Amazon Faces a $20 Billion FTC Claim Over Hidden Ad Fees

Amazon is fighting an FTC lawsuit that alleges hidden advertising surcharges cost more than 1.2 million customers over $20 billion. The August 31 complaint targets the auction machinery behind Amazon’s sponsored listings, not merely a confusing fee or isolated billing error. Amazon denies deceiving advertisers and says its relevance-based system produced better results while lowering average winning bids.

That disagreement creates a sharp conflict over how an online marketplace should price access to its shoppers. The FTC and 22 states describe a covert effort to extract more money from advertisers without changing their stated bids. Amazon says the government misunderstands real advertising markets, where customers continually adjust bids according to performance.

The Amazon story circulated through technology feeds, including an Amazon RSSHub listing of Bloomberg’s coverage. The underlying dispute is broader than any newsletter summary. It reaches into Amazon’s fast-growing advertising operation, merchant economics, consumer prices, and the transparency expected from automated auctions.

The case also extends a long-running regulatory fight. A separate federal antitrust case already challenges Amazon’s treatment of sellers, product listings, and marketplace competition. This newer lawsuit gives regulators a narrower target: specific auction mechanisms, internal records, measurable charges, and advertisers who allegedly paid the difference.

The FTC Says Amazon Secretly Repriced Ad Auctions

The lawsuit alleges Amazon preserved the appearance of a conventional auction while quietly changing what winning advertisers paid.

Amazon sells sponsored placements that appear beside shopping results on its website and mobile app. Advertisers submit cost-per-click bids, meaning they offer to pay up to a stated amount when a shopper selects an ad. Amazon ranks eligible advertisements using both the bid and the product’s relevance to the shopper.

The dispute concerns how Amazon calculates the winner’s final charge. In a traditional second-price auction, the winner generally pays slightly more than the next-highest qualifying bid. The model encourages participants to bid near their true maximum because winning does not automatically require paying that full amount.

According to the federal complaint, Amazon introduced mechanisms that pushed charges above the price set by competing advertisers. Regulators characterize those increases as undisclosed surcharges. They allege Amazon tested, adjusted, and expanded the mechanisms over more than seven years.

The complaint identifies two practices, internally called Project Green and Project Red Panda. Regulators say Project Green used an invented auction participant to raise the price paid by the actual winner. The alleged virtual bid did not represent another advertiser seeking the placement.

Project Red Panda allegedly changed how Amazon translated relevance into the final auction price. The FTC says Amazon used the system to charge winners closer to their maximum bids. Advertisers allegedly lacked the information needed to determine whether competition or Amazon’s own adjustment produced a given price.

These claims remain allegations, and Amazon will have an opportunity to challenge them in court. The complaint nevertheless presents the dispute as intentional conduct rather than an accidental design flaw. It cites internal communications that allegedly described the additional charges as hidden, nontransparent, or beneficial to Amazon at advertisers’ expense.

The FTC says the practices affected almost 1.2 million advertising customers from 2018 through 2026. That group allegedly included more than 500,000 small and medium-sized businesses. The agency estimates the disputed mechanisms extracted more than $20 billion.

The regulator also alleges that Amazon’s own sales and marketing employees did not fully understand the pricing changes. If established, that point would weaken an argument that ordinary advertiser communications provided adequate disclosure. A seller cannot evaluate an auction rule that even customer-facing employees cannot clearly explain.

The case was filed in the US District Court for the Western District of Washington. The FTC seeks monetary relief and an injunction against practices the court finds deceptive or unfair. The 22 participating states make the case geographically broader and politically harder to dismiss as one agency’s policy preference.

The action does not establish that every advertiser suffered a loss equal to every added charge. Litigation must still resolve liability, causation, calculation methods, and appropriate remedies. However, the alleged scale transforms an auction-design dispute into a material challenge for Amazon.

