Key figures: Google global search share: ~90% · Desktop share: ~84% · Mobile share: ~95% · Apple default deal: ~$20bn/year · Remedies: exclusive deals banned, data sharing ordered · Chrome divestiture: rejected, DOJ appealing

In August 2024, US District Judge Amit Mehta ruled that Google had illegally maintained a monopoly in general search and search text advertising, violating Section 2 of the Sherman Antitrust Act. It was the most significant antitrust ruling against a technology company since the Microsoft case two decades earlier.1 In September 2025, Mehta issued his remedies: Google would be barred from entering exclusive default search agreements, required to share its search index and user interaction data with qualified competitors, and ordered to offer search and text ad syndication services to rivals. He rejected the DOJ's request to force a sale of Chrome or Android.2

Google filed its appeal on January 16, 2026. The DOJ cross-appealed in early February, pushing for Chrome divestiture to be reinstated. On May 22, Google submitted a 111-page brief to the D.C. Circuit arguing that the trial court had confused harm to competitors with harm to the competitive process. Oral arguments are expected in late 2026 or early 2027.3

The legal drama is significant. But the case also provides a test of a question at the heart of competition economics: can regulation break a monopoly sustained not by predatory behaviour in the traditional sense, but by network effects so powerful that no rival can replicate them?

The Self-Reinforcing Loop

Google's dominance rests on what economists call a demand-side economy of scale. More users generate more queries. More queries produce more data on what results are relevant. More data improves the algorithm. A better algorithm attracts more users. The cycle repeats.

Network effect: A phenomenon where the value of a product or service increases as more people use it. In search, the effect operates through data: each query improves the algorithm for every subsequent user, creating a self-reinforcing advantage for the incumbent.

Google processes more than 5 trillion searches a year, around 14 billion a day.4 Its search index is more than twice the size of Bing's.5 Judge Mehta's ruling specifically identified this data asymmetry as central to the monopoly: competitors cannot match Google's quality without access to comparable volumes of user-interaction data, and they cannot attract comparable data without first matching Google's quality. The classic chicken-and-egg problem of concentrated markets.

But the aggregate 90 percent global market share masks a more revealing divergence. On mobile, where Android defaults and the Apple deal lock in Google as the pre-set option, its share sits at roughly 95 percent. On desktop, where users have more freedom to change settings and where Bing's Copilot integration has gained traction, Google's share is lower, at around 84 percent, down from around 88 percent in 2019.6

Google's search engine market share by device (%), 2018 to 2026. Source: StatCounter, annual averages.

That gap tells the story. Where defaults are hardwired and switching costs are high (mobile), Google's position is almost untouchable. Where users can more easily choose alternatives (desktop), the moat is narrower and eroding. The remedies are, in essence, an attempt to make the entire market look more like desktop.

The Default That Built the Moat

The mechanism through which Google maintained its monopoly, according to the court, was a series of distribution agreements that ensured no competitor could access the user base needed to challenge it. The most important was the deal with Apple.

Google pays Apple approximately $20 billion per year to be the default search engine on Safari across iPhones, iPads, and Macs.7 In 2023, Alphabet CEO Sundar Pichai testified under oath that Google pays Apple 36 percent of Safari's search revenue.8 The deal is not technically exclusive: users can switch to Bing or DuckDuckGo in their settings. But defaults are extraordinarily sticky. Research on consumer choice architecture consistently shows that most users never change a pre-set option. The $20 billion Google pays is, in effect, a measure of how much that default is worth: the position guarantees a volume of searches no competitor can otherwise reach.

The age old saying 'the devil is in the details' may not have been devised with the drafting of an antitrust remedies judgment in mind, but it sure does fit.

Judge Amit Mehta, December 2025 remedies ruling.9

Similar deals existed with Samsung and Mozilla. Mozilla's finances illustrate the dependency: Google's payments have accounted for the majority of Mozilla's annual revenue, making the non-profit browser maker financially dependent on the very monopoly the court declared illegal.7

Mehta's remedy bans these exclusive agreements. Google can still bid for default status, but any deal must terminate within one year and cannot contain exclusivity provisions.

Why the Remedies May Not Work

The sceptical case is straightforward. Banning exclusive deals addresses the distribution channel through which Google maintained its monopoly. It does not address the underlying network effect.

Consider what happens when the Apple deal expires. Apple will be free to negotiate with any search provider. But which provider can afford to pay anything close to $20 billion per year? Only Google generates enough search advertising revenue to justify that price. Bing, which holds roughly 4 percent of global search, does not have the revenue base to compete. The ban on exclusive deals may simply result in Google winning the same contracts on non-exclusive terms because no rival can match the bid.6

The data sharing remedy is more ambitious. By requiring Google to share its search index and user interaction data with qualified competitors, Mehta is attempting to break the data moat directly. If rivals can access Google quality data, the theory goes, they can build Google quality search engines.

