According to the article, a prerequisite for entering N1 is that candidates make a clear choice between their studies and entrepreneurship. “He particularly likes people who have dropped out. As long as you take his money, he will relentlessly urge you to drop out,” Ming Chen told us.

In September 2026, building on N1, Dong launched N1 AI School for teenagers with potential who had not yet started a business, accepting applicants as young as 14.

We learned that although admission to N1 AI School is free, entrants must sign a SAFE agreement (Simple Agreement for Future Equity) with Dong. When those admitted start a business in the future, Dong can invest no more than CNY 1.45 million in return for up to 16% of the equity. The article describes this as a very favorable investment for him. By comparison, MiraclePlus typically exchanges US$300,000 (approximately CNY 2 million) for a 7% equity stake.

“Put in a little money and make a few introductions. As long as one of these people starts a business and raises another round or two, he could make back 10 times his money in two or three months.” The investor quoted above said that, by a conservative market estimate, this cohort of “young geniuses” could raise US$30 million right at the outset. “It amounts to monopolizing the mindshare of this group of 16- to 22-year-olds. As long as they think of Kehan Dong when they start a business, he has won.”

Some investors call this an “unequal treaty,” saying the extent to which the agreement binds signatories and constrains their fundraising is “unheard of.” For example, if a founder goes on to start a second company, Dong can still obtain an additional 4% equity stake by investing CNY 200,000.

The students, however, do not necessarily see it that way. We found an eleventh-grade boy who had taken leave from school. He began attending hackathons this January, wanted to start a company less than six months later and was intensely eager to enter N1 AI School. Another boy who had already “enrolled” said that for a child from an ordinary family like his, receiving this kind of support with startup resources was “really good.” Some even compare N1 to a new-era “Academy of Athens,” seeing it as another way of imagining education in the AI era.

From fervor to mediocrity

In late 2025, Kai Zhang, then a sophomore, attended his first hackathon. Hackathons had not yet become as ubiquitous as they are now, but he already noticed something unusual: while he was buried in writing code and building a product, many participants around him were busy talking to investors. Only after the event did he discover that these participants had not followed the hackathon rule of devising a product from scratch. Instead, they had brought mature projects that were already implemented to compete and seek investment.

As the number of hackathons surged in 2026, this phenomenon became increasingly common.

“Hackathons have nearly run their course,” Sun lamented in August. After Adventure X ignited his enthusiasm for entrepreneurship in 2024, he created his own brand, Huikesong, in Hefei. But he has decided to end the brand after holding one final event at the end of the year.

“You find that no matter how many events you hold, it’s the same bunch of people competing. The products are disposable after a day, or a month if they’re a bit better,” Sun said. “And there are quite a few bounty hunters who enter just to earn prize money.”

In fact, in recent months, we have heard similar complaints at more and more hackathons: although hackathons are reaching more people, the level of innovation has not risen as imagined; instead, it has been diluted—homogeneous projects such as AI social apps, AI bookkeeping and AI pets appear again and again. “Basically every event has someone making a Pomodoro timer,” said an organizer who has held multiple hackathons.

The “democratization of technology” has turned hackathons into a mass activity. Ironically, this has also become the very reason they are drifting toward mediocrity.

Unlike Dong, who wants to build a systematic “methodology for discovering and cultivating young prodigies” through hackathons and N1, most investors go to hackathons in search of relatively mature future startup stars. When good products become hard to find and the concentration of talent is diluted as well, their enthusiasm cools accordingly.

In 2025, investor Yue Lin sat at Adventure X’s opening ceremony and heard a long succession of investment firms named as sponsors. “There were too many names even to list.” At the same opening ceremony in 2026, however, he heard the names of only two or three investment firms. Fewer and fewer investors like him are still willing to spend time sifting through competitions for promising people. “The people who used to do hackathons really had skills. The people doing them now don’t have skills,” said an investor who was once active at hackathons, even sounding somewhat nostalgic for the old days.

