Author: Kaori

"Including time cost, the return rate of being a transit station is not as good as going to work directly."

A few days ago, Sukie open sourced all the process of building her transit station. From server procurement to technical configuration to marketing and customer acquisition, she wrote a one-stop tutorial. The post brought her dozens of new users on social media in a single day. We found Sukie and wanted to talk to her. After actually doing this business, what were the differences between what we saw outside.

The transfer station circuit has gone through a complete cycle in the past year, from making a fortune in silence to fighting in the Red Sea, and then to the entry of big players to harvest. Dongcha Beating previously reported on Mo, a webmaster who only works on the B-side. Her story shows the daily operation of the transit station business and the global AI access gap it involves. This time, Sukie, the narrator, provides another perspective: Why does a person with a complete business mind still fail to make money after running through technology, marketing, and compliance.

When the threshold for a business is so low that anyone can enter, and profits are destined to grow positively only through gray operations or ecological nesting dolls, where is the space left for compliant independent operators of transfer stations?

The following is what Sukie has to say.

Accounts from compliant sources are expensive

Our number pool mainly comes from two sources. Partners in Singapore provide part of it, and we register part of it ourselves.

If you register yourself, follow the compliance path and register with real payment methods and identity information. This type of account is costly and slow to obtain, but has the best stability.

There are also some accounts with unknown origins in the industry. Some people will buy stolen accounts, use false identities, or even use robots to register in batches. It is cheap, but the risk is high. Once it is traced by the upstream platform, the entire account pool may be taken away.

This is also why our costs cannot be reduced. Accounts from compliant sources are expensive. This is a core contradiction in the transfer station industry. It is either cheap but has compliance risks, or it is expensive but can run for a long time. We chose the latter.

In 2024, the industry still has 30% to 40% gross profit margin. The top 20 large transfer stations are basically making money. The dimensions of competition are stability and customer service experience, not price.

The turning point will be in the second half of 2025. Wave after wave of new players enter the market, and each wave offers lower prices. First it was 20% off, then 30% off, 40% off, and by the beginning of 2026 some people were getting 50% or even 30% off the official price.

The current situation is that players who are willing to set prices based on healthy gross profit cannot survive because their users have all fled; players who offer ultra-low prices cannot survive because their costs cannot be reduced to that level and they can only rely on gray operations to survive.

From an industry perspective, if you use an account pool from compliant sources, plus labor and server costs, the lowest profitable price is about 20 to 20% off the official price. If it is lower than this range, there is either a problem with the origin of the number pool, or investors’ money is being burned.

10% of the official price, it is impossible to make a profit using compliance means, because the rate of the upstream account itself is more than 10%.

There is word-of-mouth circulation in the developer community. Users attracted by ultra-low prices in the short term will soon be lost if they find that there are problems with service quality, token counts are inconsistent, and model output is inconsistent. But after the loss, they left with distrust of the entire industry, which hurt the entire track.

It is really difficult to earn RMB

The biggest mistake I made was that I was in the United States but chose to earn RMB.

Our costs are settled in U.S. dollars, and the account pool, servers, and payment channels are all in U.S. dollars, but our income is in RMB, and we receive it from the most price-sensitive developer group in the Chinese market. Chinese users are extremely price-sensitive, and their willingness and ability to pay are not at the same level as European and American customers.

People in the United States should directly do business with European and American companies to earn US dollars. Selling a transit station to the Chinese market is equivalent to offsetting all the advantages of geographical location, and instead suffers double losses in exchange rate and payment.

There are problems with the payment link and fierce pricing competition, but the fundamental reason is the low payment ceiling of the customer group itself.

What you earn is hard-earned money. From buying a server to running the transfer station, a person with a little knowledge of technology can do it in a week or two. What really consumes energy and money is the operation link.

The biggest cost is account pool maintenance. Accounts have been blocked or frozen, and the status must be monitored every day, and new accounts must be added regularly. Generally, the pool of a medium-sized transfer station has hundreds to thousands of accounts.

The risk control policy of the upstream platform, whether the request pattern looks like a human being, and the quality of the account source are all factors that affect blocking. Adding these three factors together, the overall account ban rate continues to rise. Some accounts can live for a few months, while others will be subject to risk control on the day they are registered.

The second consumption is customer service. Most of the users of the transfer station are developers, and their questions are very specific, such as why the token count does not match up? Why is the response slow? Why did this model report an error today? Everyone must answer, otherwise the reputation will collapse.

Small and medium-sized enterprises contribute the largest customer unit price, AI shell entrepreneurs contribute the largest active usage, and individual developers contribute the largest user base but have the lowest unit price.

