An FA Firm Decides to Start Incubating | Entering the Game

"Time Traveler's Light."

"Entering the Game" is a recurring column from Dark Waves (暗涌Waves). It grew out of our observation that once-reliable operating models face new challenges, and the industry rules inherited from West to East have been shattered. People urgently need new maps and new orders for innovation and capital. "Entering the game" is the most precious posture one can adopt.

"Entering the Game" was born amid transformation. To summarize this column's subject in one sentence: we hope to find new players and new playbooks better adapted to a changing environment. This is the thirteenth article in the series.

By Zhiyan Chen

Today, Lighthouse Capital announced that its Lighthouse Founders' Fund (L2F) has completed an oversubscribed first close, with a final target of no less than $50 million, and has officially entered its investment period.

According to Zhile Zheng, founder and CEO of Lighthouse Capital, the fund focuses on "ultra-early stage" investment opportunities in AI and frontier technology. It will invest in seed and angel-round companies in dual currency and provide ongoing incubation support. "The plan is to make ten investments per year, each ranging from $300,000 to $2 million."

Whether judged by the efficiency and success rate of ultra-early stage investing, or by China's current lack of truly sustained and successful incubation cases, incubation seems like a "bad trade" for a top-tier financial advisory like Lighthouse Capital. But clearly, Zheng's thinking runs deeper.

In his view, AI's current development stage parallels the mobile internet era of 2009–2014: non-consensus far exceeds consensus, and abundant opportunities emerge in the early phase of this industry cycle. The primary market is fundamentally about matching tools to vintage — and incubation is the "best tool" Lighthouse has chosen to enter the market after clarifying the industry's vintage.

Currently, L2F has 15 projects in its pipeline and 5 invested projects (including planned investments). Its first investment has gone to AI mineral exploration pioneer Lingyun Intelligent Mining (凌云智矿) — led by L2F, with the project's earliest development receiving full support from Lighthouse's 3i Industrial Innovation Incubator team. Built on founder Dr. Xuance Wang's development of the world's first intelligent mineral exploration AI decision system, the company is dedicated to deeply integrating AI with geological exploration to improve productivity across the industry and reshape mineral resource survey and development models. Lighthouse conducted dozens of exchanges with university professors, senior industry experts, and leading industrial enterprises in the project's early stages, comprehensively validating the feasibility of AI mineral exploration technology and downstream demand, and assembled a cross-disciplinary entrepreneurial team (geological theory, mineral exploration, AI, capital operations, operations management) to successfully incubate Lingyun Intelligent Mining. In subsequent incubation stages, Lighthouse provided not only funding but also assistance connecting mining resources, technical experts, and local governments to accelerate product validation from lab to mine.

Zheng told Dark Waves that L2F's LP base mainly comprises: (1) successful internet and technology entrepreneurs, (2) AI industry chain enterprises and entrepreneurs, and (3) globally-minded entrepreneurs and their families with international resources. The fund has also established an advisory committee, with most entrepreneur LPs joining to provide guidance and support for portfolio founders.

"The fund's English name is L2F, Lighthouse Founders' Fund — on one hand, we want to build a Founders' Fund in the AI space," Zheng said. "L2F also means 'leap to future.' We believe the future will be an era of greatly enriched productivity, vastly improved well-being, where humans live longer and better lives. AI should create such an era, and Lighthouse should be a driving force toward it."

At Lighthouse's office in Shanghai's Lujiazui, Zheng spoke with us about why a ten-year-old FA would enter incubation and ultra-early stage investing, and his understanding of AI, cycles, and entrepreneurship.

The following conversation has been edited by Dark Waves:

Part 01

The Optimal Entry Point for AI + Incubation

Dark Waves: "AI + incubation" is starting to appear sporadically across the industry. What's driving this?

Zhile Zheng: AI is in the early phase of massive industry transformation, with abundant new opportunities emerging amid widespread non-consensus. This closely resembles the mobile internet era of 2009–2014, even 2009–2011 — "a hundred empty lots awaiting development." The primary market is fundamentally about matching tools to vintage, and incubation is the tool we believe best fits the industry's current development stage.

Moreover, AI's era differs from mobile internet's in one crucial and valuable way: mobile internet was fundamentally about connection — creating value by connecting everything, connecting all things. AI, by contrast, is productivity itself. AI is "electricity." When channeled into software and hardware, it forms end-to-end solutions that create value.

