Google asked a federal appeals court on May 22, 2026 to reverse the district court ruling that found it had illegally maintained a monopoly in online search, arguing that the trial judge applied the wrong legal standard at nearly every stage of the case.

The brief is the company's most forceful attempt yet to undo a judgment that has already begun to change how it operates. At stake is not only a finding of liability under the Sherman Act, but a six-year set of remedies that requires Google to hand competitors access to assets it spent more than two decades and billions of dollars building. The 100-page opening brief, filed with the United States Court of Appeals for the District of Columbia Circuit, was submitted by counsel from Munger, Tolles & Olson, Williams & Connolly, and Ropes & Gray. It certifies a length of 20,787 words. The appeal is docketed as Case No. 26-5023, consolidated with two related case numbers.

The filing is the next formal step in litigation that began in October 2020, when the Department of Justice and a group of states sued Google over its search distribution agreements. Google filed its notice of appeal on January 16, 2026, challenging the December 2025 final judgment entered by U.S. District Judge Amit P. Mehta. The May 22 brief now lays out the substantive arguments the company will press before the appellate panel - and it does so with an unusual rhetorical strategy. Rather than disputing the trial court's account of the facts, Google largely embraces it, then argues that those very findings, read against the correct legal rules, require a verdict in the company's favor.

The brief opens by invoking a warning from Judge Learned Hand, written in 1945, that "[t]he successful competitor, having been urged to compete, must not be turned upon when he wins." That sentence sets the tone for everything that follows. Google's case, distilled, is that it won the search market by building a better product, that antitrust law is supposed to reward exactly that behavior, and that the district court mistook the company's success for proof of wrongdoing.

US v Google - Alphabet brief - 20260522
US v Google - Alphabet brief - 20260522.pdf
1 MB
download-circle

What Google is arguing

The brief organizes the company's case into a set of distinct contentions, each presented as independently sufficient to overturn part or all of the judgment. In summary, Google argues that:

  • The district court used the wrong legal test for exclusionary conduct, applying a causation standard - whether conduct "reasonably appears capable" of maintaining monopoly power - to decide whether the conduct was unlawful in the first place.
  • The browser agreements reflected competition on the merits, because the trial court's own findings show Apple and Mozilla chose Google for its quality and monetization, not because Google blocked rivals.
  • The agreements were not exclusive dealing arrangements, since they contained no express exclusivity provision and left browser-makers free to distribute and promote competing search engines.
  • The Android agreements alone cannot sustain liability, because they accounted for a small share of queries and the court's decision depended on finding all the agreements unlawful.
  • The court defined the relevant market too narrowly by excluding specialized vertical providers, and a properly defined market would show Google lacks monopoly power.
  • The remedies were imposed without the causation findings required to connect Google's conduct to its market position, leaving no measure of whether the relief was proportionate.
  • There is no legal basis for extending remedies to generative AI companies, because such products did not exist during the challenged conduct and operate in a market the court itself called highly competitive.

The sections below set out how the brief develops each of these points, and the picture of the search market that Google asks the appeals court to accept.

The market as Google describes it

Before reaching its legal arguments, the brief spends considerable space describing how search works as a business. The account matters, because Google's legal claims rest on it.

A general search engine, in the brief's framing, is one that attempts to answer every kind of query. Google Search handles billions of queries each day in nearly 150 languages, free of charge to users. Microsoft's Bing, Yahoo, and DuckDuckGo are also general search engines. The brief is candid about Google's dominance, but frames it as long-standing and earned: by 2009, well before any conduct challenged in the case, roughly 80% of all U.S. search queries flowed through Google, and by the time the lawsuits were filed in 2020 that share had reached 89.2%.

The brief leans on the district court's own characterizations. It quotes findings that Google "has long been the best search engine, particularly on mobile devices," that it is "widely recognized as the best GSE available in the United States," and that its "superior product quality rests in part on its numerous innovations." The company, the trial court found, "hired thousands of highly skilled engineers, innovated consistently, and made shrewd business decisions." It also "foresaw that the future of search was on mobile," while Microsoft "was slow to recognize" that shift and "has been trying to catch up - unsuccessfully - ever since." The brief adds a detail meant to underline the point: one of the most popular queries on Bing is "Google.com."

