Author: Foreign Exchange Trader
Opening up the simplified Chinese circle on X and scrolling through the trending content, besides deep industry insights, what gets high likes and high reposts are posts about getting free memberships for AI services.
Posts like "Claude Freeloading Guide, Gemini Student Verification, Free GPT Plus US Military Verification" have engagement numbers that crush everything else.
Take a stroll around Xianyu (Idle Fish), and the picture is even more direct. Various "One-Year Pro Memberships" are clearly priced at ten-something,几十块 (a few dozen yuan), with store sales often reaching thousands of orders. Any AI tool you want to use has a "budget alternative" here.
There are many opinions online that attribute this phenomenon to "Simplified Chinese users are really stingy."
But that explanation is too lazy.
The real situation is: the pricing of these tools is simply not designed for the Chinese market.
ChatGPT Plus is $20 a month, which converts to almost two thousand RMB a year. For knowledge workers in Silicon Valley, this is the cost of a few lunches; for the average white-collar worker in Beijing or Shanghai, this is a month's food expenses. The price anchors are completely in different coordinate systems.
Thus, a peculiar market vacuum appears: the demand is real, but almost no one pays through official channels. And this vacuum will inevitably be filled.
The shops on Xianyu are the fillers. Their sources of supply are无非几种 (nothing more than a few types): subscriptions obtained through credit card cashback薅来的 (exploited),低价订阅 (low-price subscriptions) from regions like Turkey and Argentina resold, bulk registrations using educational discounts, and even shared accounts split and sold. It's a gray area, but it works.
You could say this is a pirating mindset. But look at it from another angle: when the official price of a product makes 90% of potential users balk, there's something wrong with the pricing itself.
Some might say, why should American companies give you a discount?
This touches on an old question: should software products have regional pricing?
Netflix does it, with Indian users paying only a quarter of the US monthly fee; Spotify does it, with much cheaper student plans in Southeast Asia; Steam is a classic example, with game prices vastly different in the Russian, Argentine, and Turkish regions.
Why are they willing to do this? Because they've done the math.
For digital products with almost zero marginal cost, one more user is one more portion of revenue. Rather than letting these users completely流失到 (drain into) the gray market, it's better to use prices that match local purchasing power to reel them back in. Even if the per-customer price is low, multiplied by a huge user base, the total revenue is反而更高 (actually higher).
In this wave of AI tool热潮 (boom), most companies haven't taken this step yet.
There are a few possible reasons. First, they're too busy. Fundraising, iterating, grabbing market share, no time for精细化运营 (refined operations). Second,担心套利 (worried about arbitrage). If the price gap is too big, members from low-price regions will be resold to high-price regions,反而冲击 (instead impacting) the core market. Third, they simply don't take the Chinese market seriously. Either they think the water is too deep, or they think the potential is too small.
But the truth is: the demand for AI tools in China might be bigger than anyone imagines.
Go look at the comment sections of those freeloading tutorials; they are full of office workers, students, entrepreneurs. They aren't unwilling to pay; they simply can't afford that price.
This is a classic case of "price discrimination failure." Money that could have been earned全部流进了 (all flows into) the pockets of resellers.
More ironically, this gray market反而帮 (instead helps) AI companies with user education. Many people use foreign AI services for the first time through these channels, get used to them, form a dependency. When their income increases, or the gray channels get shut down, a portion of them will convert into full-price users.
In other words, those shops on Xianyu are, to some extent, helping Silicon Valley with market penetration for free.
Of course, this logic has a flaw. If gray channels exist forever, users will never have the motivation to convert. So these companies will eventually face a choice: continue to放任 (let it be), handing over this huge Chinese market to resellers; or take the initiative, using reasonable regional pricing to bring users back.
Some companies have started to move. OpenAI is testing cheaper plans in some regions.
What about domestic AI manufacturers? This was a heaven-sent opportunity.