AI as productivity has two attributes. First, full industrialization: AI can act upon thousands of industries, which is why industry players increasingly recognize AI's enormous potential and rapidly growing demand. Second, globalization: productivity knows no national borders. Cross-border replication of internet products often faces challenges because transferring production relationships involves cultural, national, and other constraints — but productivity does not. British machinery works in China; American electrification experience applies to China. Similarly, as general-purpose productivity, AI gives Chinese AI companies global characteristics from day one.

Another important factor is the influx of new people and capital. Currently, and for years to come, an exceptionally strong cohort of entrepreneurs will emerge — increasingly talented people leaving major tech companies and large model companies at the -1, -2 levels. Chinese technology assets are being revalued, and more new, quality capital will enter.

Dark Waves: You're saying that doing incubation in a technology wave's early phase offers the greatest opportunity?

Zhile Zheng: Exactly. Beyond the points above, there's another crucial insight: we need to do early-stage things early, before market liquidity improves.

Money is still scarce in the market, but frontline participants can clearly feel Chinese AI assets undergoing repricing, with more capital set to flow in. So rather than waiting, it's better to enter the game now and do things that create value and serve value creation.

Dark Waves: Incubation-stage investing heavily depends on investors' judgment of industries and projects. How has your understanding of AI evolved?

Zhile Zheng: Lighthouse began tracking this AI wave in 2022. In the second half of 2022, we engaged with and served AI project DP Technology. Its essence was AI learning mathematics, physics, and chemistry to assist scientific research. Starting from this "AI for Science" point, we came to recognize that this AI wave was essentially about creating people: AI becomes a new producer, learning whatever it can predict. We can understand this "person" as a child. What should you teach a child first? Language — once they master language, they have the foundation to learn every subject. This was the origin of large language models, and their subsequent flourishing further validated our judgment. Following this logic, we wondered: if we teach AI physics and art, could it make videos?

Our bet on video stemmed from two convictions. First, we firmly believe video is a market with enormous potential, very close to application and monetization. Second, while the market generally believed China lagged behind the US in AI exploration, in video I believed Chinese people would definitely outperform Americans — the success of ByteDance and Kuaishou speaks for itself. The "AI + video" concept emerged from this, and today Aishi Technology's development proves our judgment was correct. Latest data shows Aishi Technology's video generation product Pixverse has reached #4 on the overall US App Store charts and #1 among video apps, with over 16 million global MAU and more than 60 million global users.

Dark Waves: How deeply was Lighthouse involved with Aishi Technology?

Zhile Zheng: In the Aishi Technology case, Lighthouse was deeply involved from the initial business direction and strategy discussions, through introducing various types of capital and resources, to team building. Initially, a Lighthouse colleague wanted to start a company, and we found Changhu Wang, who had previously led video technology at ByteDance's AI Lab and had already left the company. After we discussed it together, we clicked immediately — and he was indeed the most suitable person for this. AI + video has many possible directions, but from day one our goal was to build a top-of-pyramid product. During brainstorming, we jointly rejected many directions including game asset generation, until we finally settled on AI video generation models and applications — a direction that excited all of us. To this day, we have continuously invested in Aishi Technology's seed, angel, and Series A rounds. As you can see, Lighthouse's deep involvement in incubating Aishi Technology represents the叠加 of cognition, capabilities, and resources accumulated over our past decade-plus of cycle-crossing, deep primary market engagement — from mobile internet to industrial technology to artificial intelligence. This combination itself possesses certain scarcity value within the industry, and this unique scarcity is precisely the important foundation supporting our incubation work.

Dark Waves: Large models were already very hot at that time.

Zhile Zheng: At that point, the market was still quite skeptical about the AI + video market. Many people didn't believe Chinese people could build products the US hadn't yet made. But a few months later, in February 2024, Sora launched. That very day, numerous VCs and companies came to us. You could say industry consensus only truly formed then. In fact, to this day, the teams doing top-tier AI + video remain Chinese-led.

Dark Waves: I remember we spoke eight years ago, when you said you wanted to be companies' "external partner for strategy and capital." Is there a difference between being an FA and doing incubation in this regard?

Zhile Zheng: I see it as the same original intention, the same capability structure projected onto different things. The core logic is identical, but with incubation, because the chain is longer, there's more to do alongside the company, and more thinking together with founders, it manifests more completely.

Dark Waves: FAs are typically seen as having a transaction mindset, while incubation requires a creation mindset. What's Lighthouse's fit for incubation?

Zhile Zheng: First and foremost, fundraising capability. As mentioned, money is scarce in today's market, making fundraising capability hard currency for entrepreneurship. And today's capital market is more complex than before, with more diverse investor types, making fundraising a complex decision. Most entrepreneurs need someone to help them with full-process planning.