The economics of search rest on a narrow slice of activity. Search engines earn money only by attaching advertising to commercial queries - what the industry calls monetizing a query. According to the brief, only about 20% of queries on Google carry commercial intent, such as "budget-friendly Chicago hotel" or "best blender." The other 80%, including searches like "how to perform CPR," are not monetized at all. Yet that commercial fifth is enormously valuable: the brief states that Google earned $146 billion in worldwide advertising revenue in 2021.

This is where the brief introduces the players it says the district court wrongly ignored. Competing for those commercial queries are not only other general search engines but also specialized vertical providers - platforms that answer queries on a particular subject. The brief lists Amazon and eBay for retail goods, Booking.com and Expedia for travel, and Yelp for restaurants and hotels. It quotes the trial court's acknowledgment that "[n]o one disputes that an SVP can serve the same purpose as a GSE for an individual query on a particular subject matter," and that a user can turn to either Google or OpenTable to find a Japanese restaurant, or Google or Amazon to shop for a blender. The brief notes that Google itself "views competition from SVPs as 'intense for commercial clicks,'" and that the company built shopping, flights, and hotels verticals, and developed product listing ads, partly to compete with those rivals.

How search reaches users

The brief then walks through how people actually arrive at a search engine, because the distribution agreements at the center of the case govern exactly that.

A user can type a search engine's web address directly. On mobile devices, users can open a dedicated search application. Android devices carry a search widget on the home screen. And browsers integrate search into their navigation bars - Safari on Apple devices, the combined address-and-search box in Chrome, the integrated bar in Microsoft's Edge.

Because users expect browser search to work without setup, the brief states, "a 'longstanding industry practice [is] preloading a browser with a default GSE,'" and "all browsers in the United States are so designed." Browser-makers that operate their own search engines - Microsoft, DuckDuckGo, and Google itself - set their own engine as the default. The brief offers Windows as an illustration of how little a default settles: Windows PCs ship only with Edge, which defaults to Bing, yet Google's search share on Windows devices is 80%, the brief says, largely because users download Chrome for its functionality. Browser-makers without their own search engine - Apple and Mozilla - instead sell the default position as a form of promotional placement. The brief stresses that defaults are "merely preconfigured settings" and that on all major browsers a user can change the default in a few steps.

The brief also presents the default decision as a deliberate product choice by Apple, not something Google imposed. It quotes Apple executive testimony that Apple "buil[t] the Safari browser to have a single default search engine out-of-the-box," that this was "a product design decision that Apple carefully considered," and that Apple applied it across phones, tablets, and computers because "why would we give an inferior experience on a Mac versus a phone."

How Google describes the browser agreements

A large portion of the brief is devoted to the contracts at the heart of the case: the agreements that made Google Search the out-of-the-box default in the Safari and Firefox browsers, and the Android distribution agreements with device manufacturers and carriers.

The logic of the deals, as the brief presents it, is one of mutual need. A browser-maker without its own search engine has query traffic it cannot monetize; a search engine can only monetize the queries it receives. Neither can profit alone. So they strike a deal: the browser sends query traffic to the search engine through a default, the search engine monetizes those queries, and the two share the revenue.

The Apple relationship has the longest history. The brief traces it to a 2002 Internet Services Agreement under which Apple made Google Search available on Safari in exchange for a license, with no default placement and no revenue sharing. In 2005, the arrangement changed: Apple agreed to make Google the out-of-the-box default in exchange for a share of the resulting revenue. Google and Apple entered the current version of the agreement in 2016 and renewed it in 2021, with the renewed term running through 2026. Under its terms, the agreement defines the "Default" search engine as the one that "will automatically be used for responding to Search Queries initiated from the Web Browser software, unless the End User selects a different third-party search service." Google pays Apple a percentage of the revenue from queries made through that default - the brief describes it as "a very large payment in absolute terms, but one commensurate with the very large number of queries Apple sends to Google."

Google's framing is that browser-makers chose the company on the merits. The brief points to the district court's findings that Apple and Mozilla "value [Google's] quality" and concluded that Google's "search engine provides the best bet for monetizing queries." It cites trial testimony describing Google as a "no brainer" because it had "the best search engine," and a 2021 Apple internal study that found Google Search "superior to Bing on all search access points." As one Apple executive put it in testimony quoted in the brief, "we have to pick what's best for our customers, and today, that is still Google."

