For one week in November, Facebook and Instagram users around the world were shown advertisements promising a 10% weekly return on investments, an implausible figure that would translate into annual gains exceeding 14,000%. The ads showed cash and cryptocurrency raining down on a man at his computer and asked viewers if they wanted similar profits. Despite violating Meta’s policies against get-rich-quick schemes, the ads ran anyway, reaching more than 20,000 users across the United States, Europe, India and Brazil, according to data provided by Meta.
The ads were not placed by a criminal syndicate but by a Reuters reporter testing Meta’s advertising enforcement. Reuters found that by working through advertising agencies publicly endorsed by Meta as “trusted experts,” it was possible to publish investment ads that promised unrealistic returns and attract real inquiries from users. Dozens of people contacted the fake advertiser, underscoring how easily such promotions can reach potential victims.
The reporter created the ads using agencies listed in Meta’s official Partner Directory, where companies are designated as “Badged Partners.” Some of these agencies actively court clients seeking to run banned advertisements, offering access to special ad accounts that enjoy more lenient enforcement. Those accounts are often supplied by large Chinese advertising agencies that form the backbone of Meta’s China ad business and receive special protections from automated takedowns.
Although China bans its citizens from using Facebook and Instagram, it allows domestic companies to advertise to foreign audiences. Meta relies on 11 major Chinese agency partners to sell such ads. Reuters reporting shows that these partners frequently work with intermediaries in China and abroad, creating a system Meta has internally acknowledged is prone to abuse while generating billions of dollars in revenue.
In preparing the test, the Reuters reporter used a real name and was transparent with agencies about wanting to run banned cryptocurrency ads, while taking care not to take anyone’s money or cause harm. A basic website and Facebook pages were set up for a fictitious entity, “Jeff Horwitz Research,” before searching for agencies willing to assist.
What surprised the reporter was not the presence of ad accounts for sale on underground forums, but how easily such services were found within Meta’s own partner directory. One agency, Bluefocus, a Vietnam-based Meta partner linked to a major Chinese reseller, advertised tutorials on bypassing Facebook’s rules, including one titled “How to Advertise Illegal Products on Facebook.” While a Bluefocus representative denied running banned ads, Reuters later received messages from the agency offering to help create fake U.S.-based accounts.
After Reuters presented Meta with evidence that some Badged Partners openly offered to violate advertising rules, the company deleted its partner directory and said it was reviewing both the partners involved and the program itself. Meta spokesperson Andy Stone told Reuters that Chinese partners are not permitted to work with foreign resellers and that the company would investigate violations of that policy.
The reporter also contacted other resellers, including Green Orange in Hong Kong and Uproas in Estonia. All asked whether the ads would comply with Meta’s rules, and all proceeded once told they would not, simply charging higher commissions. For fees of $30 or less, paid in cryptocurrency, each reseller provided ad accounts within hours.
Once the ads were ready, Meta’s systems approved them quickly. In some cases, Meta’s own artificial intelligence tools offered to improve the ads by generating alternate images and text, including language like “Tired of living paycheck to paycheck?” and “Break the cycle and start earning a steady weekly income.” The ads moved from review to active status within minutes and logged hundreds of clicks over several days.
The experiment demonstrated how Meta’s enforcement system treats ads placed through favored partners differently, often requiring human review rather than immediate takedown. Within days, the fake ads prompted more than three dozen inquiries from users seeking investment opportunities. The reporter replied by identifying himself and warning that the promised returns were too good to be true.
The Reuters test highlights how Meta’s advertising infrastructure, particularly its reliance on partner agencies and AI-driven optimization tools, can allow clearly deceptive ads to spread widely before being stopped. It also shows how easily ordinary users searching for financial relief can be exposed to investment schemes that violate Meta’s own stated policies.