But Google argues, not unreasonably, that raw data is not the same as a functioning search product. Its appeal brief calls the data sharing order an attempt to "manufacture artificial competition" rather than reward firms that built their own capabilities.3 There is a legitimate tension: if the reward for building the best search engine is being forced to share it, what incentive does anyone have to build the next one?

The Microsoft Precedent

The closest historical parallel is the Microsoft antitrust case of 1998 to 2001. The DOJ argued that Microsoft had used its Windows monopoly to crush Netscape Navigator by bundling Internet Explorer. A trial court ordered Microsoft broken in two. The D.C. Circuit, the same court now hearing Google's appeal, reversed the breakup and imposed behavioural remedies instead.10

The conventional wisdom is that those remedies were weak and that market forces, not regulation, actually broke the monopoly: Apple's Safari and then Google's own Chrome succeeded by being better products on new platforms. The implication for the Google case is uncomfortable for regulators: if behavioural remedies failed to dislodge Microsoft and competition did the job, what reason is there to think they will work against Google?

The counterargument is that the search market is structurally different from browsers. Search engines require data at a scale that browsers never did. The barriers to entry are orders of magnitude higher. The DOJ's cross appeal for Chrome divestiture rests partly on this logic: if behavioural remedies alone cannot break the data loop, structural separation may be the only option.

The AI Complication

AI powered search is emerging as a competitive threat in a way no traditional search engine has managed. ChatGPT, Perplexity, and other AI tools collectively accounted for roughly 0.9 percent of all referral traffic in March 2026, up fivefold from 0.18 percent a year earlier, according to Similarweb data compiled by Digital Applied.11 That is still negligible. But AI search offers a fundamentally different experience rather than a marginal improvement on the same model.

Judge Mehta acknowledged this, noting that generative AI "plays a significant role in these remedies" and extending restrictions to cover Google's Gemini app alongside Search, Chrome, and Google Assistant.9 The DOJ's cross appeal argues that without structural remedies, Google will replicate its search monopoly playbook in AI: using its data advantage, distribution agreements, and bundling to ensure Gemini becomes the default AI assistant, just as Google Search became the default search engine.

The Google case is a live experiment in whether antitrust law can dismantle a monopoly sustained by network effects. The remedies attempt something rarely tried: not just punishing anticompetitive behaviour, but restructuring the information asymmetry that makes a monopoly self sustaining. Whether data sharing orders and distribution bans can achieve what decades of direct competition could not will depend on whether Google's dominance is truly a product of illegal conduct, or whether it reflects the natural economics of a market where scale begets quality and quality begets scale.

The D.C. Circuit's decision, expected sometime in 2027, will not just determine Google's future. It will set the template for how regulators approach network-effect monopolies for a generation.

Footnotes

  1. White & Case, "In Landmark Decision, D.C. Federal Court Holds Google Maintained an Illegal Monopoly in Internet Search and Advertising Markets" (opens in a new tab), August 2024. ↩

  2. CNBC, "Google stock jumps as judge rules it can keep Chrome in antitrust case" (opens in a new tab), September 2025. ↩

  3. Courthouse News Service, "Google Urges D.C. Circuit to Overturn Search Monopoly Remedies" (opens in a new tab), May 2026. ↩ ↩2

  4. Search Engine Land, "Google now sees more than 5 trillion searches per year" (opens in a new tab), March 2025. ↩

  5. UK Competition and Markets Authority, "Online Platforms and Digital Advertising Market Study, Appendix I: Search Quality and Economies of Scale" (opens in a new tab), 2020 (para 75). ↩

  6. StatCounter, "Search Engine Market Share Worldwide" (opens in a new tab), accessed June 2026. ↩ ↩2

  7. AdExchanger, "Judge Mehta's Remedies For Google's Search Monopoly Won't Cure What Ails Publishers" (opens in a new tab), September 2025. ↩ ↩2

  8. CNBC, "Alphabet pays Apple 36% of Safari search revenue, Sundar Pichai confirms" (opens in a new tab), November 2023. ↩

  9. CNBC, "Judge finalizes remedies in Google antitrust case" (opens in a new tab), December 2025. ↩ ↩2

  10. The Regulatory Review, "Antitrust Remedies After Google" (opens in a new tab), November 2025. ↩

  11. Digital Applied, "Search Engine Market Share 2026" (opens in a new tab), 2026. ↩