Meta, Coinbase, Google, Microsoft, OpenAI, Starlink, Adobe and other technology companies today published the results of a joint operation with the UK National Crime Agency, the FBI and the US Department of Justice against scam networks in West Africa. Meta removed more than 3.7 million accounts, Pages and Groups from Facebook and Instagram, law enforcement confirmed 28 arrests, and 117 people trafficked into a scam compound in Ghana were rescued.

In Short

Police from several countries and big tech companies spent a week in London in October working together to find and shut down online scam operations in Nigeria, Ghana and nearby countries. The scammers mostly pretended to be Americans and Britons on Facebook and Instagram, and some of the people doing the scamming were themselves trafficking victims held against their will. Millions of fake accounts are gone, 28 people were arrested, and the companies say they plan to repeat this kind of joint operation.

What the operation produced

The announcement, published on Meta's newsroom on October 13, 2026, describes the effort as the first joint disruption to broadly target criminal scam networks in the West Africa region and, according to Meta, "the largest operation of its kind to date." It was hosted by the National Crime Agency (NCA) at its London headquarters in partnership with the US Department of Justice, the FBI and several other US federal agencies.

Four sets of results are listed in the release.

Digital disruptions. Meta removed more than 3.7 million accounts, Pages and Groups based in West Africa from Facebook and Instagram. According to Meta, a majority of them appeared to target people in the United States and the United Kingdom using profiles that impersonated American and British users. Microsoft also suspended accounts suspected of being associated with the scam activity, though the release gives no number for Microsoft.

Arrests and rescues. Leads on more than 30 scam centers were shared. Law enforcement confirmed the arrests of 28 individuals and the rescue of 117 victims who had been trafficked as part of the criminal operations.

Victim funds. Coinbase identified more than 65 active scam victims and stopped more than $415,000 in potential losses before they occurred, according to the release.

Intelligence. The information exchanged, according to Meta, opened new opportunities to deepen investigations and let platforms widen detection and enforcement against the groups involved. No figures accompany this category.

How large is 3.7 million? Context helps. In March 2026 Meta said it had taken down 10.9 million Facebook and Instagram accounts tied to criminal scam centers, and that a separate joint operation disabled more than 150,000 accounts, figures PPC Land covered when Meta deployed AI and law enforcement against scams across Facebook and WhatsApp. The West African total from a single week is roughly a third of that earlier scam-center figure. Set against platform-wide enforcement, however, it is small: a VAB analysis found that Facebook banned about 3.5 billion fake accounts in 2025.

The release does not break the 3.7 million down by type. It does not say how many were accounts, how many were Pages and how many were Groups, nor whether any were advertiser accounts or had run paid ads. That matters for anyone trying to read the operation as an advertising story rather than a content moderation one.

Who was in the room

Beginning Tuesday, October 6, the NCA convened Meta and other technology companies at its London headquarters, in coordination with the FBI, the Department of Justice, Homeland Security Investigations and other US agencies. Ghanaian and Nigerian law enforcement attended, as did law enforcement partners from Australia, Canada, New Zealand and India. Other participating US bodies named in the release were the US Secret Service, the US Department of the Treasury and the Homeland Security Task Force (HSTF).

The week opened with remarks from law enforcement leaders, including a call to action from Jeanine Pirro, US Attorney for the District of Columbia and head of the Department of Justice Scam Center Strike Force.

There is a counting gap in the release. Its summary says Meta and 14 other technology companies took part alongside law enforcement counterparts from ten countries. The text names seven companies in its opening paragraph - Meta, Coinbase, Google, Microsoft, OpenAI, Starlink and Adobe - and the quoted statements add Chainalysis, Silent Push and TRM Labs, for ten in total. Eight countries are named: the UK, the US, Ghana, Nigeria, Australia, Canada, New Zealand and India. The remaining five companies and two countries are not identified.

Google is listed among the participants, yet the release carries no statement and no result attributed to Google. Adobe is in the same position. What either company contributed is not described.

There is also a minor discrepancy on duration. The release says the work began on October 6 and continued "over the following days." TRM Labs refers to "three days of joint work," while Meta's quoted statement places the operation "last week."

The rescue in Ghana

The most concrete physical outcome concerns the 117 people rescued. According to the release, they had been trafficked to work against their will in scam compounds in Ghana, forced into scam operations by criminal syndicates and held in a residential compound. The victims were malnourished and subjected to physical and psychological abuse, the release states. Their captors were arrested and face prosecution.

Meta draws a distinction between this region and Southeast Asia. Scam centers in Nigeria, Ghana and neighbouring countries "are generally more decentralized and less industrial than those in southeast Asia," the company wrote. "Criminals are continually seeking new havens where they believe they can operate with impunity."

Coinbase framed the shift as displacement. "When we disrupted scam compounds in Southeast Asia earlier this year, we proved that coordinated action across industry and law enforcement works," said Leah Bressack, Vice President at Coinbase. "As criminal networks try to reestablish in Africa, Coinbase's Global Intelligence team is following the money, freezing assets, and sharing intelligence in real time so law enforcement can act."

