Google Pushes Back Hard on EU Search Data Sharing Order

Google is pushing back forcefully against a European Union demand to share its search engine data with competitors, warning that compliance could compromise the privacy of hundreds of millions of European users. The escalating standoff marks one of the most significant confrontations yet under the EU's landmark Digital Markets Act (DMA).

According to Reuters, Google scientist Sergei Vassilvitskii issued the warning directly to EU antitrust regulators, arguing that opening up search data to rivals like OpenAI would create serious privacy vulnerabilities. Google's response to the EU's month-old demand was blunt and unequivocal: no.

Key Takeaways

• The European Commission ordered Google to share ranking, query, click, and browsing data with competing search engines
• Google responded with a rare, forceful refusal, citing user privacy concerns
• Google scientist Sergei Vassilvitskii plans to present an alternative proposal with broader protections
• The dispute centers on the Digital Markets Act, which designates Google as a 'gatekeeper' in search
• Competitors argue they cannot improve their services without access to the kind of behavioral data Google has accumulated over decades
• The outcome could reshape the competitive landscape of AI-powered search globally

What the EU Is Demanding — and Why

One month ago, the European Commission laid out its position clearly: Google has dominated the search market for so long that newcomers face virtually no path to fair competition. Under the DMA, which took full effect in 2024, companies designated as 'gatekeepers' must take active steps to enable competition in the markets they control.

The EU's specific demand targets the core of Google's competitive advantage. Regulators want Google to share ranking data, search queries, click-through data, and browsing behavior with rival search engines. The logic is straightforward — without access to this kind of data, competitors cannot effectively train and refine their own search algorithms.

This is not a minor request. These data categories represent the lifeblood of modern search technology. They are what allow Google to continuously improve result relevance, understand user intent, and maintain its roughly 90% share of the European search market.

Google's Rare and Forceful Refusal

Google's response has been unusually aggressive. The company is not negotiating around the edges or offering token concessions. It is flatly refusing the request as formulated.

Vassilvitskii, who has served as a Google scientist since 2012 and is a core figure in the company's research on search algorithms and data security, is scheduled to meet with EU antitrust officials. He plans to present what Google describes as a 'broader alternative proposal with more comprehensive protective measures.'

The privacy argument is Google's sharpest weapon in this fight. By framing the data-sharing requirement as a threat to European users — the very people the EU is trying to protect — Google is attempting to turn the Commission's own regulatory philosophy against it. Europe's General Data Protection Regulation (GDPR) is the world's strictest privacy framework, and any data-sharing mandate would need to be reconciled with its requirements.

Google's core argument boils down to several points:

• Search query and click data contains deeply personal information about users
• Sharing this data with third parties — especially companies like OpenAI headquartered outside the EU — creates new attack surfaces for privacy breaches
• There is no technical mechanism to fully anonymize behavioral data at the scale the EU is requesting
• Competitors could reverse-engineer individual user profiles from aggregated behavioral signals

The Hidden Asset: Why 'Signals' Matter More Than Raw Data

To understand why this fight matters so much, it helps to look beyond the surface-level concept of 'search data.' Google's real competitive moat is not just the queries people type — it is the behavioral signals surrounding those queries.

Consider a simple example. A user searches for 'flights from London to Paris.' That is a search query. But the real value lies in what happens next: how long the user stays on a result page, whether they click back, whether they immediately search for 'Paris hotels' or 'Charles de Gaulle airport to city center.' These sequential behavioral patterns — or signals — are what allow Google to understand intent, context, and satisfaction.

Over 2 decades, Google has accumulated an unparalleled library of these signals across billions of users and trillions of searches. This dataset creates a compounding advantage: better signals produce better results, which attract more users, which generate more signals. It is a flywheel that no competitor has been able to replicate.

This is precisely why the EU wants to intervene. Regulators believe this self-reinforcing cycle has made the search market effectively uncontestable. Without some form of data access, even well-funded competitors with strong AI capabilities — like Microsoft's Bing, DuckDuckGo, or OpenAI's emerging search products — cannot close the quality gap.

The AI Search Arms Race Adds Urgency

The timing of this dispute is no coincidence. The search market is undergoing its most dramatic transformation in years, driven by the integration of large language models and AI-powered answer engines.

OpenAI launched ChatGPT Search in late 2024, directly challenging Google's core business. Perplexity AI has attracted millions of users with its AI-native search experience. Microsoft has poured billions into integrating AI into Bing. Even Apple has been exploring partnerships to bring AI search features to Safari.

For these challengers, access to Google-scale behavioral data would be transformative. It would allow them to:

• Train more effective ranking models without needing years of organic user growth
• Understand query intent patterns across languages and regions
• Benchmark their own result quality against Google's established standards
• Accelerate the development of AI search features that require massive behavioral datasets

Google clearly recognizes that sharing this data in the current AI climate would be far more consequential than it might have been 5 years ago. In the era of LLMs, behavioral search data is not just useful for traditional search ranking — it is valuable training fuel for AI systems that could directly compete with Google's own Gemini-powered search experience.

Privacy Concerns or Strategic Maneuvering?

Critics will inevitably question whether Google's privacy argument is genuine or merely a convenient shield for its business interests. The truth likely sits somewhere in between.

The privacy risks are real. Search data is among the most sensitive information a tech company holds. Queries related to health conditions, financial situations, legal troubles, and personal relationships flow through Google's search engine every second. Sharing even anonymized versions of this data carries genuine risks of re-identification, particularly when combined with other datasets.

However, Google's track record complicates its position as a privacy champion. The company has faced repeated fines and regulatory actions in Europe over its own data practices, including a landmark $170 million fine from the FTC for collecting children's data on YouTube and multiple GDPR enforcement actions. Regulators may be skeptical of a company that monetizes user data at unprecedented scale suddenly positioning itself as the guardian of user privacy.

The EU will need to navigate a genuine tension. The DMA was designed to open up digital markets, but the GDPR was designed to protect personal data. When these 2 frameworks collide — as they do in this case — there is no easy resolution.

What Happens Next

Vassilvitskii's meeting with EU officials will be a critical moment. Google's alternative proposal could offer a path to compromise — perhaps involving aggregated, differential privacy-protected datasets that provide competitive insight without exposing individual user behavior.

The European Commission has several options:

• Accept Google's alternative if it meaningfully addresses competitive concerns
• Reject it and enforce the original demand, potentially triggering legal challenges
• Negotiate a middle ground, perhaps with tiered data access and strict privacy safeguards
• Impose fines of up to 10% of global revenue for DMA non-compliance — which for Google parent Alphabet could exceed $30 billion

The stakes extend far beyond Europe. Whatever framework emerges from this dispute will likely influence how regulators in the United States, United Kingdom, Japan, and other jurisdictions approach the same fundamental question: can a dominant search engine be required to share the data that makes it dominant?

Looking Ahead: A Precedent for the AI Era

This confrontation between Google and the EU is about more than search data. It is a test case for how societies will regulate the competitive dynamics of AI.

The companies building the next generation of AI products — from search engines to autonomous agents — depend on massive datasets to train and refine their systems. If regulators can compel dominant players to share certain categories of data, it could fundamentally alter the economics of AI competition. If they cannot, the current concentration of power among a handful of tech giants may become permanent.

Google's aggressive stance suggests the company views this as an existential issue. The search giant has typically preferred negotiation and gradual concession in its dealings with European regulators. This time, it drew a hard line. That alone tells you how much is at stake.