Why Amazon’s Advertising Business Is Under Pressure

Regulators are attacking a business that has become central to Amazon’s growth and to sellers’ ability to reach customers.

Amazon reported $68.64 billion in advertising services revenue for 2025, up from $56.21 billion during 2024. Its annual filing says those services include sponsored ads, display advertising, and video advertising. Revenue is recognized as impressions or clicks are delivered.

That expansion matters because advertising connects several sides of Amazon’s business. Shoppers arrive with purchasing intent. Sellers compete for visibility around relevant searches. Amazon controls the marketplace, the ranking system, the auction, and much of the performance data used to judge results.

This structure gives Amazon information that an individual advertiser cannot reproduce. The company sees competing bids, shopper behavior, placement availability, conversion patterns, and auction outcomes across the marketplace. A seller sees its own campaign settings and the reports Amazon chooses to provide.

Information asymmetry is not automatically unlawful. Advertising platforms routinely use private ranking signals, fraud controls, quality scores, and pacing systems. Those systems can improve relevance and prevent the highest bidder from placing an unsuitable advertisement.

The legal risk grows when a platform represents one pricing method while allegedly applying another. Auction mechanics determine whether an advertiser’s maximum bid acts as a ceiling, a strategic signal, or the price itself. A hidden platform-generated adjustment can change how safely an advertiser interprets that number.

Small businesses face particular pressure because advertising is often tied to discoverability. A merchant can improve its listing, product quality, and fulfillment, yet still struggle when sponsored placements dominate valuable search positions. Pulling back from advertising can mean surrendering traffic to better-funded competitors.

That dependence complicates Amazon’s claim that customers could respond to performance. Advertisers can lower bids, pause campaigns, or shift budgets. However, sellers that rely on Amazon for distribution may have fewer practical alternatives than large brands operating across many retailers.

The FTC argues that added advertising costs can also reach shoppers. A seller facing higher customer-acquisition costs can accept a smaller margin, reduce other spending, or raise product prices. The exact response differs by category and seller, so broad consumer harm will require evidence rather than assumption.

The complaint arrives as retail media expands beyond sponsored product listings. Major retailers now use purchasing data to sell ads across search pages, websites, streaming services, and third-party inventory. Amazon’s enormous transaction dataset gives it a strong position in that market.

Walmart, Target, and Instacart also operate retail-media businesses. Google and Meta remain major destinations for performance advertising, but they do not control Amazon’s marketplace checkout in the same way. Amazon can connect an advertisement with a purchase inside one commercial system.

That advantage raises the stakes of the lawsuit. A ruling about hidden auction adjustments could influence how other retail platforms disclose pricing logic. It could also encourage advertisers to demand clearer records showing how bids, relevance, and platform adjustments created each charge.

The immediate pressure falls on Amazon’s legal and advertising teams. The longer-term pressure falls on the business model itself. Amazon must defend both its conduct and the idea that better campaign performance can answer concerns about pricing transparency.

Amazon Says Relevance Saved Advertisers Money

Amazon’s defense separates auction mechanics from advertising value, arguing that outcomes matter more than simplified descriptions of bid pricing.

In its Sponsored Ads response, Amazon calls the lawsuit misguided and denies running a companywide effort to deceive customers. It says the FTC selected a small number of simplified communications from roughly 1.5 million reviewed pages.

Amazon agrees that relevance affects which ad wins. It argues that this feature benefits advertisers and shoppers because the highest bid does not necessarily represent the most useful product. A relevant lower bidder can win over a less relevant higher bidder.

The company says approximately 92% of placed advertisements do not go to the highest bidder. It also claims average winning bids for Sponsored Products search ads fell 50% from 2019 through 2025. Those figures are company estimates and have not been adjudicated.

Amazon further estimates that its relevance-based system saved advertisers more than $8 billion between 2021 and 2025. That calculation assumes advertisers respond to actual campaign performance, not merely to a published description of auction rules.

