For most of the past two decades, betting against Meta Platforms has been a losing trade. This week, the market started charging more for that bet than it ever has.
Earlier today, Rhode Island Attorney General Peter Neronha sued the owner of Facebook, Instagram and WhatsApp in Providence Superior Court, accusing it of knowingly profiting from fraudulent advertising while assuring users it was working to keep them safe. It is the first scam-ad case brought against Meta by a US state, but it is not the first of its kind. California's Santa Clara County filed a similar action in May, the US Virgin Islands, a territory, sued over scam ads in December 2025, and the Consumer Federation of America filed a proposed class action in April.
Also today, the cost of protecting Meta's bonds against default has climbed to an all-time high. Five-year credit default swaps, effectively an insurance policy on a company's debt, widened 7.6 basis points to 114.79, according to Markit data cited by Reuters. The move followed Meta's decision to lift the floor of its 2026 capital spending forecast to $135 billion. The shares fell about 8%, their steepest one-day drop since July.
Put the two together and you get a company squeezed from both directions: an AI arms race that is eating its cash, and a legal backlog that keeps getting heavier. For insurance professionals, both halves of the story matter.
The complaint leans on internal Meta documents that Reuters brought to light last year. Those records estimated that the company served around 15 billion "higher risk" scam ads a day and earned roughly 10% of its 2024 revenue from fraudulent e-commerce, dubious health products and other banned promotions. Meta has called that reporting a distorted picture of its anti-fraud work.
Rhode Island goes further. It alleges that Meta's own AI ad-building tools sometimes helped scammers sharpen their pitches, and that instead of reliably blocking suspected fraudsters, the company at times charged them higher rates. The state says many schemes were aimed at older users, citing research by the nonprofit Center for Countering Digital Hate that flagged nearly 94,000 scam ads. One example in the filing promised to reverse Alzheimer's disease in 45 days and was running in more than 2,000 versions.
The state is asking for civil penalties for each violation, restitution for consumers, disgorgement of profits and a court order forcing Meta to change course. Neronha said his office would do "everything within our authority to hold them to account." Rhode Island was also among the states that sued Meta in 2023 over harms to young users.
Meta called the new suit "unfounded." Spokesperson Andy Stone said the company had taken down more than 159 million scam ads in the past year and added new safeguards, including broader advertiser verification and WhatsApp warnings about risky device-linking requests. He also criticized the attorney general for bringing in an out-of-state private law firm rather than working with the company.
Read next: Meta's insurers win child harm defense battle in Delaware court
Rhode Island's filing has joined a crowded docket. On Wednesday, a New Mexico judge ordered Meta to pay the state $25 billion after a jury found it had misled Facebook users about data privacy, a case rooted in the Cambridge Analytica scandal. Judge Francis Mathew wrote that state law would have supported a far larger penalty before he reduced it. Meta has said it will appeal.
In August, Meta reached an agreement with 52 attorneys general over teen mental health claims that provides for payments of up to about $18.59 billion, according to its third-quarter earnings release. It booked a $9.89 billion charge for that deal in the quarter. Earlier this year, New Mexico also won $375 million in jury-awarded penalties and a $567 million abatement order in a separate child-safety case, pushing that matter close to $1 billion.
For casualty carriers, the key question is who ends up paying, and so far the answer has largely been Meta. A Delaware Superior Court judge ruled earlier this year that Hartford, Chubb and more than a dozen other liability insurers had no duty to defend Meta in the youth-addiction litigation, on the reasoning that a deliberately engineered product isn't an "accident." That ruling dealt only with the duty to defend and can still be appealed, but it explains why design-based liability claims are making Meta's insurers so wary.
Consumer-protection cases like Rhode Island's may be even harder to shift onto an insurer. Civil penalties and disgorged profits are frequently excluded from liability and D&O wordings, and many states treat them as uninsurable on public policy grounds, although the result always turns on the policy language and the governing law. Coverage counsel will be watching to see whether plaintiffs try to recast scam-ad harm as a negligent failure to police the platform rather than intentional conduct.
That is also where the youth-harm coverage fight is heading. Rosehana Amin, a Clyde & Co partner who has advised on coverage questions arising from the Meta litigation said insurers are likely to lean on intentional-conduct and "expected or intended injury" exclusions wherever complaints allege deliberate design choices with known harms. Rhode Island's claim that Meta knowingly helped build and optimize scam ads fits that pattern closely.
The directors and officers side is also active. Meta is still defending a securities class action, first filed in 2021, over its disclosures on content enforcement and user well-being, according to its latest quarterly report. Each new allegation about what the company knew internally adds material that shareholder plaintiffs can draw on.
Read next: Appeals court keeps Meta, Google and TikTok on the hook over young users
The scam-ad allegations also matter for personal and commercial lines. Losses reported to the FBI's Internet Crime Complaint Center reached about $20.9 billion in 2025, up 26% on the previous year, according to the bureau's 2025 Internet Crime Report. Americans aged 60 and over accounted for roughly $7.7 billion of that total. The FBI also logged more than 22,000 complaints involving AI, with close to $900 million in losses.
Much of that money never touches an insurance policy. A growing slice does, through social engineering and funds-transfer fraud cover in crime and cyber programs, and through identity-theft endorsements on homeowners policies. Rhode Island's complaint describes scammers using the name and image of a senior Bank of America executive to steer Facebook and Instagram users into bogus stock trades on WhatsApp. That is exactly the kind of impersonation scheme that deepfakes are making harder to spot, and that cyber and crime underwriters have been repricing.
Thursday's selloff, though, was about spending, not scams. Meta's third-quarter results showed free cash flow of negative $2.12 billion as capital expenditure reached $34.91 billion, the first time its cash flow has turned negative since 2012, according to Reuters. The company also unveiled a new class of non-voting shares, due to take effect in January.
Steve Sosnick, chief market analyst at Interactive Brokers, told Reuters that investors saw the new stock as a possible "backdoor" for future issuance that could dilute existing holders. The pain spread across the AI trade. CoreWeave sank more than 8%, the Philadelphia Semiconductor Index posted its biggest one-day decline since July, and default swaps on Oracle edged toward their own record.
That matters to insurers in their role as investors, not just as underwriters. AM Best has warned that carriers with private credit or financing ties to the data center boom may be carrying more balance-sheet risk than they realize, even as the same buildout fuels demand for property, builders risk and business interruption cover.
Read next: AM Best warns P&C insurers to rethink data center coverage
Meta still ended September with $82.45 billion in cash and marketable securities, more than enough to cover the penalties announced so far. What the default-swap market now seems to be pricing is whether the bills, legal and technological, will keep arriving faster than the returns.