Overseas products are highly priced, have high payment barriers, and there's the Great Firewall for access. In theory, domestic AI applications should be躺着都能接住 (able to effortlessly catch) this spillover demand.
But the reality is, most domestic AI tools are also模仿 (copying) Silicon Valley's pricing posture.
Kimi, Tongyi Qianwen, Zhipu, Minimax—although cheaper than overseas options, they aren't cheap enough to eliminate psychological barriers.
More crucially, they haven't created a differentiated price perception.
What users perceive is: "The domestic ones are a bit cheaper, but not much cheaper, and the capability is a notch worse." Once this perception forms, it's hard to turn around.
Actually, domestic manufacturers could take another path: "Price it so low that people feel不好意思 (embarrassed) to freeload."
Think about how Pinduoduo fought Taobao. It wasn't 10% or 20% cheaper, but so cheap that you felt comparing prices was a waste of time. Once the price drops below a certain threshold, the user's mental accounting undergoes a qualitative change—from "I need to compare which is more cost-effective" to "Why hesitate at this price?"
AI tool subscriptions are similar. If a domestic tool dared to price its Pro membership at 9.9 RMB a month, or even lower, directly piercing the user's decision-making cost, what would happen?
First, those gray market shops on Xianyu instantly lose their reason to exist. Go through the trouble of finding a reseller, worry about account bans, only to save a few bucks—who would bother?
Second, user mindset gets locked in. Once accustomed to a tool, the switching cost is extremely high. AI assistants aren't like video sites where you can just switch. The conversation history, usage habits, even its "understanding" of you accumulated in this tool are all assets. Use low prices to圈住 (lock in) users first, then adjust prices slowly after the ecosystem is built—this is basic internet strategy.
Third, reverse educate the market. When domestic tools drive prices to the extreme, the high pricing of overseas products will seem even more outrageous. Users will start to question: Why does ChatGPT cost over a hundred RMB a month? Once this question is planted, the competitive landscape changes.
Of course, low price isn't a万能药 (panacea). If the product isn't good, no one will use it even if it's free. But currently, the capabilities of several leading domestic AIs are already sufficient for the daily needs of most ordinary users—writing copy, looking up information, translation, brainstorming. What's lacking isn't technology, but market strategy.
Another overlooked opportunity: the enterprise market.
Individual users are price-sensitive, but enterprises are different. Enterprise purchasing decisions are based on ROI. If you can prove an AI tool saves employees an hour a day, a monthly fee of a few hundred RMB is根本不是问题 (not a problem at all).
Domestic AI manufacturers should walk on two legs: extremely low prices on the C端 (consumer end) to grab users and cultivate habits, and standardized products on the B B端 (business end) to reap profits. Use the buzz from the C端 to feed sales on the B端, use the revenue from the B端 to support subsidies on the C端. This strategy has been验证过 (validated) by Meituan, Didi, Pinduoduo.
But what do we see now? Domestic manufacturers want both the high pricing of Silicon Valley and the large scale of the Chinese market. Wanting both, they end up with neither.
The deeper problem is: many domestic AI companies are still fundamentally to-VC in their thinking.
In fundraising stories, high customer unit price implies a high ceiling, implying valuation support. If you set the membership fee at 9.9 RMB, investors will ask: Can this make money? How does the financial model work?
Thus, they fall into a paradox: to make the books look good, they don't dare price too low; if the price isn't low, users flow to the gray market; if users flow to the gray market, growth data looks bad; if growth looks bad, it affects the next round of fundraising.
Finally, a vicious cycle forms.
Breaking this cycle requires courage. Someone needs to stand up and say: I'm not playing this game anymore, I'm going to use price to crush everything, get the user base to the maximum size first, and then consider how to monetize.
Whoever figures this out first will eat the biggest红利 (dividend) in the Chinese AI application market.
After all, users who are willing to疯狂找 (frantically search for) freeloading guides on Xianyu and the internet aren't actually unwilling to pay; they are just waiting for a reasonable price.