Can the new path represented by hackathons really give the majority of ordinary people a story of rising against the odds? We asked some investors about the probability that hackathon projects ultimately secure financing. Their general answer was: extremely low.

“Definitely less than 1%,” Lin said, offering a rough estimate. Over the past three years, his firm has invested in around a hundred projects, fewer than five of which originated at hackathons. Yet his fund is already among the investment firms most active in discovering talent through hackathons.

Now, as hackathon fervor rapidly fades, investors have returned to waiting for top researchers with big-tech or elite-university credentials. Wherever they look for entrepreneurial talent, what investors want has always been people who are “good enough.”

Although hackathons give all participants the same starting line, investors’ underlying selection logic has not actually changed. “Early-stage investing still depends on the person, but hackathon registration doesn’t screen the person,” said entrepreneur Haitian Yuan. “So even when many hackathon projects have been built, they still can’t actually receive investment, because the person isn’t good enough.”

The old evaluation system may be loosening, but selection itself has not disappeared. Hackathons, seemingly free and rebellious, have not truly escaped the influence of traditional education either. People still have to prove their worth against some standard; the only difference is to whom they must prove it.

Venture capital firms favor people who are “special,” but that preference has its preconditions. Lin gave an example: of the roughly 20 projects he judged at this year’s Adventure X, three seemed good to him. On following up with one project, he discovered that the three team members had not only been high-school classmates but had all gained admission to excellent universities, including Tsinghua’s Yao Class. “The more you dig, the more you find.” Lin quickly established contact with them.

At the same time, he has found more and more people beginning to perform being special: “For example, deliberately making some strange decision, like dropping out of middle school.” This kind of withdrawal earns no extra points with Lin. Dropping out of middle school means not having passed the traditional system’s selection process, while giving it up carries little cost.

Confusion and questions are emerging too. At the end of August, a thousand-person hackathon in Beijing that billed itself as “Asia’s largest” left many judges and participants confused about the process because of its excessive scale and insufficiently clear mechanisms. Numerous participants reported that they had struggled for a long time to find judges to score their projects. And when the awards were announced, a mature project from a sponsor had actually won first prize.

In the participants’ group chat, challenges and questions continued for a full day and night. Some complained on social media: “So all you need to do is rent a venue, get a few sponsors and set a few themed tracks, and you can run a thousand-person hackathon.” “Make so many people travel from all over the country to add luster to the organizing team’s résumés.” Holding a large hackathon can strengthen an organizer’s résumé when applying to elite universities overseas. That was also Adventure X founder Haoyu Zhu’s original motivation for holding hackathons. Today, Zhu has gone to study in the United States and has raised US$4.5 million for his startup.

Looking back at the atmosphere of a hackathon—the countdown on the screen, the impassioned speeches, the applause from the audience—it feels as though being there envelops you in a beautiful illusion. But this call to everyone ultimately still opens the door only a crack, for a small number of winners.

Tudou still often thinks about everything connected with hackathons. Sometimes, late at night, he posts previews on WeChat Moments, declaring that he will “do something absolutely huge” at Adventure X in 2027.

Since dropping out, he has supported himself by building websites for others or modifying phones, sometimes earning only a little over CNY 1,000 a month. So far, the most hopeful moment of his life, he feels, was when a startup idea occurred to him on the subway: bring Vision Pro’s way of interacting to ordinary computers, eliminating the mouse and keyboard. In the carriage, with people coming and going, one glittering fantasy after another surfaced. He saw himself lying on a big bed in a villa, exercising in a private gym… The subway kept moving faster, as if it would soon carry him toward a future within reach.

(Ming Chen, Xin Wang, Kai Zhang and Yue Lin are pseudonyms.)

This article comes from “Intelligent Emergence,” by Sijia Xiao and Xinyi Wang.

Related topics

After reading

Join the AINRK community

Read how to join the community and connect with people working in AI.

Community information (opens in a new tab)