EU's Google Endgame Turns Search Data Tap to

EU's Google Endgame Turns the Search Data Tap Toward AI Rivals

The European Commission is days from binding Google to share its search query, click and ranking data with rival engines — including AI chatbots — under the Digital Markets Act, a move that reframes antitrust from policing a results page to redistributing the data pipeline that feeds the next discovery layer.

The European Commission is set to conclude as early as next week that Google violated the bloc's Digital Markets Act by steering users toward its own shopping, travel and hotel services, with fines reportedly running into hundreds of millions of euros across two cases and a 60-day window to comply or face daily penalties, according to the Financial Times, as reported by Noah Intelligence. But the fines are the headline; the substance is a binding data-sharing order — due by July 27 — that would force Google to hand rival search engines, including AI chatbots with search functions, anonymized access to the query, click and ranking data it uses to optimize its own results. The real target is not the blue links. It is the data pipeline feeding classical search and AI retrieval alike.

The Weapon: Article 6(11), Not Just a Fine

The fines matter politically but are structurally secondary. Under the DMA, the Commission can levy up to 20 percent of a gatekeeper's global annual turnover for repeated non-compliance, and daily penalties of 5 percent of average daily turnover for drag-out defiance, as CSIS notes in its assessment of the Act's first years. Alphabet has so far evaded DMA fines even as Apple was hit €500 million and Meta €200 million for anti-steering and consent breaches.

The binding instrument is the specification order. On January 27, 2026, the Commission opened proceedings (case DMA.100209) to define exactly how Google must comply with Article 6(11) of the DMA, which requires gatekeepers operating search engines to "provide to any third-party undertaking providing online search engines, at its request, with access on fair, reasonable and non-discriminatory terms to ranking, query, click and view data." On April 16, 2026, the Commission sent Alphabet its preliminary findings — draft binding measures that, after a public consultation that closed May 1, become law by July 27.

The measures apply a "parity principle": Google must share all the query, view, click and ranking data it collects to optimize its own search, subject only to anonymization. The data must flow through an API "at a frequency on par with Alphabet's own frequency of access," for a duration of the beneficiary's choice up to five years, with latency of no more than seven days and no more than Google's own internal lag, per the Commission's proposed measures. The measures do not require Google to share its algorithms — only the raw behavioral data that makes its algorithms better.

The Twist: AI Chatbots Are Eligible Beneficiaries

The most consequential line in the preliminary measures is the eligibility clause. The Commission specifies that Alphabet "shall not exclude third-party undertakings, such as undertakings providing AI chatbots with OSE functionalities," so long as they meet the definition of an online search engine under EU regulation. The Commission's developer portal is explicit that "AI chatbots rely on search retrieval systems to fetch most recent information from the web and ensure accuracy (so called 'grounding')."

This is where the DMA quietly leaps from a competition remedy for 2010s-style vertical search — the Kayak-versus-Google-Flights fight — into the AI layer. A chatbot that grounds answers in live web retrieval is, for DMA purposes, a search engine. The same data Google uses to rank blue links is the data that trains and retrains retrieval-augmented generation. The Commission is not ordering Google to open Gemini's model weights; it is ordering Google to open the data tap that any rival — OpenAI's search product, Perplexity, a European challenger — could plug into to improve web grounding.

Competition Commissioner Teresa Ribera framed the stakes in the Commission's April announcement: "Data is a key input for online search and for developing new services, including AI. Access to this data should not be restricted in ways that could harm competition." The DMA staff working document released in April 2026 places the data-sharing obligation alongside self-preferencing (Article 6(5)) as the two pillars of "fair online search."

The Counter-Argument: Privacy and Power

Google has two lines of defense, and both are weak in EU law. The first is privacy: forcing Google to share search data raises re-identification risks that the Brookings Institution argues are non-trivial, because de-identified query logs are notoriously hard to anonymize without degrading usefulness. The Commission's answer is a two-layer anonymization regime — technical measures complemented by contractual bans on re-identification attempts, with eligibility screens limiting data to "trustworthy" search providers. Google itself is not held to the same anonymization standard when using its own data, a asymmetry Brookings endorses as the point.

The second is jurisdictional overreach. Google has argued that forcing data sharing exceeds the Commission's powers. This argument lost in the ECJ on July 2, 2026, when the court dismissed Google's final appeal against the €4.1 billion Android fine — confirming that the Commission's view of what constitutes an abuse of dominance in search ecosystems is legally sound, as BBC News and The Economist reported. The Commission's accumulated confirmed fines against Google now stand at roughly €9.45 billion across the Shopping, Android and ad-tech cases.