The Southeast Asia operation she refers to is listed on Meta's newsroom as a June 3, 2026 announcement. Earlier PPC Land coverage tracked the build-up: in December 2025, Meta said it had removed 134 million scam ads in 2025 and disrupted nearly 12 million accounts tied to scam centers in the first half of that year, and in March 2026 the company reported 21 arrests by the Royal Thai Police after a second Joint Disruption Week.

West Africa is not new territory for Meta's enforcement teams either. In July 2024, Meta removed about 63,000 Instagram accounts linked to financial sextortion from Nigeria. In February 2025, Meta said it had taken down more than 408,000 romance scam accounts in 2024, mainly tied to West African countries. And the March 2026 update recorded seven arrests in Agbor, Delta State, Nigeria, in an effort involving the Nigeria Police Force and the NCA.

Seven scam types

The operation identified and disrupted a range of fraud types, which the release lists:

  • Investment scams, where criminals build fake romantic or professional relationships before steering victims into fraudulent investment schemes, including cryptocurrency;
  • Job or task scams, which offer fake employment or online tasks to extract upfront fees or transfers;
  • Law enforcement impersonation, in which criminals pose as police or government officials to extort victims;
  • Military impersonation, with criminals posing as service personnel to gain trust and solicit money;
  • Non-delivery scams, where sellers take payment for goods or services that never arrive;
  • Romance scams, built on fake romantic relationships; and
  • Scam recovery, which targets prior victims with promises to recover lost money for an upfront fee, then defrauds them again.

Several of these map directly onto paid media. Investment scams are the largest category by losses in US data. When the Federal Trade Commission published its Data Spotlight in April 2026, social media scams cost Americans $2.1 billion in 2025, with investment scams accounting for $1.1 billion. Those typically began with an ad or post promoting an investment course. Non-delivery scams overlap with shopping fraud, which the FTC found was the most frequently reported type, with more than 40% of victims saying it began with an ad. Job scams, meanwhile, hit younger users hardest: LinkedIn survey data showed Gen Z loses money to job scams at five times the boomer rate.

What the law enforcement agencies said

Officials gave figures that do not appear in the release's own results section. Kash Patel, Director of the FBI, said: "Thanks to brilliant partnership with the Department of Justice and the private sector, across multiple African countries, we arrested dozens of criminal actors, seized hundreds of devices, identified at least $11 million in potentially freezable funds, and truly saved lives."

Two points stand out. The FBI refers to "dozens" of arrests across "multiple African countries," while the confirmed number in the release is 28. And $11 million is described as potentially freezable, not frozen or seized. Meta's own statement mentions "cryptocurrency seizures" without a figure.

Nick Sharp, Deputy Director of Fraud at the NCA, said: "Fraudsters can operate from anywhere and will target victims both domestically and globally. The NCA will continue to work with international and private sector partners to tackle fraudsters working upstream, overseas, and online."

Pirro, described in the release as founder of the Scam Center Strike Force, said she had "partnered with the NCA and the Crown Prosecutor for England and Wales to push forward cases on both sides of the Atlantic; criminals are not hindered by borders and neither are we."

Lord Hanson, the UK Minister for Fraud, said: "Criminal gangs operating overseas are behind most of the scams targeting the British public." He added that a new Online Crime Centre "will build on this work to dismantle criminal networks and bring offenders to justice, wherever in the world they might be." The release gives no launch date or budget for that centre.

Infrastructure, money and satellites

The technical detail in the release comes mostly from smaller specialist firms rather than the large platforms.

Silent Push, a threat intelligence company, said it mapped "more than 35,000 malicious domains, dedicated internet infrastructure and nearly 700 communication channels tied to scam centers operating across Africa." According to Michael Sweeney, Director of Preemptive Cyber Defense at Silent Push, pairing that mapping with attribution analysis "helped law enforcement trace these networks back to physical locations and the operators running them, which led directly to enforcement action on the ground." The geography differs: the 35,000 figure covers Africa broadly, not only West Africa.

TRM Labs, a blockchain analytics firm, said it helped trace scam proceeds toward freezing and seizure and "mapped the infrastructure behind dozens of fraud sites." Chris Wong, Mission Director at TRM Labs, said the work exposed "a large-scale peer-to-peer laundering network" and identified operators running fraud across the UK, the US and West Africa.

Chainalysis, another blockchain analysis company, did not give numbers. "Fraud networks now run at industrial scale, using AI to target victims and move proceeds across chains faster than any manual process can follow," said Will LaForest, its Chief Customer Officer.

Starlink's statement was notably defensive. "As with any consumer product, the risk of abuse is always present and is not something that Starlink contends with in isolation," said Lauren Dreyer, Vice President of Starlink Business Operations at SpaceX. She said Starlink's Acceptable Use Policy prohibits fraud and that the company is "diligent in disabling any kits involved in illegal activity." The release does not say how many Starlink kits, if any, were disabled during the operation.