For 2026, Amazon projects advertisers will generate at least 58% more sales and achieve at least 46% better return on advertising spending under its model. It compares those outcomes with a system that ranks advertisements using bid amounts alone.

These claims address an important economic question. A higher charge for a more productive placement can still leave an advertiser better off. If improved relevance increases conversion enough, the campaign can generate more profit even when individual clicks cost more.

However, that does not resolve the FTC’s legal theory. Performance and disclosure are separate questions. An auction can produce valuable results while giving participants an incomplete account of how the platform sets their charges.

Amazon also says advertisers never pay more than their submitted bids. That is a meaningful boundary, but it does not settle whether they paid more than disclosed auction competition required. A maximum bid is permission to charge up to a limit, not necessarily consent to every undisclosed calculation below it.

The parties therefore disagree about the correct counterfactual. The FTC compares actual charges with prices that competing bids allegedly would have produced without Amazon’s adjustments. Amazon compares its system with auctions that give less weight to relevance and potentially deliver worse advertising results.

Both comparisons can produce large figures without directly answering the other side. The FTC’s $20 billion estimate measures allegedly improper surcharges. Amazon’s $8 billion estimate measures benefits it attributes to relevance-based allocation. Courts will need to examine assumptions behind both calculations.

The evidence will likely include experiments, technical documentation, advertiser disclosures, employee communications, and auction records. Experts may reconstruct prices under alternative mechanisms. They will also need to separate legitimate relevance adjustments from alleged price increases unrelated to another advertiser’s bid.

This is the central tradeoff in the case. Amazon wants freedom to optimize a complex marketplace around predicted usefulness and campaign performance. Regulators insist that optimization cannot conceal material information about what customers pay.

The conflict is not simply Amazon versus an old auction model. It is Amazon’s outcome-based defense against the FTC’s disclosure-based accusation. That distinction will shape discovery, economic testimony, and any eventual settlement or judgment.

The $20 Billion Claim Depends on Auction Counterfactuals

The headline number is consequential, but proving it requires more than adding every disputed adjustment across billions of auctions.

The surcharge allegations describe likely extraction exceeding $20 billion. New York’s attorney general uses a more direct formulation, saying advertisers were overcharged by that amount. The complaint itself must support the figure through evidence acceptable to the court.

Calculating harm requires a baseline. Regulators need to show what advertisers would have paid under the represented auction rules. They must then compare that amount with actual charges caused by the challenged mechanisms.

That process becomes difficult when relevance determines both eligibility and ranking. A lower-priced advertisement might never have won without a relevance adjustment. Removing one component can change the winner, not just the final price.

Advertiser behavior also changes over time. Campaign managers raise and lower bids based on conversion, inventory, competition, seasonal demand, and automated recommendations. A pricing change can cause later bid changes, which then alter future auctions.

Amazon relies heavily on this dynamic response. The company argues that advertisers optimize for results and would not keep bids fixed while campaign economics changed. The FTC must show why its counterfactual appropriately accounts for that behavior.

Automation adds another layer. Many merchants use Amazon’s dynamic bidding controls or third-party software rather than setting every bid manually. Those systems react to performance data, but they depend on signals produced by Amazon’s marketplace.

A campaign can appear healthy at the aggregate level while individual auction prices remain difficult to audit. Return on advertising spending combines prices, clicks, and attributed sales. It does not reveal whether a platform-created bidder raised a particular charge.

Attribution can also complicate the defense. Amazon has substantial control over which sales receive credit for an advertisement. Better reported returns can reflect stronger relevance, different attribution, increased consumer demand, or several factors at once.

None of these complications invalidate the complaint. They show why internal documents matter. A message describing a deliberate surcharge or an effort to prevent detection can reveal intent that aggregate campaign data cannot.

Context will matter too. Engineers often use shorthand when discussing experiments. A provocative phrase in an internal document does not automatically prove the company adopted an unlawful scheme. The court will examine who wrote each statement, what system it described, and how that system operated.

