Advertisers lost $32.6 billion to ad fraud in 2025, according to Spider Labs’ 2026 Ad Fraud White Paper, which scanned 6.05 billion clicks and $6.2 billion in tracked ad spend across 174,483 domains in 242 countries. Inside that data, campaigns optimized by AI carried up to twice the fraud rate of an average campaign. That detail matters more than the total, because it points at where the next round of waste is coming from.
Made for advertising sites are not a side problem anymore. Spider Labs reports that placements on MFA sites, pages built mainly to carry ads rather than serve readers, rose 14 times over the prior year. Spider Labs CEO Satoko Ohtsuki summed up why that is worse than ordinary waste: “If fraud enters the system as a learning signal, advertisers are not only losing budget, but they are teaching campaign algorithms to optimize toward the wrong outcomes as well.”
Performance Max buyers are already running into it. An Adweek investigation found one ecommerce brand had 30 to 40 percent of its Pmax display and video ads land on MFA sites and dead domains, many running cookie cutter templates filled with AI-written text. The brand did not catch this through Google’s own reporting. It found out from DeepSee.io, a third-party measurement firm, which separately told Adweek the count of AI-generated MFA sites had more than doubled in a matter of months. The ads were running through Google video partners, a network Google describes as meeting high quality standards.
Generative tools removed the bottleneck on the supply side. MFA operators used to need writers and editors to produce enough pages to qualify for ad networks. Industry groups including the Association of National Advertisers have tracked MFA sites for years using criteria like heavy ad to content ratios, auto-refreshing placements, and templated layouts, but the pool of qualifying sites has widened now that one operator can spin up hundreds of them with AI-written copy and no editorial staff at all.
What to check in your own plan. None of this means pulling spend out of automated buying. It means not treating a platform’s own quality claims as the only check on where that spend lands. Independent verification tools, domain-level exclusion lists reviewed on a schedule rather than once, and a habit of spot-checking actual placements are the practical offset while the industry works out a fix for AI-generated supply.
Sources referenced in reporting: Spider Labs, 2026 Ad Fraud White Paper, Adweek, 40% of an Ecommerce Brand’s Pmax Buys Were Served on AI-spun MFA Sites and Dead Domains, Fraud Blocker, Made-For-Advertising Websites Are Still Harming Advertisers.