The brief argues that Microsoft's search engine was a poor commercial alternative for Apple. Even though Microsoft offered Apple 100% of search advertising revenue if Bing became the default, Google says Apple expected to earn less, because users would abandon a Bing default and route their queries back to Google. The brief quotes the district court's finding that there was "no price that Microsoft could ever offer [Apple]" to make Bing the default that would be more profitable. Apple, the filing states, viewed Bing as "horrible at monetizing advertising" and a "business risk."

The agreement with Apple has long been the largest single deal in the dispute. Court documents in the earlier proceedings showed Google paid an estimated $20 billion to Apple in 2022 for default placement on Safari, and the scale of that arrangement made Apple a participant in the remedies fight over how the deal might be unwound.

The Mozilla account follows a similar arc, and the brief presents it almost as a natural experiment. According to the filing, Mozilla and Google agreed to make Google the default for the Firefox search bar in exchange for a share of advertising revenue. A former Mozilla chief executive testified that the company chose Google between 2004 and 2014 because "Google was way ahead" of the competition. In 2014, Mozilla switched the default to Yahoo for a three-year term. It found "almost immediately," the brief says, that "users did not like the Yahoo! Search experience," that search volumes declined as people shifted away from the default, and that Yahoo responded by increasing the number of ads it placed, degrading the experience. Mozilla reverted to Google, the brief states, because "our users made it clear that they look for and want and expect Google." Since then, the brief says, Mozilla has periodically tested switching the default and consistently found that within days a substantial number of users abandoned the alternative for Google; its "takeaway was that switching the Firefox default to Bing would result in missing revenue targets." The relative scale of that relationship is far smaller than Apple's; during the remedies phase, Mozilla's finance chief testified about the browser ecosystem risksof cutting off Google payments estimated in the hundreds of millions of dollars annually.

The brief also describes the Android agreements. Under a Mobile Application Distribution Agreement, a device maker such as Motorola or Samsung pays no fee for a license to certain Google applications and agrees to preload and place specified Google apps, including Chrome and the Google Search Widget. Separate Revenue Share Agreements with manufacturers and wireless carriers provide for Google to pay for the promotion of Search and Chrome on the device.

From these findings Google draws a legal conclusion: the agreements reflected what the brief calls "competition on the merits," not exclusion. The company also argues that the agreements were not exclusive dealing arrangements at all, noting the district court's acknowledgment that they contained "no express exclusivity provision" and that "[u]sers are free to navigate to Google's rivals through non-default search access points." Apple, the brief notes, sells homepage bookmark placement to Bing and Yahoo in exchange for revenue share, and users who select a different engine can make it their default.

Google's central argument concerns how courts are supposed to identify illegal conduct under Section 2 of the Sherman Act. The company contends that the trial court collapsed two separate legal questions into one, and that the error tainted the entire liability finding.

The brief lays out the structure it says the law requires. Monopolization has two elements: possessing monopoly power in a relevant market, and the willful acquisition or maintenance of that power "as distinguished from growth or development as a consequence of a superior product, business acumen, or historic accident." The second element, the brief continues, itself has two parts: the conduct must be exclusionary, and there must be a causal link between that conduct and the maintenance of the monopoly.

The brief draws a sharp line between two kinds of behavior. Antitrust law, it says, distinguishes between "making a better offer to a consumer" and "depriving a consumer of the freedom to accept a better offer." The first is competition on the merits and cannot be exclusionary, however much it costs rivals their sales. The second can harm the competitive process. The brief quotes the 2001 Microsoft decision that "offering a customer an attractive deal is the hallmark of competition" and that "the antitrust laws do not condemn even a monopolist for offering its product at an attractive price." It also borrows a phrase from a 2013 opinion written by then-Judge Neil Gorsuch: "Competition, after all, is a ruthless process." The point, in the brief's telling, is that every sale won is a sale a rival loses, and harm to competitors alone never proves an antitrust violation.

