Google Search Data: A New DMA Enforcement Test

The April 16 measures, in detail

The Commission’s April 16, 2026 announcement is the first time DMA enforcement has produced a field-level data-access specification. Google, as a designated gatekeeper, must allow third-party search engines and AI chatbots with search functions to access four specific categories of data: ranking signals, search queries, click logs, and view data. Access has to be on fair, reasonable, and non-discriminatory (FRAND) terms — the same legal frame typically applied to standard-essential patents in telecom — applied here to a gatekeeper’s most commercially sensitive logs.

The technical depth matters. Brussels published a 29-page specification document that defines, at the field level, what data must flow, how it must be anonymised, how it can be priced, and what audit regimes will govern access. The document covers six concrete dimensions: who qualifies as a “data beneficiary,” the scope of shareable data, sharing methods and frequency, anonymisation rules for personal data, pricing parameters, and the governance process for access disputes. That is the sort of specificity that turns regulation from a principle into a workable obligation.

The procedural calendar is just as important. The Commission opened formal Article 6(11) proceedings on January 27, 2026. The April 16 announcement is a preliminary finding, not a binding remedy. A public consultation opened on April 17, 2026 with a deadline of May 1, 2026, and the Commission must adopt a final decision within six months of the proceeding’s opening — so by July 27, 2026. Google has the right to respond, and the final remedy may differ from the April text, but the trajectory is clear.

Why search data is the fight that matters

The DMA designated Google Search as a “core platform service” in September 2023 and identified Article 6(11) data-sharing obligations from the start, but until April 2026, those obligations existed mostly on paper. The shift is not just enforcement intensity. It is the recognition that search-derived data — what users typed, what they clicked, what ranked above what — is now the foundational input for an entire generation of AI products.

That is the most interesting move in the April text: AI chatbots with search functionality are explicitly listed among eligible data beneficiaries alongside traditional search engines. Perplexity, OpenAI’s ChatGPT search, You.com, Mistral’s Le Chat, and other AI-native search interfaces all depend on real-time access to web-ranking signals to compete with Google’s own Gemini. If those companies can obtain Google’s ranking and click data on FRAND terms, the cost of building a credible Google-quality search experience drops materially. If they cannot, the AI-search competitive race is decided before it really starts.

The Commission’s framing reads as a deliberate response to that reality. Margrethe Vestager’s successor team — Teresa Ribera now holds the competition portfolio, with Henna Virkkunen as Executive Vice-President for Tech Sovereignty — has been explicit that DMA enforcement should keep pace with AI-era market structures rather than relitigate the desktop-search disputes of the 2010s.

Implementation is where regulation lives or dies

Large regulatory frameworks routinely fail at the implementation layer. The DMA’s core text is only a few hundred pages, but its credibility depends on documents like the 29-page Google search specification — the sort of obscure technical annex that determines whether a rule survives litigation. Anonymisation thresholds, query-frequency caps, audit rights, dispute-resolution timelines, and pricing benchmarks all sit in that document. Each one is a place where compliance can soften under pressure.

Google has already signalled disagreement with the proposed scope, arguing in its public response that some of the requested data fields raise privacy and competitive risks. The May 1 consultation will produce written submissions from challengers including Microsoft Bing, DuckDuckGo, Qwant, Ecosia, and AI search providers, and from privacy regulators concerned about whether anonymisation is sufficient. The text the Commission adopts in late July will reflect that pressure, in either direction.

That implementation focus is also why the case matters beyond search. Article 6(11) is one of nine data-related obligations in the DMA, and the way the Commission resolves field-level questions for Google will shape how analogous obligations are enforced against Apple’s App Store data, Meta’s advertising-feedback signals, and Amazon’s marketplace data flows. A credible Google specification creates a template; a watered-down one creates a precedent regulators will struggle to escape.

Why this case is bigger than Google

If the Commission lands the July 27, 2026 decision with substantive teeth, it sets a precedent for how platform-derived data flows under regulation. AI-era competition increasingly depends on access to behavioural data accumulated inside dominant platforms. Whether that data is treated as a private asset of the platform or as essential market infrastructure subject to access rules is one of the defining policy questions of the next five years.

The case also has implications outside the EU. The UK’s Competition and Markets Authority is running a parallel “strategic market status” review of Google Search under the Digital Markets, Competition and Consumers Act, and South Korean regulators are examining similar data-access questions. Ex-EU jurisdictions, including Algeria, generally watch the Brussels precedent closely because EU enforcement creates the global compliance baseline that multinational platforms tend to roll out worldwide rather than fragmenting by region.

For Algerian regulators, the takeaway is not whether to import the DMA, but to study how a serious data-sharing remedy is constructed: who counts as a beneficiary, how access is priced, what audit rights exist, and how disputes are resolved. Those are the same questions Algerian competition policy will eventually face when domestic digital markets concentrate, and the answers being worked out in Brussels in April-July 2026 are the most detailed reference any regulator has.