The most difficult to serve are AI shell entrepreneurs, who are extremely price-sensitive, move on monthly price comparisons, and have the lowest retention rate; the number of requests fluctuates greatly, and a product suddenly becomes a hot search. A 100-fold increase in requests in a day can crush the account pool. Feedback is also rapid. In the developer community, the saying "it's not easy to use" spreads much faster than that of B-side customers.

The best service is for small and medium-sized enterprises with specialized technical contacts. They have stable traffic, timely payment, invoices and contracts. Long-term cooperation can be established after a few back and forth.

From a channel perspective, RMB is also more difficult to earn. This time I open sourced the transfer station methodology on X and got dozens of new orders. Overseas developers’ willingness to pay and decision-making speed are one order of magnitude higher than domestic developers. Community fission and agent distribution conversion are the best, because the trust cost of transfer stations is extremely high, and the efficiency of recommendations from acquaintances is more than ten times higher than cold traffic.

SEO and Xiaohongshu are slow but steady channels. SEO users have the strongest purchasing intention, but the traffic is small and the growth is slow. It takes three to six months to take effect. Xiaohongshu has large traffic and medium conversion, and its users are mainly developers and early entrepreneurs in the AI ​​application layer.

The companies with the worst input-output ratios are Douyin and Xianyu. Douyin's algorithm is not friendly to technical content, and it is better to invest in traffic than to grow naturally. Xianyu users expect to solve problems for tens of dollars, but our cost structure does not allow that kind of pricing. More people come to consult and fewer orders are placed, and each consultation is time-consuming.

The third one is compliance and payment. Every cross-border payment must be handled carefully. For example, our academic customers want contracts, overseas customers use the US dollar channel, and domestic customers use RMB to solve the subject and invoice issues. There are separate compliance processes for each type of customer.

Another area is anti-profit. As soon as open source and free trials are opened, people will immediately write scripts to register for free tokens in batches, and the defense strategy must be updated every week.

The most common thing for newbies to underestimate is the details of these operating sectors. Low threshold means fast roll-up, and fast roll-up means no moat.

Make the market more volume

In the past two weeks, Brother Sun, Fu Sheng and even the Trump family all stopped working in the transit station business. This business has reached an inflection point. Sun Ge focuses on privacy and anonymity with the lowest prices on the Internet. Fu Sheng focuses on 15% off the entire store. The Trump family directly cuts 35% off. The most expensive package is $9999 and comes with Mar-a-Lago tickets.

The customer groups of the three people are completely different, but they essentially add a shell to the API of the upstream model and resell it, and no one really lives on the transfer price difference.

Brother Sun's main dish is the payment precipitation on the chain. The money collected by the transfer station is settled through the blockchain, which is equivalent to creating transaction volume for Tron. Fu Sheng’s main dish is the story of Cheetah Mobile’s business transformation, telling the story of AI to the capital market. The main dishes of the Trump family are WLFI tokens and USD1 stable coins, and the transfer station is just a traffic diversion tool.

For the three parties, the transit price difference is not a source of profit at all. They can use transfer stations to attract users at a loss, and then make money from users in terms of ecology, tokens, and brands.

This is why this business is not friendly to individuals. Those who make money in the industry are either eating the information gap of gray operations, or using transfer stations to feed another larger business. Neither end is us.

So this is why I chose to make the transfer station business public. On the one hand, it can be used for marketing. The spread rate of open source posts in the past two days is much higher than the previous advertising posts.

The second is the contribution of the industry. Transfer stations have been severely demonized. The outside world thinks that they are all gray products, marginalia, and liars. But there are people who do it legally. We make the construction process public to let more people know that technology is not a black box.

The moat of this business is already thin, but some black-box players have maintained high gross profits through information gaps. After open source, retail investors can build it themselves, and the technical threshold premium for black box players will disappear. The remaining real skills are the stability of the account pool, customer service experience, and compliance capabilities.

So the real meaning of making the industry more flexible is to shift the price of products to the quality of services. For us, this is pushing the battlefield in our favor.

It is no longer meaningful to continue to increase prices in the C-end retail battlefield. We open source the tools and let small players build their own. Instead, we can focus on high-margin customers in niche markets.

However, I advise friends who still want to do this industry not to get involved easily.

If you just want to make some pocket money, use it yourself and sell it to a few friends around you, you can try it. The technical threshold is not high, you can run one yourself for your own use, and earn back the server money by the way.

If you want to invest in this full-time as a startup project, don't do it. High-end players have their own ecology behind them, and ultra-low-price players are using gray methods. If you comply with the regulations, you cannot negotiate the price with them.

If you have already entered the market, focus on niche markets, such as B-side, academic institutions, or overseas markets.

There are so many more worthy investment directions in the AI industry, and there is no need to bet on a track without a moat.

The transfer station is just a ticket, not the destination.