Against that backdrop, Google says the district court used the wrong test. The brief states that the trial judge treated the "key question" for identifying exclusionary conduct as whether Google's agreements "reasonably appear capable of significantly contributing to maintaining Google's monopoly power." That test, Google argues, was lifted from Microsoft - but from the wrong part of it. In Google's reading, the "reasonably appears capable" standard exists only to assess causation, the second step, after a court has already found that conduct was exclusionary. Using it to decide whether conduct was exclusionary in the first place, the brief says, "should be obvious" as error, because it "would condemn any monopolist for succeeding on the merits by offering the best product at the best price - the very thing the antitrust laws seek to encourage."

That is not a minor technical complaint. If the appeals court agrees, the liability finding could fall on that ground alone, before the panel reaches any factual dispute about the distribution deals themselves.

The arithmetic of the queries

The brief reinforces the argument with a breakdown of where U.S. search queries actually go, framing it as a demonstration that, under the correct standard, the district court's own findings require judgment for Google.

By the brief's account, slightly more than half of all U.S. queries to general search engines flow through channels where no party claims Google did anything wrong. Another 28% flow through Safari or Firefox - the browser agreements that Google says were lawful competition on the merits and were not exclusive in any event. The remaining 19% are made on Android devices. The Android agreements, the brief argues, were never a freestanding basis for liability and cover too few queries to sustain the case on their own. The district court's decision, in other words, depended on every category of agreement being found unlawful.

The exclusive dealing question

Google's third line of attack is that the browser agreements were not the kind of contract that exclusive dealing law even reaches. The brief argues that "[t]he 'prerequisite to any exclusive dealing claim' is 'an agreement to deal exclusively,'" and that an exclusive dealing arrangement is one that "forbids the buyer from purchasing the contracted good from any other seller."

Google's agreements, the brief contends, did no such thing. Apple and Mozilla remained free to promote and meet user demand for rival search engines, and they did. The brief returns to two of the district court's own findings - that the agreements had "no express exclusivity provision" and that "[u]sers are free to navigate to Google's rivals through non-default search access points" - and argues those facts "should have ended the exclusivity inquiry."

The trial court instead found the agreements exclusive on a de facto basis, reasoning that users overwhelmingly ran queries through the preset Google default. Google's answer is that users did so because they preferred Google, a preference the court's own findings confirm, and that user preference cannot be the basis for antitrust liability. The brief also notes a procedural point: it says the plaintiffs framed the case as an exclusive dealing case from the outset, and that only after trial did they ask the court to apply a broader Section 2 standard - a shift the district court rejected as an "[u]nexpected" and "dramatic post-trial shift."

The market definition argument

Google's next independent attack targets how the court defined the relevant market. Monopoly power, the brief notes, can only be measured against a properly defined market, and Google contends the trial court drew the boundaries too narrowly.

The district court placed specialized vertical providers outside the general search market. According to the brief, the court did so because those providers use different "business models" to attract users, even after finding that they "respond to queries" and are useful for commercial searches.

Google argues that this reasoning inverts settled antitrust principle. Market definition, the brief states, must turn on "commercial realities" and on which products are "reasonably interchangeable" by consumers for the same purpose. Because the trial court found that users can turn to vertical providers for commercial queries, and that this option has pressured Google to improve quality, Google contends those providers belong inside the market. In a market defined that way, the company says, it lacks monopoly power - which would independently require reversal. The original August 2024 liability ruling found Google held roughly 90% of online search and that its agreements foreclosed about half the general search services market by query volume.

Challenges to the remedies

Even if the liability finding survives, Google argues, the remedies must be vacated. The brief frames this as a separate failure of legal method rather than a disagreement about policy.

Antitrust remedies, Google contends, must be tied to a causal connection between the unlawful conduct and the defendant's market position. The brief argues that the district court made no findings - "none, at any stage" - about how much of Google's position came from the challenged agreements as opposed to what the company calls its lawful advantages, business foresight, superior product, and superior monetization. Without such a measurement, Google says, the court had no yardstick for deciding whether forcing the company to aid competitors was proportionate. A remedy, the brief argues, "should be tailored to fit the wrong," and the court cannot order relief beyond stopping the unlawful conduct without first weighing that causal connection.

