The Scene Everyone Pretends Not to See
Somewhere in Shenzhen, China, a small advertising firm spins up a Facebook account using a fake name, a borrowed birthday, and a crypto-funded wallet. Within hours, an ad promising miracle supplements or guaranteed investment returns is running in English, aimed at retirees in Ohio or small business owners in Ontario. The victims will never know where the ad came from. Meta will know almost immediately and will do very little, if anything about it.
Inside Meta, this reality is not a mystery. It is documented, quantified, and debated in internal presentations. It is discussed in dashboards and spreadsheets. It is flagged, tracked, and ultimately tolerated.
Over the past four years, internal documents reviewed by Reuters show that the company quietly concluded that a significant share of its China-linked advertising business was built on fraud. Nearly a fifth of Meta’s China revenue, more than $3 billion annually, was tied to scams, illegal gambling, and banned products. Executives understood the scale of the harm. Engineers built tools to reduce it. Safety teams proved those tools worked.
Then the effort stopped, seemingly overnight.
What followed was not a failure of detection, nor a technological shortcoming. It was a deliberate recalibration of risk. Meta did not lose control of its platforms to fraudsters. It chose to coexist with them.
A Market with No Consequences
China occupies a singular position in Meta’s global ecosystem. The country bans Facebook, Instagram, and WhatsApp for domestic users, yet allows Chinese firms to advertise freely to foreign audiences. The result is a paradoxical arrangement: Meta cannot operate in China, but it earns more than $18 billion a year from Chinese advertisers, over a tenth of its global revenue.
Internally, Meta staff has dubbed China the company’s top “Scam Exporting Nation.” Roughly a quarter of all scam ads worldwide were traced back to Chinese advertisers. Even China’s national calendar affected global fraud levels; during Golden Week, when millions of Chinese citizens travel, scam volumes on Meta platforms declined worldwide.
Yet Chinese advertisers face no domestic risk. Because the victims are overseas, Beijing largely looks away. The consultants Meta hired to study the problem reached a blunt conclusion: China’s ad fraud ecosystem thrives because it operates in a space where enforcement is weak, oversight is fragmented, and accountability is virtually absent.
Meta did not create this environment. But it adapted to it, accepted it and, in doing so, became a central node in its operation.
The System Profiting Off Abuse
The architecture of Meta’s China advertising business is not incidental. It is foundational to the problem.
Rather than dealing directly with advertisers, Meta relies on a small group of “top-tier” Chinese ad agencies. These partners receive commissions, preferential treatment, and special protections. Their accounts are whitelisted, meaning flagged ads are not immediately removed. Instead, they undergo secondary human review, often delayed, sometimes never completed.
That delay, is all scammers need to execute their operations.
Below these top-tier agencies sits a shadow market of second-tier resellers. They sublease access to Meta’s platforms to hundreds of smaller firms, many of which openly advertise their ability to bypass enforcement. Accounts can be purchased for as little as $30. Identity verification is minimal. When one account is burned, another appears.
Internal documents show Meta understood this system was nearly impossible to police effectively. Consultants warned that Meta’s own incentives, commissions, growth targets, and uneven enforcement, were actively encouraging abuse. Competitors such as Google and TikTok imposed stricter identity checks and faster takedowns. Meta did not.
The Pivot That Opened the Floodgates
In early 2024, Meta briefly chose a different path. A dedicated team was formed to tackle Chinese ad fraud. New detection and enforcement tools were deployed. Within months, the share of banned ads tied to China was cut in half.
The success was measurable, immediate, and financially inconvenient.
According to internal records, the intervention threatened billions in revenue. By late 2024, after what documents describe as an “Integrity Strategy pivot” involving CEO Mark Zuckerberg, the crackdown was halted. The China-focused enforcement team was dissolved. Restrictions on new Chinese ad agencies were lifted. Safeguards that internal testing showed were effective were shelved.
The company’s stated objective shifted. Meta would no longer aim to bring Chinese advertisers into parity with global standards. Instead, it would seek to “maintain the percentage of global harm.”
Fraud was no longer framed as a systemic failure. It became a tolerable externality.
When Harm Becomes a Line Item
This tolerance has consequences measured not in abstract metrics, but in real losses borne by users. Scam ads originating from China have fueled investment frauds that wiped out savings across North America. In one case cited by U.S. prosecutors, Facebook and Instagram ads funneled victims into WhatsApp groups run by fraudsters posing as American investment advisers. The FBI seized more than $200 million linked to the scheme.
Meta cooperated with investigators. Accounts were removed. The ecosystem regenerated.
Internally, staff flagged advertiser after advertiser with violation rates exceeding 50%. Some ranked among Meta’s top global spenders, alongside brands like BMW and American Express. Rather than sever ties, Meta raised their advertising costs, a “penalty” that functioned less as deterrence than as a licensing fee for misconduct.
As one internal assessment noted, shutting down abusive accounts might temporarily reduce harm, but “it’s likely the revenue will return.” The system is self-healing, in the worst possible way.
The Pattern of Offshore Harm
There is something deeply familiar about this arrangement. Like polluting industries relocating to weakly regulated jurisdictions, Meta’s fraud problem has been effectively offshored. The perpetrators operate in a space with little domestic oversight. The victims live in countries with strong consumer protections. The profits accrue in Silicon Valley.
By allowing privileged intermediaries to shield advertisers, by accepting elevated abuse as the cost of market access, and by prioritizing revenue stability over user protection, the company has created a two-tier system of accountability. One set of rules applies to advertisers in regulated markets. Another applies where enforcement is inconvenient.
The Real Risk for Meta
The immediate risk is regulatory. U.S. senators have already urged investigations by the SEC and FTC. Evidence that Meta knowingly tolerated fraud to preserve revenue could invite serious scrutiny, particularly if investors were not fully informed of the scale of “high-risk” advertising income.
But the deeper risk is structural. Meta’s platforms depend on trust, trust that ads are legitimate, that enforcement is fair, and that harm is not silently priced into the business model. Each time the company chooses growth over integrity, that trust erodes.
Conclusion: The Cost of Looking Away
Meta often positions scam proliferation as an arms race against ever-more sophisticated criminals. The internal documents tell a more uncomfortable story. This was not a failure to keep pace. It was a decision to pause.
Like exiled power brokers funding instability from afar, Meta’s leadership has chosen containment over resolution, tolerating harm so long as it remains profitable, distant, and politically diffuse. The victims are scattered. The responsibility is abstract. The revenue is concrete.
The question now is not whether Meta can reduce fraud, it has already demonstrated that it can. The question is whether it is willing to accept the financial cost of doing so consistently. Because in the end, this is not just a story about China, scammers, or even technology. It is about what happens when a global platform decides that some victims matter less than quarterly growth and builds a system capable of sustaining that choice indefinitely.
This briefing is based on information from Reuters.