The Transatlantic Mirror — and the Trump Wedge

The EU move runs directly against the grain of Washington's posture. In a February 2025 memorandum titled "Defending American Companies and Innovators from Overseas Extortion," President Donald Trump directed agencies to prepare tariffs against foreign regulations deemed to burden US tech firms, as Brookings documented. On June 27, 2026, Trump threatened a 100 percent tariff on any country implementing a digital services tax on US companies, with the levy explicitly designed to "supersede" existing trade deals, per BBC News and Al Jazeera. In January 2026, the State Department imposed visa restrictions on five European officials involved in drafting the DMA and DSA, CSIS reported — the most direct US action against EU regulators to date.

The Commission has not blinked. The new head of the competition directorate, Anthony Whelan, told the Financial Times that enforcement "needs to be even-handed, fact-based, constrained by whatever is the legal framework" and will not be influenced by "how much noise others make." The EU-US trade deal finalized in July 2025 capped tariffs on most EU goods at 15 percent but conspicuously excluded digital regulation from its scope. Brussels has treated that silence as a green light.

The divergence is sharpest on remedies. In the US, Judge Amit Mehta's September 2025 ruling ordered Google to share search index and click-and-query data with qualified competitors for six years, overseen by a five-person technical committee, as NPR detailed — but Mehta stopped short of a Chrome divestiture and declined to extend remedies to advertising data. The EU's parity-principle data order is broader on the data side and narrower on the structural side: no breakup, but a deeper, API-level, five-year data tap open to any eligible search provider including AI chatbots. The two regimes are converging on data sharing as the remedy of choice, but Brussels is setting the terms.

Who Benefits

The immediate beneficiaries are the vertical-search plaintiffs who have litigated Google's self-preferencing for over a decade. Trivago filed an antitrust damages claim against Google in a Hamburg court on May 5, 2026, seeking compensation for traffic lost to Google's hotel metasearch since 2014, building on the ECJ's September 2024 confirmation of the Shopping fine, as the FT reported. Two first-instance damages awards have already been issued by a Berlin court in November 2025. The DMA fines and binding data order would convert these civil actions from uphill battles into near-open-and-shut liability findings.

The second-order beneficiaries are AI-search challengers. A European Commission explicitly inviting "AI chatbots with search functionality" to apply for Google's data is handing OpenAI, Perplexity, Mistral and any future EU-founded retrieval model a regulated right to the same behavioral signal Google uses to rank results. The Commission's own framing — that grounding accuracy depends on this data — positions the measure not as a search-market fix but as an AI-market intervention.

The UK is moving in parallel but more cautiously. On June 17, 2026, the Competition and Markets Authority ordered Google to rank organic results using "objective and non-discriminatory criteria," allow data portability to authorized third parties, and give publishers an opt-out from AI Overviews, according to Al Jazeera and the BBC. The CMA gave Google six months for fair-ranking compliance and three for data portability. London is following the Brussels template but with bilateral flexibility toward Washington — a hedge Brussels is not making.

What to Watch

• July 27, 2026 — Statutory deadline for the Commission to adopt the final, legally binding data-sharing measures under Article 6(11). This is the hard date. If the measures land as drafted, Google has an implementation window to build the API and eligibility process.
• Week of July 20 — The Financial Times-flagged announcement window for the two DMA non-compliance fines (self-preferencing and, separately, Google Play anti-steering). The fine figures and the 60-day remedy clock will set the compliance calendar.
• DMA.100220 — A separate, less-reported specification proceeding on Android interoperability with third-party AI services, still at the preliminary stage. If the Commission applies the same parity logic here, Google could be compelled to open Android-level access points for rival AI assistants — extending the data-sharing principle from the search box to the operating system.
• ECJ appeals track — Google will almost certainly challenge the binding measures in the General Court, but the July 2 Android ruling narrows the legal runway. The Commission is winning the precedent war.
• US response — Whether the Trump administration links the DMA data order to its tariff threat, or treats it as separable from the digital-tax fight, will determine whether this becomes a trade incident or a regulatory routine.

Diplomat View

The EU is not fining Google for the 2010s sin of self-preferencing. It is building a regulated right of access to the data layer that underpins both classical search and AI retrieval, and it is doing so under a statute — the DMA — that the ECJ has now twice declined to constrain. The decisive move is the eligibility of AI chatbots as data beneficiaries: that single clause turns a search-competition remedy into an industrial policy for the AI discovery layer, and it does so at exactly the moment the US government is attempting to shield American platforms from foreign regulation via tariff threat. The forecast is that the July 27 binding order lands substantially as drafted, Google appeals and loses in the General Court within 18 months, and at least one AI-search challenger — most likely a European model — secures a data-access agreement by mid-2027. What would change this: a Trump administration decision to fold DMA enforcement into the tariff calculus, forcing an EU member-state revolt against the Commission's enforcement appetite. Absent that, the data tap opens, and the search market's next chapter is written in Brussels.