On September 16, TechNode joined Asia Capital Exchange (ACE), Lighthouse Capital and BEYOND in Beijing to host “ACE Gathering x Light Spot: The First Year of AI Startups,” a closed-door event.
The roundtable was moderated by Qiao Jianlong, editor of AI Insider’s Demo Club. Joining him were Zheng Jinliang, co-founder of BASAL INTELLIGENCE (本溯智能); Song Haocong, founder and CEO of Aurano; Zhou Junting, CEO of Lingqi Xiyuan (灵启犀元); and Li Zhengwei, managing director at Lighthouse Capital and co-head of its 3i incubation business.
Qiao recalled introducing a former classmate with a strong background to an investor friend. The investor’s private response was blunt: startups that come knocking are often just there to make up the numbers.
Yet another observation from the same discussion seemed to point in the opposite direction. In any funding round, Li said, the investors most likely to commit are often those who approached the startup first.
The tension between those two remarks gets at a difficult question for founders in their first year: What does it take to make investors, talented people and customers seek you out?
This article is based on remarks made during the roundtable.
When Fundraising, Chasing Investors Can Work Against You
Li has seen fundraising from several sides. He joined Lighthouse Capital in 2016 and spent six years advising startups on financing. He later left to start a business, incubate companies and make personal investments before returning to Lighthouse Capital earlier this year to work on early-stage incubation and financial advisory services for growing technology companies.
Founders often ask him how they can raise a large amount of money quickly. But today’s funding market is sharply divided: some startups have investors lining up, while others struggle to attract any interest.
Drawing on his experience, Li said investors who reach out to a company themselves may be at least twice as likely to invest as those introduced by a third party. An introduction backed by someone the investor trusts may, in turn, be at least twice as effective as a founder’s own unsolicited approach. For an individual founder, contacting large numbers of investors directly is often the least productive route.
His advice is to find someone influential who understands the company and is willing to vouch for it. That person can make thoughtful introductions to three to five investors whose support would carry weight. Founders without access to such a person might instead work with a capable financial adviser.
Trust also helps explain why a startup’s first round often comes from friends and family. But it cannot carry a company indefinitely. Qiao suggested that, during the first six months, founders may be drawing on the credibility and relationships they built before starting the business. If they still have no convincing product, or at least no meaningful progress to show, after that period, raising the next round becomes harder. Even a highly influential supporter can put their reputation on the line only so many times.
For Li, the order matters. Founders need to know what they are building and who can build it with them. When those questions have credible answers, fundraising becomes easier. Reversing the order is one reason the process can feel so painful.
Look Beyond the Obvious Direction
Before a startup can raise money, it has to decide where to go. A strong introduction will not rescue a weak premise.
Aurano founder Song Haocong came to that question after working at Spark Education, which reached the pre-IPO stage, and at foundation model company 01.AI. Since the arrival of GPT-3.5 in 2022, he has watched more human-computer interaction move into a text box. Whether a user types or speaks, the input often becomes text; the model then responds with another long stretch of text. It can be tiring to say, write and read.
Song wanted to find a more natural way to interact with AI at the application layer. Rather than follow the most visible trend, he uses what he calls a “three-step” approach to thinking through a product opportunity.
If a founder looks only at what today’s technology can do and builds the most obvious product, many others will arrive at the same idea. Looking one step further reduces the number of competitors. Looking three steps ahead requires more assumptions to be tested, but it can reveal a product that seems small today and may become much more important later.
“From day one, I look for a direction that is too small for others to care about today, but could grow as the conditions around it change,” Song said.
Zheng Jinliang, co-founder of BASAL INTELLIGENCE, is approaching the question from embodied AI. His company focuses on the “brain” of an embodied system rather than the robot hardware itself. It is particularly interested in on-device self-improvement and whether in-context learning could open up a new technical approach.
Zheng is a doctoral student at Tsinghua University. After three to four years of embodied AI research at the university’s Institute for AI Industry Research, he decided to start a company in July. He sees the hardware side of robotics as crowded, while the intelligence behind it still leaves room for more fundamental innovation.
He believes the field’s prevailing approach may be nearing a ceiling. Instead of beginning with better robots and more data, his team asks what problem a model should solve. From there, it works backward to the form of data needed and then to the hardware that could produce it.
Zhou Junting has chosen a third route: AI for Science. An undergraduate at Peking University’s Yuanpei College, he wants Lingqi Xiyuan to build infrastructure for an AI lab that can improve through its own work.
His proposed system has two layers. A “Science Agent” would help move from a research hypothesis to an experimental protocol. A “Physical Agent” would carry out that protocol in the real world. By interacting repeatedly with laboratories, the system could accumulate specialized knowledge and improve both agents.
Zhou sees this as a step beyond AI copilots, agents and the more recent idea of the AI scientist. If AI is to contribute to scientific discovery, he argues, it must enter the full research loop, including physical experiments and the testing of results.
