Crazy post. Makes you think how many harms is Anthropic willing to create in the world to prevent one ostensible harm?
https://www.chinatalk.media/p/how-to-buy-cheap-claude-tokens-in
Anthropic's selective passport-plus-selfie checks for certain Claude features have prompted Chinese operators to set up transfer stations that pool accounts and forward requests, letting users pay in RMB for access priced at roughly one-tenth the official rate while the proxies handle the overseas verification layer.
Crazy post. Makes you think how many harms is Anthropic willing to create in the world to prevent one ostensible harm?
https://www.chinatalk.media/p/how-to-buy-cheap-claude-tokens-in
Operators rely on bulk registration, shared subscriptions, and occasional credential theft to maintain supply, yet the exact mix of legitimate versus fraudulent sources is not detailed in available reporting.
The same middle servers can log user inputs and potentially resell them, though the downstream volume and uses of that data are still unknown.
Many users objected to exposing gray-market proxies for cheap Claude tokens because Anthropic's verification rules make them the only affordable access option for broke users.
No Digg Deeper questions have been answered for this story yet.
A crazy blog.
Chinese developers are buying Claude access through gray-market API transfer stations that can sell tokens at 5% to 10% of official prices while hiding the real user from Anthropic.
A transfer station is a middle server that takes a user’s prompt, sends it to Claude through overseas accounts, returns the answer, and collects payment through WeChat or Alipay.
The transfer station collects many Claude accounts through free credits, discounted accounts, shared subscriptions, overseas payment workarounds, fake verification, or sometimes stolen-card accounts.
It connects all those accounts behind one proxy, so Chinese users do not talk to Anthropic directly and only pay the proxy in RMB.
The cheap price comes from account farming, free-credit abuse, resale of unused quota, subscription splitting, possible stolen cards, and a darker trade where user prompts and outputs become training data.
So the price hugely cheap not because Anthropic is giving a discount; it is cheap because the transfer station lowers its own cost and creates extra hidden revenue.
The user thinks they are buying cheap inference, but the proxy may swap Opus for weaker models, inflate token use, or store private code, tool calls, reasoning traces, and business data.
The proxy may store user prompts, code, outputs, and tool traces, then sell or reuse that data for model training.
This breaks a core assumption behind KYC, account bans, and abuse monitoring: the AI company sees the proxy, not the real person, so banning one account leaves the upstream supply chain alive.

How to Buy Cheap Claude Tokens in China The Transfer Station Economy, Explained
https://www.chinatalk.media/p/how-to-buy-cheap-claude-tokens-in

@matt_leho 😅

@deliprao conclusion: as ai systems adopt stricter KYC, the risk of incentivizing malicious workaround

@rohanpaul_ai 5% pricing creates shadow infra fast. Demand is real. Trust risk is the part that bugs me.

@rohanpaul_ai it's cheap though and fuck anthropic

@rohanpaul_ai At 5-10% of list price that isn't efficiency, it's someone reselling credits they aren't paying full freight for - abused trials, leaked keys, whatever's upstream. Cheap right up until the tap gets shut off.

@rohanpaul_ai 这绕法也太互联网缝合怪了

@deliprao that framing skips the real question though
are we even sure the "ostensible harm" is the one actually getting prevented?

@rohanpaul_ai @rohanpaul_ai this is wild. never thought about how deep the gray market stuff goes with AI access. kinda impressive and a little scary too tbh.

@rohanpaul_ai Why are you exposing this?? It’s the only way for broke people like me to have access to these models at a reasonable price so please just allow it 😭😭

@rohanpaul_ai it has the ring of truthiness, but it could also be government disinformation designed to whip up support for regulation trust no one

@rohanpaul_ai Adding this to my 'try once' list. I test everything once. 🙂