Event Recap | How AI Startups Can Drive Overseas Growth and Business Model Innovation

👋 Event Recap

Today at Linear is a regular sharing series hosted by Linear Capital. We hold events in major cities to help founders better understand industry trends and explore opportunities at the frontier of technology.

On June 28, Linear Capital hosted an in-person "Linear AI Sharing Session" in Shanghai, focused on overseas growth and go-to-market strategies for AI startups, as well as business model ideas for AI applications. The discussion explored how AI startups can approach international growth and business model innovation from multiple angles.

After the event, we immediately compiled notes from our two speakers. We hope you find them useful.

📝 Session Notes

01 How AI Startups Can Achieve Overseas Growth and GTM Strategy

Speaker: Davin, former growth and monetization lead for consumer-facing businesses at Tencent and Baidu; former growth marketing machine learning tech lead at Instagram

For AI startups pursuing international growth, success hinges on focusing on specific markets and professional use cases, rapidly validating market feedback, choosing effective distribution channels, and investing seriously in creative and marketing assets. Here are the key takeaways from Davin's session:

When taking our product international, what should we do and not do?

1. From a product-market fit perspective, AI adoption varies significantly across industries. Founders should focus on vertical sectors rather than trying to do everything at once.

2. From a product format perspective, building an AI product as an app may not be cost-effective right now. A web page or browser extension might be a better approach.

3. Products at the 0-to-1 stage need to validate the market quickly and get user feedback fast. You can create an Instant Page on Facebook before even building the product — within two days, you'll know if the idea holds up.

4. When expanding globally, consider three factors:

   • First, target the two largest single markets with sufficient scale: China and the US.
   • Second, consider markets that may seem smaller but have limited competition and high growth rates — Latin America, Russia, the Middle East, and India are all promising options in the AI era.
   • Third, choose a vertical industry direction, perhaps a particularly niche language, and build specifically for that language ecosystem.

During the growth, monetization, and path-to-profitability phase, what should we do more of? What should we avoid?

1. When building a global product, figure out where your target users actually are and which platforms they actively use — don't just think in terms of which country they're in.

2. User retention matters, but for early-stage AI products, potential new users will outnumber existing ones. Focus more attention and energy on acquisition.

3. Never use Product A to drive traffic to Product B within a small portfolio. Countless experiments have proven this inefficient — unless you can do the math and make it work.

4. The best and most efficient way to validate PMF from 0 to 1 is buying traffic through paid search ads, not asking friends and family to try your product.

   • From a financial perspective, the economic cost of roping in friends and family isn't actually lower than buying ads.
   • From a logic perspective, users who come in from the real world are your real target users — they'll surface real problems.

5. How should AI products at the 0-to-1 stage choose media and channels?

   • My recommendation: stick to globally pervasive platforms like Google and Meta. When budgets are tight, don't spread yourself thin.
   • Exception: if what you're building is genuinely imaginative, consider TikTok, YouTube, and Instagram.

6. For products going from 0 to 1, creative and marketing assets are critical conversion drivers. Private communities, KOLs, social media, and content operations can come later.

7. When you need new users or customers to test your product, don't hesitate and don't be stingy. Pay to test traffic. Iteration speed is everything.

Photo: Davin presenting

02 Business Model Ideas for Generative AI Applications

Speaker: Hu Xiuhan, founder of NieTa, former Meta video technology

Generative AI applications can maximize technical advantages and fully unlock commercial value through token economics, content asset monetization, user experience fees, advertising revenue, and broad application scenarios. This model creates new business opportunities while offering users rich, personalized experiences.

Hu Xiuhan's presentation demonstrated the enormous commercial potential of generative AI. Through token economics and content asset monetization, generative AI will bring entirely new experiences and business opportunities to creators and users. His talk centered on several core elements:

1. Generative AI business models rely heavily on token economics. Tokens serve as the cost and value measurement tool for generated content, driving the entire business model.

   • Token consumption reflects not just usage but is also a key metric for evaluating growth.
   • As technology advances, token costs have dropped significantly, lowering user barriers and accelerating adoption.

2. Generative AI applications have a unique business model: generating content assets and monetizing them.

   • Users pay for the unique experience of generated content, enjoying personalized and customized services.
   • Through embedded advertising and brand partnerships, generative AI applications can generate revenue from advertisers, driving value appreciation of content assets.

3. Application scenarios for generative AI include, but are not limited to, the following:

   • Player-generated licensed fan content and hidden skins: Players can use generative AI to create officially licensed derivative works and hidden skins, enhancing the gaming experience.
   • IP-customized hidden storylines: Combining IP licensing with generative AI, players can experience customized hidden storylines and pay for them.

4. Technical advantage is key to successful generative AI business models.

   • Advancing technology continuously improves content quality and generation efficiency.
   • By optimizing models and assets, and combining multiple information-rich assets as model inputs, we can reduce users' prompting burden while enhancing content expressiveness and interactivity.

5. Challenges and opportunities in commercializing generative AI applications

   • Previous-generation recommendation platforms treated token consumption as content production cost; new-generation generative platforms treat token consumption as fuel for monetizing content assets.
   • Current content creators receive very low revenue shares on platforms; generative AI may change this.
   • Generative AI can achieve commercial value through control and combination of content assets, rather than relying on traffic distribution.
   • As generative AI evolves, the boundaries of content production and user experience will continuously expand.

Photo: Hu Xiuhan presenting

✨ Upcoming Events

That wraps up this Linear AI Sharing Session. We'll continue hosting events on related themes — follow Linear Capital's WeChat official account for the latest updates.

About Linear Capital

Linear Capital is an early-stage investment firm focused on "frontier technology + industry" — applying frontier technologies such as data intelligence, digital infrastructure, next-generation robotics, and new technological transformations in traditional sectors (biotech, materials, energy, etc.) to vertical industries to dramatically improve efficiency, solve pain points, and enable industrial upgrading. We capture outsized returns through substantial increases in industrial value. We currently manage ten funds with approximately $2 billion in total AUM.

Our investment stage focuses primarily on leading angel to Series A rounds, with check sizes ranging from $1 million to $10 million (or RMB equivalent).

We have made early-stage investments in over 120 teams, including Horizon Robotics, Kujiale, Sensors Data, Tezign, Rokid, Guandata, and Agile Robots. The combined valuation of Linear's portfolio companies is approximately $20 billion.

In the near term, Linear Capital is working to become the best "Data Intelligence Technology Fund." In the long term, we aim to build the most influential "Frontier Technology Application Fund."