The specific remedies at issue are significant. The final judgment, the bulk of which took effect on February 3, 2026 and runs for six years, requires Google to make a one-time disclosure of its search index to so-called Qualified Competitors. The brief describes that index as a database of publicly available web pages "critical to returning high-quality search results," built with what an earlier finding called "billions of dollars" for "just the technical infrastructure." The judgment also requires Google to transfer user-side interaction data on a rolling basis - signals such as which links a user clicks, how long the user hovers over a result, and how quickly the user clicks back. A separate provision compels Google to syndicate its search results, and its search text ads, to competitors for five years, so that rivals can monetize their own offerings using ads from Google's advertiser network. The forced disclosure of Google's RankEmbed model and associated training data was among the most contested elements of the September 2025 remedies decision.

A Technical Committee was established to resolve the unsettled particulars of the syndication and data-transfer requirements. The brief notes that the district court denied Google's motion to stay those remedies as premature, reasoning that license terms, security and privacy safeguards, and the identity of Qualified Competitors were "far from established." Google had earlier asked the court to pause the data-sharing requirements while the appeal proceeds.

The brief is careful to distinguish what Google did and did not propose during the remedies phase. According to the filing, Google itself offered remedies that would bar it from distribution agreements containing terms the court deemed exclusive, and that would expressly let browser-makers set another general search engine as the default for certain devices or modes. What Google opposed, the brief says, was relief designed to "affirmatively aid rivals or hobbling Google." The company also notes that the trial court rejected some of the plaintiffs' proposals, including a ban on payments to distribution partners, on the ground that cutting off those payments would impose "substantial - in some cases, crippling - downstream harms" on partners and consumers. The plaintiffs had pressed for far more aggressive relief, with the Department of Justice's final remedies proposal seeking a Chrome sale and a ban on search-related payments.

The fight over generative AI

The most forward-looking section of the brief concerns generative AI products, and it is here that the appeal reaches beyond traditional search.

The district court decided that generative AI companies which "plan to invest and compete in or with" the general search market could qualify as Qualified Competitors entitled to Google's data and syndication. Google argues there is no legal basis for that extension, calling that portion of the injunction "pure prospective regulatory policymaking in a new and vibrant market, untethered to any liability determination."

The brief states that generative AI products "did not even exist at the time of Google's challenged conduct," that no one claims Google holds a monopoly over them, and that the trial court itself described the generative AI space as "highly competitive," with "numerous new market entrants" and "constant jockeying" for the lead in quality. The filing cites a figure from the remedies record: a finding that a leading competitor, OpenAI, reported a market share of approximately 85%. Google's point is that a market where another firm holds the dominant share cannot be one where forcing Google to share assets corrects a competitive wrong it caused. The brief also notes the trial court's observation that generative AI products from companies other than Google had already found success obtaining distribution with device makers and other companies.

The brief reaches back to the Microsoft remedy of two decades ago to define the limit it says the district court ignored. In that case, the brief argues, relief was tied to restoring competition inside the monopolized market: courts ordered Microsoft to aid products resembling the "middleware" central to liability, and refused to extend that aid to products that were not similar. Here, by contrast, Google says the plaintiffs never argued and the court never found that Google had blocked generative AI products that could have helped other search engines compete. Even acknowledging that "GenAI products have emerged as a competitive threat to the traditional GSE," the brief argues, "turbocharging outside-the-market GenAI products was a remedy unrelated to any wrong found at the liability stage."

The brief closes that section with an argument about the limits of courts. It quotes the trial court's own description of "these fast-moving times, where GenAI technologies are breaking barriers seemingly at light speed," and argues that judges "are neither economic nor industry experts" and should not act as "central planners." Continuing supervision of a detailed decree in such a market, the brief warns, "could wind up impairing rather than enhancing competition." Whether antitrust relief should reach AI systems has been a recurring question since the remedies phase, when observers noted the case could reshape the artificial intelligence landscape well beyond search.

Procedural background

The two consolidated cases below were brought by the United States with 11 states, and separately by a group of 38 states. The matters were split into a liability phase and a remedies phase. A liability trial took place in 2023, after which the district court issued its findings of fact and conclusions of law. A remedies hearing followed in 2025, and the court entered final judgment in December 2025.

