# United States v. Google Two separate federal monopolization cases against Google, tried before different judges in different districts, both won by the government on liability, and both of which then **declined to break anything up**. Together with [[FTC v. Amazon]] they are the test of whether the 2020s antitrust revival can produce structural relief, and so far the answer is no. ## Case 1 — Search. D.D.C., Judge Amit Mehta **Filed 20 October 2020** with eleven state attorneys general; a broader companion suit led by Colorado, joined by 38 states and territories, was filed in December 2020 and consolidated for trial. Trial ran from September to November 2023. **Liability, 5 August 2024.** A 277-page opinion found Sherman Act §2 monopoly maintenance in **two** markets — **general search services** (89.2% share overall, 94.9% on mobile) and **general search text advertising** (about 88%). The line everyone quotes: *"Google is a monopolist, and it has acted as one to maintain its monopoly."* What the government **lost** is less reported and matters: the broader "search advertising" market was accepted as plausible but Google was found to lack monopoly power in it; a proposed "general search advertising" market was rejected outright; and the state plaintiffs' claim about Search Ads 360 failed on the ground that there is no antitrust duty to deal. **The conduct** was exclusive and near-exclusive default distribution — the Apple agreement making Google the Safari default, Android distribution and revenue-sharing agreements with manufacturers and carriers, and browser defaults at Mozilla and Opera. Mehta rejected the "defaults are one click away" defense, finding the deals foreclosed the most efficient distribution channels and created a scale feedback loop. > [!note] Get the payment years right > Google paid **$26.3 billion** for defaults across all distributors in **2021**, and roughly **$20 billion** to Apple alone in **2022**. The widely repeated "$10 billion a year to Apple" is a stale pre-trial estimate and should not be used for the trial-record years. ### The remedy, 2 September 2025 Mehta **ordered**: no exclusive agreements for distributing Google Search, Chrome, Assistant or **Gemini**; no conditioning Play Store or app licensing on placement; **maximum one-year terms** for default agreements, renegotiable annually; **data sharing** of portions of the search index and user-interaction data to court-certified "qualified competitors" at marginal cost, with ads data excluded; search and search-text-ad **syndication licenses** on commercial terms for about five years with declining usage caps; advance notice of auction and algorithm changes; a five-member technical committee; and a six-year decree. He **declined**: **Chrome divestiture** — "the complete divestiture of Chrome is a poor fit for this case"; contingent Android divestiture; and, critically, **a ban on paying for defaults.** Google may keep paying Apple, Mozilla and manufacturers, provided the deals are non-exclusive and short. A ban, he reasoned, would "pose a substantial risk of financial harm to browser developers, wireless carriers, and cellphone manufacturers." ⭐ **Two reasons carried the decision, and both will be reused by every defendant facing a remedies phase this decade.** **The causation standard.** Under *Microsoft*'s "significant causal connection" test, Mehta found "ample evidence that lawful conduct... played an important role in the maintenance of Google's monopoly." Where engineering investment and genuine product quality contributed alongside the illegal conduct, divestiture does not follow from a liability finding. **Generative AI as an off-ramp.** He found the remedies record materially different from the liability record: AI assistants are "a new and potentially serious competitive threat to Google's dominance," in a space that is "highly competitive" with well-capitalized entrants. So he declined structural relief on the theory that the market may disrupt itself — while extending the conduct remedies **to GenAI products**, specifically to stop Google reproducing default exclusivity in AI. The remedy is calibrated to a disruption that has not yet happened. **Appeal.** Final judgment entered 5 December 2025; Google appealed in January 2026 to the D.C. Circuit (Nos. 26-5023, 26-5047, 26-5049, consolidated), attacking market definition, monopoly power and the remedies. DOJ and the states responded and **cross-appealed on 28 July 2026, challenging only the refusal to ban payments for defaults** — arguing Mehta never asked whether relief without a ban can actually restore competition. Notably **they did not cross-appeal for Chrome divestiture.** No oral argument had been scheduled as of late September 2026. ## Case 2 — Ad tech. E.D. Va., Judge [[Leonie Brinkema]] **Filed 24 January 2023** in the "rocket docket" with eight states. Trial September 2024. **Liability, 17 April 2025.** Unlawful monopolization of the **publisher ad server** market (DoubleClick for Publishers) and the **ad exchange** market (AdX), plus unlawful **tying** of the two. The alleged advertiser-ad-network market was rejected as a distinct market, and the historic DoubleClick and AdMeld acquisitions were not held independently anticompetitive. **Remedy, 2 September 2026** (opinion unsealed 16 September 2026, 106 pages). **AdX divestiture denied** — no viable buyer, migration estimated at five to eight years, disruption to small publishers, and appellate uncertainty. Brinkema's line is the one to keep: *"Distrust of Google alone is not a lawful basis for structural relief."* She ordered instead a six-year worldwide behavioral package: API integrations exposing AdX and DFP to Prebid; AdX bidding into rival publisher ad servers on the same terms as DFP; real-time win/loss bid data to all publishers; elimination of First Look, Last Look and Unified Pricing Rules for indirect demand; per-exchange price floors; no preferencing of Google ad tech in Google Ads; and a monitor with source-code access. **No final judgment had been entered as of 27 September 2026** — proposed judgments were due 2 October. Google has said it will appeal liability to the Fourth Circuit. > [!note] Not the same as Epic Games v. Google > Epic's case is private, was won before a jury in December 2023, produced a Play Store injunction affirmed by the Ninth Circuit in July 2025, and was substantially resolved by settlement in 2026. It is frequently cited as though it were one of the government cases. It is not. ## The through-line Three consecutive remedies decisions against Google produced **zero structural relief**: no Chrome, no Android, no AdX. Both federal judges applied *Microsoft*'s causation test at the remedy stage and found lawful contributors to monopoly sufficient to defeat divestiture. And the [[AT&T Breakup]] analogy that breakup advocates reach for is weaker than it looks. The 1982 divestiture was a **consent decree AT&T agreed to**, not structural relief imposed over a defendant's objection. The one genuinely imposed modern breakup — *Microsoft*'s 2000 structural order — was **vacated on appeal in 2001**, which is precisely the appellate risk Brinkema cited in declining to order one. The 2020s record is therefore coherent and somewhat bleak for structuralists: **monopolization findings are winnable; breakups are not.** [[FTC v. Amazon]] is next, with trial set for 2027 and liability and remedies bifurcated. On this pattern, an FTC liability win would most likely yield conduct rules rather than a Prime/FBA separation. ## See Also - [[FTC v. Amazon]] — the next test of the same question - [[AT&T Breakup]] — the structural precedent, and why it is not the precedent it is taken for - [[Consumer Welfare Standard]] · [[New Brandeis Movement]] — the doctrinal stakes - [[Antitrust Division]] · [[Leonie Brinkema]] - [[US Antitrust and Monopoly Timeline]] · [[Legal Cases Hub]]