Another aspect is industrial resources. Under the AI + N trend, entrepreneurs have enormous hunger for scenarios and industries. Lighthouse has accumulated resources from over 100 listed companies that can help AI entrepreneurs and scientists validate technology — some directly convertible to orders.

To summarize, it's three things: scarcity of cognition, scarcity of capability, and scarcity of resources.

Dark Waves: Entrepreneurs need deep industrial resources with real voice. Can an FA really deliver that?

Zhile Zheng: Let me give you two examples. Just in one week this April, I took the founder of a leading embodied intelligence company to visit a trillion-level and a hundred-billion-level industrial giant in succession, meeting directly with their controlling shareholders, and securing strategic cooperation and order commitments — with Lighthouse fully participating in the cooperation scheme design throughout.

For Lingyun Intelligent Mining, our incubation fund's earliest investment, in its initial incubation stage Lighthouse introduced top-tier first-in-command entrepreneurs from both AI and industrial circles.

I can say this: among FAs in the industry, only Lighthouse can systematically and sustainably facilitate direct dialogue between first-in-command entrepreneurs and first-in-command industry giants, forming strategic alliances between startups and industry leaders, and driving implementation through capital and business cooperation.

Dark Waves: How does Lighthouse ensure influence with industry chain leaders?

Zhile Zheng: Through genuinely covering and serving over 100 listed companies across industry chains over the past decade. This coverage isn't just knowing people — we've actually done business with these listed companies across M&A, business spin-offs, and innovative project financing.

I previously proposed the concept of "Three Wave Convergence" — the tight interconnection and interaction between globalization, intelligentization, and industrialization, where each element actively promotes the other two. Because we've genuinely engaged with, served, and cooperated with these companies, we find that the "wave leaders" of each wave are extremely eager for新锐 in their respective waves. When these "first-in-command wave leaders" want to connect today, as their channel to新锐, Lighthouse has already built strong links and trust networks on both sides, establishing pathways through our tool chain system — this is precisely our core competitiveness for systematically achieving ecological linkage.

Dark Waves: Most people still see FAs as essentially "Party C" [service providers].

Zhile Zheng: I understand what you mean. But Lighthouse's relationship with industry players goes beyond transactions.

Take Sany Group as an example. Lighthouse Capital has deeply participated in its strategic transformation through a dual-wheel drive model of "capital empowerment + industrial resource import": promoting the independent development of Sany's incubation projects, leading financing for a series of Sany ecosystem projects including Rootcloud and Yigongpin, and assisting Sany Heavy Truck in completing its nearly 1 billion RMB Series A financing, introducing industrial investors like GLP and China Merchants Capital to strengthen new energy logistics ecosystem synergy and inject innovative momentum.

Due to massive information, cognition, and resource asymmetries, Lighthouse has opportunities to deeply participate in industry players' AI or globalization upgrade and transformation processes, and jointly assess new initiatives they want to pursue, and how to realize them "in-house" or "externally."

You can understand it this way: when cooperating with industry players, Lighthouse is not only a partner for capital and strategy, but also their innovation partner.

Part 02: Super Engine Pulling a Small Cart

Dark Waves: This frontier fund completed its oversubscribed first close in two months. How was that achieved?

Zhile Zheng: During fundraising and preparation, my expectation was that this fund must be small and lean. The initial target was $30 million — my intention was to quickly launch by "pooling money" with successful entrepreneur friends around me. After discussing the fund concept with friends, they quickly oversubscribed. Considering that ongoing follow-on investments in quality portfolio companies during the investment period would require greater capital, the fund size was ultimately set at no less than $50 million. The entire fundraising process went very smoothly.

Dark Waves: Lighthouse has been running an incubator since 2024. What's the relationship between this incubation fund and the incubator?

Zhile Zheng: The incubator's core is talent discovery and intellectual support; the incubation fund deploys capital for ultra-early stage investments.

Lighthouse's incubator team comprises young people with both entrepreneurial passion and industrial insight. When entrepreneurs have only an idea, they can jointly sketch technology commercialization paths — helping entrepreneurs launch projects through deep collaboration of "strategic foresight + resource pre-configuration." This is why entrepreneurs are willing to deeply integrate with Lighthouse's incubator.

L2F is independent from the incubator yet strategically coordinated with it — another tool to better manifest our forward-looking deployment approach in the AI + N space. The incubator and L2F fund together constitute Lighthouse's "dual-wheel drive synergy engine" for ultra-early stage opportunity deployment.

