Spending money just to end up with a banned account: the agonizing experience of Chinese Claude users
Is Anthropic a barbaric company? If you ask any Chinese user who has lost their account for no reason, the answer will most likely be yes.
Under normal business logic, if a company carries out indiscriminate, high-frequency account bans for users in a specific region, even banning accounts immediately after users recharge, it will most likely be quickly spurned by consumers and face mass unsubscribes. However, in Anthropic's world, Chinese users have not left at all. Instead, they have flocked to the platform one after another, buying a new account for dozens of yuan every ten days or half a month, and each account is attached with at least a Pro version subscription, willingly sending money across the ocean to Anthropic.
The most classic widely shared Claude account ban meme on the internet
Most of those unexpected account ban moments that caught Claude users off guard have actually been summed up by netizens. Claude's strict risk control system often accurately targets domestic users from multiple dimensions.
The first is the cleanliness and consistency of the network environment. Slightly flawed proxy nodes and frequently jumping IP addresses will bring risks to users; the second is abnormal payment behavior. Domestic users cannot bind overseas credit cards directly, so they turn to top-up services on Taobao, use virtual credit cards or SMS verification platforms for registration. These abnormal bill associations and batch operations will all be marked as risky behaviors;
Furthermore, a series of factors including system time zone, browser language, and login time are also reportedly included in the judgment criteria. The official will not announce the weight of these factors, but some users have long set their system time to the European and American time zones to use their accounts safely, and then follow a dedicated Claude schedule: it is best not to use it during the day, 9 a.m. Beijing Time is the most dangerous time, you can log in after work, and start heavy use from night to early morning.
Finally, there is the almost metaphysical "behavior-identity recognition". On social platforms such as Xiaohongshu, Bilibili and Zhihu, jokes about Claude banning accounts because users use Chinese behavior and language habits have long been widespread. Some people joke that there is no category of "suspected Chinese" in Claude's underlying database, only "confirmed Chinese people and Chinese people who cover up well". In those famous Claude jokes, typing a Chinese comma in your prompt, using a dash with a gap in the middle, or even just telling Claude that you have to work overtime on weekends and that the house has shared public area, may accidentally trigger the alarm and lead to the end of your account.
How many true words are spoken in a fake email
Anthropic's official report shows that from January to June 2026, a total of 11.4 million accounts were banned worldwide, and only 42,000 of the 398,000 appeals were successful. It is unknown whether the extremely low appeal ratio and success rate prove that these more than ten million accounts were indeed "banned with just cause", or that even loyal Anthropic users cannot "guess the will of the authority" before the mysterious iron fist falls.
But in the eyes of consumers, this has to some extent become the best reverse advertisement for Claude — the more fierce the account ban, the more powerful the algorithm is; the more powerful the algorithm is, the more it highlights its absolute dominance in the model ecosystem.
Under the continuous domestication of Anthropic, domestic users who use Claude now are all experts proficient in the cat-and-mouse game. If you enter an AI community or exchange group and ask how to prevent Claude account bans, countless solutions are available for you to choose from: residential IP, exclusive node, virtual card, fill in Wyoming as the billing address... If you are not sure whether your account is safe, you can also go to an independent developer webpage called "Fuck Claude" for self-test to see if you are a pure-blooded Chinese.
It can be seen that in addition to the account ban reasons mentioned above, whether using domestic brand devices, domestic browsers and Chinese fonts, and even the rendering style of Emoji are also within the judgment criteria.
There are many similar pre-detection websites for Claude account ban
Why do people keep sending money to Claude even though they know their accounts will be banned? Hedgehog Commune (ID: ciweigongshe) interviewed several heavy AI users in various industries, and their answers are different but quite bitter:
As a personal witness who has encountered frequent account bans, Sonia, an overseas student, can no longer count how many times her account has been banned. "I didn't pay attention to the purity of the IP for the first account, and it was banned soon; the second account was even more outrageous. I was in Hong Kong at that time, connected to an overseas node, but when I logged in, it prompted me to verify with my mobile phone, and the account was gone immediately after I scanned the code."
Most of the account ban memes are not real chat records, but these pictures reflect the same theme: making one mistake in front of "Lord Claude" may lead to unexpected disaster.
After getting the third account, Sonia began to learn lessons and checked a large number of strategies before using it: use incognito browsing, strictly switch the computer and mobile phone to the same node, and carefully log out of the web page first and then disconnect the node after each use. However, accidents happen unexpectedly — she took a nap at noon while running a project, and the computer screen off caused the node to disconnect automatically. Sonia guessed that maybe the system detected the change of network environment, and when she woke up, the account had already been banned.