Three Signals Hidden in the DMA Enforcement Structure

The April 16 measures are not merely a Google compliance story. They reveal how AI-era competition policy is being redesigned in real time. Each of the three structural signals below has implications beyond EU borders — for regulators, AI builders, and platform strategists watching how the next wave of competition rules is assembled.

Signal 1: Behavioral Data Is Being Reclassified as Market Infrastructure

For two decades, a platform’s query logs, click streams, and ranking signals were treated as proprietary assets — the data exhaust of running a service, privately owned and unregulable at the field level. The April 16 specification reverses that treatment. By requiring Google to share ranking, query, click, and view data under FRAND terms with any qualifying beneficiary — including AI chatbots — the Commission is reclassifying behavioral data as essential market infrastructure. This is the same conceptual move applied to telecom interconnection in the 1990s: a company’s proprietary signals become regulated inputs once their absence structurally prevents competitors from entering the market. The key question the July 27 decision will answer is whether the reclassification survives Google’s litigation challenge — and if it does, it sets a template that the UK CMA, the Korean FTC, and eventually non-European regulators will reference for years.

Signal 2: AI Chatbots Are Recognized as a Distinct Competitive Layer — Not a Google Feature

The April 16 text explicitly lists AI chatbots with search functionality among eligible data beneficiaries alongside traditional search engines. This is a precise and consequential framing choice. It acknowledges that Perplexity, OpenAI’s ChatGPT Search, Mistral’s Le Chat, and You.com are not Google’s downstream customers — they are a parallel search market whose competitive viability depends on access to the same ranking signals that power Google Gemini. By granting them beneficiary status, the Commission is treating AI-mediated search as a structurally distinct market that requires the same data-access remedies as classic blue-link search. The practical implication: any AI search product that previously had to reverse-engineer Google’s ranking signals by crawling at scale can now negotiate FRAND terms for direct access — dramatically lowering the cost of building a credible Google-quality search alternative.

Signal 3: Enforcement Depth Now Lives in 29-Page Technical Annexes, Not Headlines

The most important document in this proceeding is not the Commission’s press release — it is the 29-page specification defining field-level data scope, anonymisation thresholds, pricing parameters, and audit rights. This is where the regulation either works or is quietly arbitraged into meaninglessness. Anonymisation thresholds determine whether the data beneficiaries receive enough signal to train competitive models or only a degraded representation. Pricing parameters determine whether access is genuinely FRAND or nominally compliant but commercially prohibitive. Audit rights determine whether the Commission can verify compliance without a three-year litigation cycle. Regulators outside the EU — including Algerian competition-policy researchers — should read the July 27 final annex as a masterclass in how to convert a high-level data-access obligation into a technically enforceable remedy, rather than the press release that describes its political intent.

The Antitrust Question

The deeper antitrust question underneath the Google search data case is one that will follow every AI-era regulator: when does a platform’s accumulated data advantage become a structural barrier to competition, and at what point does that barrier require a remedial access obligation rather than just market observation? The DMA answers this for European markets with a designated-gatekeeper model — thresholds based on revenue, users, and platform type trigger the obligation before harm has to be proven. Other jurisdictions are watching the July 27, 2026 outcome to see whether the DMA’s approach is enforceable or whether Google’s litigation will blunt the remedy enough to make it an impractical template.

The UK Competition and Markets Authority’s parallel strategic market status review under the Digital Markets, Competition and Consumers Act is the most closely watched alternative test. South Korean regulators are examining comparable data-access questions. What all three proceedings share is the recognition that behavioral data — query logs, click streams, ranking signals — is no longer just operational exhaust from running a service. It is a resource that competing AI products cannot replicate from scratch in commercially viable timescales, which is precisely the condition that justified mandated interconnection in telecommunications, payment networks, and power grids in earlier eras.

The antitrust question for the next decade is whether that logic holds at the European Court of Justice level, and what happens to AI-native search competitors like Perplexity, Mistral’s Le Chat, and OpenAI’s ChatGPT Search if it does not. The July 27, 2026 final decision will be the most significant regulatory precedent of the AI era so far. Regulators, competition-law researchers, and platform strategists — in Algeria and everywhere else — should read the full technical annex, not just the press release.

Frequently Asked Questions

Sources & Further Reading

• Commission proposes measures to Google on sharing search engine data — European Commission
• Commission proposes measures — Digital Markets Act portal
• Commission opens proceedings to assist Google in complying with DMA obligations — European Commission
• DMA.100209 — Consultation on the proposed measures for Google Search data sharing
• EU proposes new Google search-data sharing measures under DMA — Digital Watch Observatory