Dark Waves: The incubation stage's core is judgment of people. What kind of entrepreneurs do you want to find?

Zhile Zheng: This incubation fund focuses on three core incubation directions: entrepreneurial incubation (young serial entrepreneurs or high-potential entrepreneurs from major tech companies), scientific incubation (scientists with industrial thinking and entrepreneurial spirit), and industrial incubation (industry players pursuing new initiatives in AI or globalization directions).

Dark Waves: Domestic incubation has never had truly successful cases. Will Lighthouse have new approaches?

Zhile Zheng: We've broken a company's early 0-to-1 process into three stages —

The first stage is from leads to people-mission fit. "Leads" means someone wants to start a company but doesn't know what to do yet, though we judge them as excellent; or we've identified a highly promising opportunity but haven't found the right person. At this stage, we use an "incubator direct equity" model. Annually, we encounter approximately 1,000 potential projects or entrepreneurs; after strict screening and deep assessment, roughly 10 successfully reach the people-mission fit stage.

The second stage is people-mission fit to product-market fit. For example, when we encountered Galaxy General in 2023, it was at this stage — the team was assembled, committed to embodied intelligence large models, and Lighthouse's fund and FA business would engage.

The third stage is after achieving product-market fit, where FA business and growth-stage business deeply participate.

Dark Waves: For both Aishi Technology and Galaxy General, Lighthouse invested at the people-mission fit stage. Looking back, why were you willing to pull the trigger before seeing product-market fit signals?

Zhile Zheng: First, through the incubation accompaniment process, we've established deep cognition of team capabilities and project potential. Second, relying on industrial methodology accumulated in the AI + N domain, we possess structured capability to precisely预判 whether product-market fit can be achieved. Additionally, as I mentioned, Lighthouse is one of the few industrial investment banks that has crossed cycles, with very complete past understanding of mobile internet and industries, enabling us to well judge the possibilities of AI as productivity combined with industries. Meanwhile, our沉淀 of extensive industrial resources can help us achieve multi-dimensional assessment from technical feasibility, scenario adaptability, and commercialization paths.

L2F's LP base has already gathered numerous heavyweight investors, including multiple top internet and technology entrepreneurs, leading industrialists and ecosystem giants in AI and frontier technology, and globally-minded entrepreneurs and their families with international resources. Going forward, most entrepreneur LPs will enter the fund's advisory committee to provide guidance and support for portfolio founders. This precisely embodies Lighthouse's original intention for this fund: to be the best incubation/seed-round investor for entrepreneurs, and the best co-investment partner for early-stage institutions.

Dark Waves: What does the incubation business mean for Lighthouse now and in the foreseeable future?

Zhile Zheng: Many people see the incubation fund as a marginal business given Lighthouse's current scale. But I believe it's a critical deployment affecting Lighthouse's future strategic depth and industry transformation process. We aim to build it into one of the industry's most boutique ultra-early stage funds.

Our incubation team is small, the incubation fund only $50 million, yet it can deliver rich resources and professional tools to early-stage entrepreneurs. The core support comes from Lighthouse's unique ecological engine — a "super power system" built by over 100 professional investment bankers. I call this: "super engine pulling a small cart." With this super engine, the small cart can go very fast.

Dark Waves: How does an FA known for mid-to-late stage financing build ultra-early stage investment judgment and future foresight capabilities?

Zhile Zheng: This capability is in Lighthouse's DNA. When we did Kuaishou's Series A, the team was just ten people, crammed into a residential apartment in Huaqing Jiayuan. At that time, Hua Su told me: "The future will be an era of 1 billion users watching 100 billion videos daily." I had met over a dozen short video teams then; everyone was talking about influencer operations, traffic playbooks, product design — only he spoke of endgame, that kind of endgame vision cutting through industry noise. I decided to go all-in on the spot.

Later with Bilibili in 2015, it was still an underwater company; no one believed it would become mainstream. But we saw Gen Z cultural fission potentially birthing an "internet version of Disney" — another case of endgame thinking established at an early stage.

So I don't see this as a capability that needs rebuilding; it feels more like traveling back ten years. Compared to ten years ago, we've now built complete investment banking capabilities, formed insights across different stages, and accumulated rich industrial resources. The "reborn, I'm doing XX at XX" narrative is popular now — I feel this is "reborn, Lighthouse doing incubation at the wave's early stage."

---

Recommended Reading

Where Money Flows, People Rise and Fall