As for the fourth account, Anthropic showed its rogue nature even more, and banned Sonia's account within one minute after she recharged. Sonia's case shows that your loyalty to Claude cannot get kindness from Claude in return. "The first sword when you reach the shore cuts down the person who treats you sincerely" — according to Anthropic's IPO prospectus submitted to the U.S. Securities and Exchange Commission, in the first nine months of 2026, about 5% to 7% of Anthropic's revenue came from accounts that were later identified as violating the rules.
Overseas, there are also websites collecting cases of users being banned by Claude. Source: bannedbyanthropic.com
Most of the banned users registered on this website are developers. In addition to Chinese people, there are also Americans, Singaporeans, Kuwaitis, etc.
Another user, Xiao W, is a designer, and Claude is a daily production tool for him. He was banned for the first time at the end of 2025. The account had been used for about four months, and one day when he logged in, he was suddenly prompted that the account was restricted. During that period, he did not change nodes, did not change payment methods, and did not think he said anything out of line. Xiao W said that the only change was that he took an urgent job that week, and logged in and used it more frequently than usual. "Maybe the system thought I was not a normal user, but I was just rushing to finish the work."
When he was banned for the first time, Xiao W carefully wrote an appeal letter, but it did not succeed, and he only received a template reply. Later, Xiao W never considered appealing again — in his eyes, an account that can be used for half a month is worth the cost, and an account that can be used for a month is a great blessing. In my opinion, Xiao W's situation on Claude is similar to that of an overseas illegal worker who cannot safeguard his labor rights.
Now Xiao W's way of using Claude is very practical: he only recharges monthly, exports important materials immediately, and is always on guard against Anthropic's move to pull the rug from under his feet. But even after experiencing many account bans, Xiao W still has not migrated his main work to other AI tools. For him, the workflow and work habits he has accumulated on Claude for a long time, and the production level he can achieve by using Claude, cannot be matched by other large models.
A netizen's matching picture in the post about his Claude account being banned. The first impression Claude's work leaves on many users is simplicity and elegance. Source: Xiaohongshu @James
Another creative worker, Zoe, said that after returning to China from New Zealand, she gradually stopped using Claude. When she was in New Zealand, she used Claude in very clear scenarios: when she needed to sort out the background of an event in cross-cultural communication, or when she only wanted a definite answer to a certain question, and did not need AI to diverge her thoughts. Claude's restraint, stability and unwillingness to add unnecessary content at will are great advantages.
After returning to China for work, Zoe uses ChatGPT more to help her brainstorm and share some personal life thoughts. Zoe told me that her best New Zealand friend has always been heavily dependent on Claude: a typical science and engineering student, who strictly follows the life logic that A is A and B is B, and seems to have little need for emotional and perceptual support.
From Anthropic's attitude towards domestic users, Claude's product style and its user portrait, we can roughly see the underlying logic of this "abusive relationship": The relationship between Anthropic and users is not like a traditional service relationship, but more like a screening mechanism. Anthropic is like an arrogant luxury shopkeeper who takes pleasure in humiliating customers, with well-dressed clerks standing behind the glass, treating everyone as if they just come in to enjoy the air conditioning.
A true NPD is best at pretending to be innocent when questioned
But even after repeated setbacks, domestic professional users still cannot completely give up Claude. From the feedback of users on social media, we can see that after Claude 3.5 Sonnet became popular among the public, its subsequent versions have built highly recognizable barriers in logical reasoning, complex code architecture generation and long text understanding. The irreplaceability of productivity tools gives Anthropic extremely high premium capability, and also constitutes its confidence to treat users in a barbaric way. On the geek highland of commercial competition, when the technology generation gap is significant enough, user experience is squeezed to a secondary position.
I love you, but it's time for me to leave.
Behind the "Claude account ban jokes" full of laughter and bitterness are the real needs and helpless self-mockery of the domestic heavy AI user group. A netizen compared Claude to a "proud but sensible genius young lady from a wealthy family", and he also said that between users and Claude, there is a "perverted and unreasonable stepfather", which is Anthropic.
The two-dimensional image that netizens gave to Claude
Of course, even overseas, Anthropic is a controversial, mixed and alternative Silicon Valley image. This company always stirs the nerves of the entire technology industry with a nearly paranoid, even Quixotic "sect" attitude. They hold high the banner of Constitutional AI, and brand themselves as the ultimate whistleblower of human existential risk. Whenever they release a new model, they always put on a cautious and trembling posture, claiming that because they are worried that the technology is too powerful and may be abused, they have to carry out long and strict safety alignment.
On European and American technology forums, the ironic jokes about Anthropic are even more exaggerated than the "account ban literature". Independent developers in Silicon Valley are also denouncing Anthropic's black-box review mechanism and high API pricing. Some people joked: "Anthropic's safety vision is: before ensuring that humanity is not destroyed by super AI, let all legitimate developers go bankrupt because they cannot afford