From Madrid to Washington, the same charge keeps landing: that Google used a gateway it already controlled — search ranking, Android defaults, the Play Store, the ad-tech stack — to box out rivals. The newest judgments no longer just fine the company. They order it to rebuild.
Google's most consequential losses cluster in competition law, with major adverse rulings in the EU, the United States, India and Australia. Across them, courts repeatedly accepted variants of one theory: that Google leveraged control of a critical access point to preference its own services, foreclose rivals, or preserve a monopoly. The recurring labels are self-preferencing, exclusive defaults, tying, leveraging dominance into adjacent markets, and restrictions on interoperability.
Outside antitrust, the losses came from privacy and consumer-protection design choices rather than the mere fact of collecting data — cases where product architecture or disclosures made processing opaque, misleading, or insufficiently consented to.
The remedy trend matters as much as liability. Older losses ended in fines or delisting duties. Newer ones impose structural conduct constraints: limits on default-search and preload deals, bans on conditioning Play access, catalog-sharing and rival-store obligations, data-sharing requirements, and technical oversight.
Courts are moving from “pay and stop” to “re-architect the platform.”
In Shopping it was the results page; in Android and US Search, the path that preloads apps and assigns defaults; in Epic and India, the app-store and billing bottleneck; in ad tech, combined control of ad-serving and the exchange. Courts ruled against Google when they could tell a simple story: it used a bottleneck it already held to make rivalry meaningfully harder.
Google Spain and the CNIL case changed obligations around delisting and consent. The newer competition cases go further: the DC search judgment rewrites Google's contracting and mandates competitor access; the Epic injunction forces Play Store openness; India targets billing and distribution. The future risk is business-model redesign, not just money.
The vocabulary differs — Article 102 in the EU, Section 2 in the US, Section 4 in India, consumer law in Australia — but the concerns line up: self-preferencing, tying, discriminatory access, default-placement payments, foreclosure by design. EU-style gatekeeper theories now travel fast, and US courts impose remedies once thought distinctly European.
Borrowing from information geometry — the study of families of related objects as points on a curved space — each ruling becomes a point, and a jurisdiction's legal vocabulary (Article 102, Sherman §2, India's §4, Australian consumer law) is just one coordinate chart describing it. The interesting question isn't where a case sits in any one chart, but the shape of the space all the cases share.
To draw that space, a divergence is defined between any two cases: how much they disagree on legal theory, remedy, era and doctrinal area — with jurisdiction deliberately down-weighted, so geography can't dominate the geometry. Projecting those divergences into two dimensions (classical MDS) places doctrinally similar rulings near each other. Proximity is shared theory; the skeleton lines are each case's nearest neighbours — the local geodesics along which the law actually moved.
Recolour the same points by remedy or era to read different scalar fields over the manifold. This is an analytical lens, not a claim that law is literally Riemannian — but the structure it exposes is real, and it's drawn straight from the thirteen rulings above.
The defaults-and-preload cases from the EU, India, the US and Australia are mutual nearest neighbours. Down-weight jurisdiction and the borders dissolve — what remains is a single bundling ridge that four legal systems independently walked onto.
Across the privacy–IP boundary, Google Spain and Equustek are each other's closest points (the gold span). Different doctrines, one shared act: ordering the search intermediary to remove links. The geometry finds the kinship the labels hide.
Switch the colour to remedy and a gradient appears across the competition basin: fines on one flank, structural conduct orders on the other. The manifold tilts — and the recent rulings all sit downhill, on the “re-architect” side.
The space isn't uniform. Rulings fall toward two attractors — a competition basin and a privacy/intermediary basin — with the consent cases knotted tightly together. Google's legal risk lives in distinct regions, not one cloud.
Its largest losses now form a coherent body of law across competition, privacy and consumer protection, and intermediary obligations — but the center of gravity is unmistakably antitrust. Courts increasingly treat Google's ecosystem-design choices as exclusionary when they make rivals depend on Google-controlled gateways. That thread runs from Shopping and Android to DOJ Search, Epic, and the Indian cases.
Two risk trends follow. Distribution deals and default placements are now a primary litigation hazard, especially when reinforced by payments or cross-product conditioning. And remedies are hardening — aimed at opening interfaces, unbundling defaults, and enabling rival access. That is a greater long-term threat than even very large fines.
Scope note: this record applies a significance threshold — final appellate or merits judgments, major adverse rulings of operational weight, or court-enforced settlements with admitted liability — rather than cataloguing every local dispute. Several matters are flagged as major adverse rulings rather than final global endpoints where remedy or appeal stages remained active in the reviewed sources: EU Android, US ad tech, India Play Billing, and the implementation phase of US Search.