The three companies are pursuing different opportunities: a new form of human-computer interaction, a different foundation for embodied AI, and a link between computational work and physical experiments. What they share is a reluctance to compete solely by improving what everyone else is already building.
Hire People With Conviction and Independent Judgment
Once a direction is set, founders need people who can pursue it. The three entrepreneurs described different ways of finding them.
Aurano initially needs engineers who are comfortable exploring uncertain problems; it may bring in more algorithm researchers later. Song looks for conviction. With new models and industry headlines arriving every week, he said, people who change course with every development can lose sight of what the team is trying to build.
He also values a research mindset. Aurano keeps textbooks close at hand. When the team runs out of ideas, its members return to books and papers for a deeper way into the problem.
Zhou assembled his team before seeking outside funding. For a student founder, a campus can be a good place to find people who already believe in the same goal. That shared belief, he said, may bring more commitment than a hurried hire from the open market.
He is drawn to people with a distinctive way of thinking. Someone might be quiet but deeply serious about a technical problem, or see an angle that others miss. “I don’t want a team made up only of people who are excellent in a generic way,” he said. High grades matter less to him than an ability to form an independent judgment.
Zheng’s first seven team members had already worked together in a laboratory on embodied AI research. They knew one another’s strengths and how to divide the work. When they meet potential recruits now, they rely on their technical ideas and research results to attract them. People who have spent time at the frontier of embodied AI, he said, can tell whether a team understands the hard questions.
Li framed the hiring question from an investor’s perspective. In the internet and mobile internet eras, founders were often judged chiefly on their understanding of user demand and their ability to execute. Those qualities still matter, but he now pays particular attention to technical judgment and business sense.
The combination is crucial. A team must be able to turn a research result into a product and a company. Otherwise, it risks building an impressive laboratory that never becomes a business.
Find the Opportunity Big Companies Overlook
Competition with large technology companies is difficult to avoid in a startup’s first year. For founders building applications, the immediate threat may come from foundation model companies. Those companies already interact with users, and each improvement to their models can absorb features that once supported a standalone product.
Song said investors ask him about this frequently. His answer goes back to choosing a direction that looks too small to justify a large company’s attention today, while showing signs of future demand. Big companies often need a quicker or larger return to approve a project internally.
“I need to find an opportunity where I can build what looks like a lightweight product now,” he said, “but one that can later connect to capabilities many other companies are racing to develop. Others may already be building those capabilities. The entry point is what remains open.”
Zhou faces a similar question in AI for Science, where companies such as Google DeepMind are also active. He does not assume they must be rivals. Better foundation models, more accessible interfaces and a market that understands the technology could all help a smaller company. His team’s immediate task is to make the loop between computational work and real experiments function inside a laboratory.
Zheng sees embodied AI as too large a field for any one company to address in full. Large firms can assign departments to individual parts of a long supply chain. The question, in his view, is whether the field has defined those parts and their underlying problems correctly.
Since 2023, embodied AI has advanced in data collection, hardware and model training, he said. But the basic approach to learning has not changed as much. A small team cannot match a large company’s resources; it may, however, be better placed to test a new approach in a compact, complete system.
Make Others Want to Join You
What would count as getting through the first year?
After reviewing many startups, Li said the companies that falter almost immediately often lack a committed core team. Some founders chase whichever area is attracting attention. Others start a company because colleagues have done so, or because they see a chance to turn an existing resource into quick money. A team that genuinely believes in a difficult, unfashionable direction can be more resilient.
The risks become more varied on the way to a Series A round. A technical shift can invalidate an early assumption. The expected customer need may not exist. Key people may leave, funding may run out, or the product may never be adequately tested.
Lighthouse Capital’s 3i incubator has set a goal of seeing half of its projects reach Series A within two years. Li acknowledged that this may be higher than the success rate many early-stage investors actually achieve. To him, a Series A round can indicate that a company has tested whether its product meets a real market need. Few startups go all the way from identifying a direction to building a product and raising that round without setbacks.
Some mobile internet startups had an easier starting point: their products had already been developed and tested inside a larger company before being spun out. A team starting from zero has no such advantage. Li suggested that founders might do better to value the process itself rather than count only on a particular outcome.
Qiao closed on a more optimistic note. The mobile internet era often followed a winner-takes-all pattern, but AI and hard technology have more links in the chain, more applications and more specialized niches. No company, including a foundation model provider, can do everything.
In the first year of an AI startup, the scarcest resource may not be money or model capability, but the ability to make others believe in what the team can do. The aim is to have investors seek you out, talented people want to join, users actively choose your product, and opportunities emerge—while big companies have yet to take notice of your niche.
That is both a survival rule before a Series A and the niche for AI and robotics startups in their first year.