The appellate panel will review legal conclusions without deference, under the de novo standard, and will set aside factual findings only if they are clearly erroneous. Remedies are reviewed for abuse of discretion, though the brief notes the principle that a legal error is itself an abuse of discretion. The brief ends with a single-sentence request: that the district court's judgment be reversed. Oral argument has not yet been scheduled.

The case sits alongside a second antitrust defeat for Google. A separate federal court in Virginia found in April 2025 that the company had unlawfully monopolized digital advertising technology markets, a matter that proceeded to its own remedies proceedings later that year. The two proceedings together place sustained regulatory pressure on a single company across search and advertising.

Why this matters for marketers

For the advertising and search marketing community, the appeal keeps a wide set of questions unresolved rather than closing them.

The remedies that took effect in February 2026 touch the mechanics of how search results and search advertising are produced and distributed. Search syndication and the forced sharing of index and interaction data could, over time, support competing general search engines and AI-based answer products, changing where queries are answered and where advertising inventory sits. The September 2025 ruling also required Google to disclose changes to its advertising auctions, a transparency measure with direct implications for campaign management. If the appeal succeeds in narrowing or overturning the judgment, much of that potential change does not arrive.

The default-placement question carries its own weight. The agreements at issue governed roughly $26 billion in annual payments tied to default search placement, a figure reported across the earlier proceedings, with approximately $18 billion of that directed to Apple. The outcome of the appeal will influence whether those payment structures persist and on what terms - a matter that shapes the economics of mobile device distribution and browser revenue, and by extension the placement of search advertising that those defaults feed.

The market definition fight has consequences that reach beyond this single case. If the appeals court accepts that specialized vertical providers belong inside the relevant market, it would change how dominance is measured for platforms that compete with retailers, travel sites, and review services for commercial queries - the queries that carry advertising. A narrower or broader market definition shifts the baseline against which future enforcement is judged.

The generative AI dimension is the newest variable. The district court tied AI answer products into the remedy framework; Google's brief argues they do not belong there. The resolution of that specific dispute will help define whether antitrust remedies built around traditional search reach the AI systems that increasingly mediate how users find information, and whether those systems gain access to Google's index and data. Reaction to the original remedies was mixed within the industry, with some competitors arguing the package did not go far enough; the DuckDuckGo chief executive called the remedies insufficient shortly after the September 2025 decision.

Appellate review of a case this size typically runs many months, and oral argument has not been set. For now, the remedies remain in force while the litigation continues, leaving marketers, publishers, and competing platforms to plan against an outcome that will not be known until the D.C. Circuit rules.

Timeline

Summary

Who: Google LLC, represented by counsel from Munger, Tolles & Olson, Williams & Connolly, and Ropes & Gray, filed the brief. The opposing parties are the United States, 11 states, and a separate group of 38 states. The case was decided at trial by U.S. District Judge Amit P. Mehta and is now before the United States Court of Appeals for the District of Columbia Circuit.

What: Google filed a 100-page opening appellate brief, certified at 20,787 words, asking the court to reverse the district court's finding that Google illegally maintained a search monopoly under Section 2 of the Sherman Act, and to vacate the remedies. The brief argues the trial court used the wrong legal standard for exclusionary conduct, wrongly treated the browser agreements as exclusive dealing, defined the relevant market too narrowly by excluding specialized vertical providers, failed to measure how much of Google's market position came from the challenged agreements, and had no basis for extending data-transfer and syndication remedies to generative AI companies.

When: The brief was filed on May 22, 2026. It follows a January 16, 2026 notice of appeal, a December 2025 final judgment, a September 2, 2025 remedies decision, and an August 2024 liability ruling. Oral argument has not been scheduled.

Where: The appeal was filed with the United States Court of Appeals for the District of Columbia Circuit, as Case No. 26-5023, consolidated with two related cases. The underlying cases were tried in the U.S. District Court for the District of Columbia.

Why: The case determines whether Google's distribution agreements that made its search engine the default in major browsers and on Android devices violated antitrust law, and whether the resulting remedies - including forced sharing of search index data, user interaction data, and search results - will stand. The outcome affects roughly $26 billion in annual default-placement payments, the competitive position of rival search engines and AI answer products, the way market dominance is measured in search, and the degree to which advertising auctions and search distribution change for marketers and publishers.

Share this article
The link has been copied!