Dialogue Immediately 2.5 Lab AI application technical lead: When building AI applications, we will prioritize profits first.

Author | Huang Yu

Every time an AI foundation model is upgraded, it means that a batch of AI application products that rely on single-point capabilities will have their product barriers further compressed, and may even lose the value of existing independently.

This keeps AI app entrepreneurship under a cloud of uncertainty.

Among the AI application products that Xia Junchen, head of technology at Jike’s 2.5 lab, is focusing on developing now is an AI travel translation app called kulikuli. His biggest worry is that his competition is Google Translate.

Xia Junchen told Wallstreetcn, “Every day, the team faces pressure from a rapidly changing market environment,” which is a problem that every AI application team will face.

But he believes that foundation model makers and AI application entrepreneurs are fighting two different battles, with different demands. Small companies care more about how to build a profitable product. In addition, some tracks that startups focus on are easy for big companies to marginalize; once marginalized, small companies still have a gap to survive.

The most successful product incubated by Jike is its podcast platform, Xiaoyuzhou. About four years ago, Jike founded 2.5 lab. This department mainly works on some AI innovation projects.

Xia Junchen told Wallstreetcn that the reason behind the department’s name is to create products for the 2.5% early adopters who are eager to try, and to explore which problems AI can solve that could not be solved before, or to re-optimize problems that already exist.

In the past few years, 2.5 lab has incubated several products in sequence, including the open-source AI client Chatbox, ChatHub, which emphasizes “using multiple large models in one interface and comparing outputs side by side,” kulikuli, and the self-discipline tool “Self-discipline Shitou,” among others.

Xia Junchen revealed that although the user base isn’t very large, the products that 2.5 lab has launched are already profitable.

Between customer acquisition at scale and profit, Xia Junchen prioritizes profit.

For more than a decade, internet products have almost followed the same growth formula: get users for free, quickly expand DAU, then look for a business model.

But in the AI era, this approach starts to fail. The reason is simple—cost.

Unlike traditional internet products, AI services cannot expand infinitely at extremely low marginal cost. Every model call and service delivery incurs real costs. As call volume grows, Token costs quickly become a “value-destroyer.”

Xia Junchen uses his real-world experience to prove this.

He told Wallstreetcn that now almost all AI applications face the same issue: Token costs often account for around 70% of total revenue. If revenue is 1,000 yuan a day, maybe 700 yuan has to be paid to cover model fees. Under such a cost structure, trying to grow by “burning money to buy users” in the way done in the internet era is almost impossible.

Meanwhile, the primary market also won’t be willing to pay purely for user scale. Compared with DAU and downloads, investors care more today about whether revenue, ARR, and commercialization efficiency actually hold up. Only after the business model is proven—showing that people truly are willing to pay for the product—does further growth make sense.

“Growth must come after commercialization is validated.”

In Xia Junchen’s view, this has become a new order for AI app entrepreneurship. And kulikuli’s ability to grow from a “toy project” that was once close to being abandoned into a profitable product with several million users is precisely the result of following this path.

It is understood that the first three months after kulikuli launched were free. At that time, Xia Junchen was already unsure where the project should go due to cost pressure, and even discussed “maybe we should stop this project.” Later, they tried adding paid features, and it turned out that real people were willing to pay—only then did they turn the project profitable.

According to Xia Junchen, kulikuli now has more than 3 million total users, mostly overseas users. The team is about 10 people, and it has a decent level of profitable revenue.

Xia Junchen believes that because Google Translate’s translation quality still cannot reach kulikuli, kulikuli still has room to survive.

As for why Google Translate ends up doing less well than a small company’s product, Xia Junchen thinks that travel translation scenarios are too remote and not profitable enough for a big company like Google. Meanwhile, kulikuli can do well because “the team puts a lot of effort into this, and we also have some engineering accumulation.”

Besides product competitiveness, how to reduce model costs as much as possible is also something the development team must consider.

Xia Junchen revealed that while keeping quality unchanged, it is feasible to cut the cost of mature scenarios by 50% to 70% by using services similar to Tencent Cloud’s TokenHub.

Once survival is solved, ambition will grow out. Xia Junchen said this is a bit like “leveling up to fight monsters”—after climbing a mountain like kulikuli, the team doesn’t want to climb the same one again, and they will want to find a higher one.

In terms of launching new projects, 2.5Lab actually doesn’t have a fixed track, nor does it set up any so-called objective evaluation system. Xia Junchen said, “It’s more subjective.” Nowadays, with agent tools like WorkBuddy, anyone in any role—if they come up with a product idea, build a prototype, and can keep pushing it forward—has a chance to form a team and truly build the project out.

Risk Warning and Disclaimer Terms

        The market is risky; investment requires caution. This article does not constitute personal investment advice, and it has not considered any special investment goals, financial conditions, or needs of specific users. Users should consider whether any opinions, viewpoints, or conclusions in this article align with their specific situation. Invest at your own risk and assume responsibility.