Microsoft and OpenAI kept their remarks general. Steven Masada, Associate General Counsel and General Manager of Microsoft's Digital Crimes Unit, said: "No single organization sees the full picture, which is why this kind of collaboration matters." Jack Stubbs, Lead Scams Investigator at OpenAI, said OpenAI is "using AI to disrupt scam operations and help people spot scams."

The role of AI cuts both ways. Research covered by PPC Land in July 2026 found that AI cut the entry cost of a scam farm to $5,000, with an operator able to rent the whole stack for $2,970 a month - well within reach of the smaller, decentralized operations Meta describes in West Africa.

Why this matters for advertisers

The release does not use the word "ads" once. Its results are measured in accounts, Pages, Groups, arrests and rescues. For marketers, that omission is the first thing to notice, because the scrutiny Meta faces over fraud has centered on paid placements.

In November 2025, Reuters reported internal Meta projections that put scam and banned-goods advertising at about 10% of 2024 revenue, roughly $16 billion. PPC Land's reporting at the time detailed how Meta charged suspected fraudsters premium rates while earning billions from scam ads, using penalty bids rather than bans and requiring 95% certainty before removing an advertiser. Those figures have since appeared in litigation and trade-body disputes. IAB Sweden expelled Meta in March 2026, saying its work against deceptive ads was not enough. In April, the Consumer Federation of America sued Meta in the District of Columbia Superior Court under the local consumer protection statute. And in September, a German court found that Meta faces up to 250,000 euros per fake Finanzfluss ad, ruling that its ad auction prevented it from relying on the hosting exemption.

The joint operation addresses one part of that picture: the organic and impersonation side of fraud, where fake profiles posing as American and British users initiate contact. It says nothing about the malvertising channel or about how the 3.7 million removals relate to Meta's advertiser systems. Meta had already stated a goal of having verified advertisers account for 90% of ad revenue by the end of 2026, up from 70%, as part of its wider advertiser verification push. Whether accounts removed in London had passed or failed those checks is not addressed.

Brand impersonation is the other point of contact. Legitimate advertisers compete in the same auctions and feeds as fraudulent ones, and impersonation of real people and brands has been a recurring theme in Meta's enforcement. In February 2026, Meta sued scam advertisers in Brazil, China and Vietnam over celeb-bait and cloaking, one case involving discounted "Longchamp" items. The West African networks, according to the release, relied on impersonating ordinary users rather than brands, which places them closer to the romance and investment schemes that drive the largest losses in FTC data.

Google's absence from the detail is also worth noting. The company is named as a participant, but no Google figure appears. Its most recent annual report, covered in April 2026, said Gemini blocked 8.3 billion bad ads in 2025 and that 602 million ads were removed for violations most closely associated with scams. How much of Google's inventory or infrastructure was linked to the West African networks is not stated.

What comes next

Meta closes the release by saying meaningful progress "ultimately depends on real consequences for the criminal organizations behind these schemes," and that it and its partners are committed to the model's "ongoing strength and repeatability." No date for a further operation is given.

Prosecutions are the clearest open thread. The 28 arrests and the captors in the Ghana compound are described as facing prosecution, but the release names no defendants, charges or courts. The $11 million in potentially freezable funds cited by the FBI may or may not be frozen. And the UK's Online Crime Centre, which Lord Hanson tied to this work, remains undescribed.

The operation took place days before many advertisers start planning for the fourth quarter shopping season, the period in which non-delivery and shopping fraud typically rise. A June 2026 report showed how this plays out around major events, when Meta's World Cup safety playbook described a Visa-assisted takedown of a network using spoofed FIFA sites. Whether a disruption week in October produces fewer scam contacts in November and December is something the release cannot yet answer.

Timeline

Summary

Who: Meta, Coinbase, Google, Microsoft, OpenAI, Starlink, Adobe, Chainalysis, Silent Push, TRM Labs and other unnamed technology companies, working with the UK National Crime Agency, the US Department of Justice, the FBI, Homeland Security Investigations, the US Secret Service, the US Treasury, the Homeland Security Task Force, and police from Ghana, Nigeria, Australia, Canada, New Zealand and India.

What: A joint disruption against West African scam networks that led Meta to remove more than 3.7 million accounts, Pages and Groups, produced leads on more than 30 scam centers, 28 confirmed arrests and the rescue of 117 trafficking victims, and allowed Coinbase to stop more than $415,000 in potential losses. The FBI cited at least $11 million in potentially freezable funds.

When: The operation began on October 6, 2026, and its results were published today, October 13, 2026.

Where: Coordinated from NCA headquarters in London, targeting scam centers in Nigeria, Ghana and other West African countries, with victims mostly in the US and the UK.

Why: Scam networks are moving into regions where they expect less enforcement, and the companies and agencies say no single organization can see a network across platforms, payment systems and jurisdictions. For advertisers, it adds to a year of pressure on Meta over fraud, though the release reports account removals rather than any figures on scam ads.