The FTC’s case appears stronger if Amazon’s external explanations remained materially inconsistent with internal pricing rules. It becomes weaker if disclosures adequately described flexible pricing and advertisers understood how relevance influenced payment.

Amazon’s defense grows stronger if campaign records demonstrate broad, durable gains tied to the disputed mechanisms. It becomes weaker if those gains came from unrelated improvements or if the alleged surcharges can be isolated without reducing relevance.

The number of affected advertisers presents another challenge. Harm can vary widely across sellers, categories, keywords, and time periods. A large consumer brand with an advertising team occupies a different position from a merchant launching its first sponsored campaign.

Any monetary remedy may require grouping customers or calculating individualized amounts. The court could also focus on disgorgement, restitution, civil penalties, or prospective restrictions, depending on which claims survive and what the law permits.

Readers should therefore treat $20 billion as the government’s allegation, not an established judgment. They should apply the same caution to Amazon’s claimed $8 billion in savings. Both figures express opposing models of what would have happened under different rules.

That uncertainty is not a reason to dismiss the case. It identifies the real evidentiary contest. The lawsuit will test whether Amazon can document pricing choices clearly enough to overcome internal language that regulators portray as evidence of concealment.

This Is Amazon’s Latest Clash With the FTC

The advertising case narrows a much wider argument about whether Amazon uses marketplace control to extract value from dependent businesses.

The FTC and state attorneys general sued Amazon in 2023 over alleged monopolization of online superstores and marketplace services. That case challenges several business practices, including seller conditions, product visibility, and mechanisms regulators say suppress price competition.

The newer case does not require the government to prove every part of that broader monopoly narrative. It focuses on representations to advertising customers and the mechanics behind specific charges. A court can evaluate alleged deception even while wider antitrust questions remain contested.

Still, the two cases reinforce each other’s stakes. Amazon operates the marketplace where many sellers reach customers. It also sells fulfillment, collects referral fees, ranks products, and auctions prominent search placements.

A merchant can choose whether to advertise in a formal sense. In practice, the visibility of sponsored results can make that choice commercially difficult. The FTC uses this dependence to explain why hidden pricing changes would carry substantial consequences.

Amazon disputes the idea that its services trap or harm sellers. It argues that businesses choose advertising because it generates measurable value. The company can point to competition from other marketplaces, retailer networks, search engines, and social platforms.

The legal theories are nevertheless distinct. The 2023 action concerns market power and exclusionary conduct. The 2026 action accuses Amazon of unfair and deceptive auction practices. Success or failure in one case will not automatically determine the other.

The government’s history with Amazon extends beyond these proceedings. The company agreed to resolve earlier FTC cases involving Prime enrollment and cancellation practices, privacy at Ring, and children’s voice data handled through Alexa. Each matter involved different conduct and legal standards.

That record makes the advertising lawsuit part of an institutional relationship, not a one-off disagreement. Regulators increasingly scrutinize the interfaces, algorithms, and defaults through which large platforms convert user dependence into revenue.

The current case is especially significant because advertising has become one of Amazon’s largest reported service categories. Its 2025 advertising revenue exceeded the company’s subscription-services revenue. Weakening advertiser confidence could affect a growth engine that reaches across retail and media.

A mandated disclosure remedy could force Amazon to explain price calculations more clearly. A stronger remedy could restrict specific auction adjustments or require records that let advertisers reconstruct charges. Monetary relief would raise separate questions about eligibility and distribution.

Amazon might also change disclosures without conceding wrongdoing. Platforms often revise interfaces, documentation, or reporting while litigation continues. Such changes can reduce future risk but create debate about whether earlier explanations were sufficient.

Competitors will watch closely. Retail-media networks use varied combinations of first-party shopping data, relevance scores, reserve prices, and bidding systems. A detailed ruling could establish expectations that extend beyond Amazon’s own marketplace.

Advertisers should resist reading the dispute as proof that all algorithmic auctions are manipulated. Automated auctions require hidden signals because ranking depends on fraud detection, predicted engagement, relevance, and limited inventory. The issue is whether material pricing rules were misrepresented.

They should also resist assuming that good performance eliminates disclosure duties. Customers need enough information to set rational bids and compare channels. A platform’s claim that advertisers benefited does not grant unlimited discretion over price formation.

For knowledge workers tracking the litigation, source discipline matters. Saved newsletters, court filings, company responses, and financial records can quickly become difficult to reconcile. A searchable personal knowledge base can preserve the distinction between allegations, company estimates, and established findings.

That distinction will remain essential throughout the case. Headlines compress the conflict into $20 billion versus $8 billion. The litigation will turn on mechanisms, representations, evidence, and legal definitions that do not fit into either number.

Three Signals Will Show Which Side Is Winning

The next meaningful developments will come from the courtroom, Amazon’s advertiser disclosures, and measurable changes in campaign economics.

The first signal is the federal court’s treatment of Amazon’s initial response. Amazon can seek dismissal of some or all claims before full discovery. A ruling that preserves the core deception theory would strengthen the government’s position and expose more internal records to examination.

A ruling narrowing the case would weaken the FTC’s leverage, especially if the court rejects its theory of auction harm. The precise reasoning will matter more than whether either party describes the result as a victory.

The second signal is Amazon’s treatment of auction transparency. Advertisers should watch for revised help pages, clearer price explanations, new reporting fields, or controls that separate relevance effects from pricing adjustments.

More detailed disclosures would not prove the complaint’s allegations. They would show that Amazon sees value in reducing uncertainty around auction charges. No meaningful change would suggest the company remains confident that existing explanations are sufficient.

The third signal is advertiser performance data. Amazon’s 2026 claims about higher sales and improved return on advertising spending create a testable defense. Independent agencies and large advertisers can compare cost-per-click trends, conversion rates, attributed sales, and placement quality.

Those comparisons need careful controls. Category demand, competition, inventory, seasonality, and attribution changes can move every metric. Evidence becomes more persuasive when similar campaigns show consistent results across long periods and advertiser types.

Merchants should preserve campaign exports and change logs now. Historical records can help them understand whether bid increases came from their own decisions, automated tools, rising competition, or platform-level pricing changes. Aggregate dashboards alone may not provide enough detail later.

Advertisers should also review how maximum bids are chosen. A maximum should reflect the value of a click under realistic conversion and margin assumptions. Treating it as a harmless ceiling becomes risky when the pricing mechanism remains disputed.

The case does not require businesses to abandon Amazon Ads. Sponsored placements can still generate profitable sales, and Amazon’s shopping audience remains valuable. The practical response is closer measurement, more disciplined experimentation, and clearer documentation of bidding decisions.

Large brands can compare Amazon campaigns with Walmart, Google, Meta, and other retail-media channels. Smaller sellers may have fewer alternatives, but they can still separate branded and nonbranded campaigns, test lower bid limits, and monitor changes in incremental sales.

Consumers should watch the case because marketplace advertising affects what they see. A product’s position can reflect sponsorship, relevance, bid strategy, organic performance, or several signals together. Pricing pressure on sellers can also influence product availability and retail prices.

The FTC must ultimately convert striking internal language and a large estimate into admissible proof. Amazon must explain why the challenged mechanisms were fair, adequately disclosed, and beneficial without treating performance as a substitute for transparency.

That is why this lawsuit is more than another regulatory clash. It asks whether a platform can defend opaque price formation by showing that its optimization produced better outcomes. The answer will shape Amazon’s ad business and the wider retail-media market.

Follow the docket, disclosure changes, and independent campaign data rather than either side’s headline number. Those signals will reveal whether the Amazon advertising case becomes a narrow communication dispute or a lasting constraint on platform-run auctions.