French Newspapers File Competition Complaint Over Google AI Overviews

Full analysis

Nearly 300 French newspapers just asked France's competition regulator to make Google pay for AI Overviews, the AI-written summaries that now sit at the top of search results and answer the query so the user never clicks through. The publishers want a compensation deal before the traffic bleeds out. This is the opening shot in a fight that will decide who pays whom for open-web content in the AI era, and every operator whose revenue depends on Google referral traffic has a stake in the answer.

What's actually being decided: not whether French papers get a check. Whether the Autorité de la concurrence forces Google to negotiate collectively with publisher coalitions over AI-scraped content, and whether that template travels to Germany, Spain, and the rest of the EU.

Reversibility: Type 1 for Google. A collective-bargaining precedent under EU competition law is hard to unwind once set. Type 2 for individual publishers, who can adjust traffic strategy quarter to quarter.

Forcing function: whether the regulator grants interim measures (a provisional order that binds Google while the case drags on) or lets Google appeal into multi-year delay.

The Market Analyst. For a generalist: this is about whether a regulator can make the biggest search company share revenue with the news sites it summarizes. Google's French ad money is not the thing at risk here. Branded and commercial searches, the queries that actually carry yield, mostly don't trigger AI Overviews. What the complaint creates is regulatory drag that makes rolling AI Overviews across Europe more expensive to defend. Modest headwind for Alphabet's EU story, not an earnings event. The clear winners sit elsewhere: walled gardens and retail media networks never lived on open-web referral traffic, so every session Google keeps for itself pushes budgets toward inventory those players already own.

The Skeptic. Steelman the case against this mattering. French publishers ran this same play over neighboring rights in 2019 and settled for deals most people called symbolic. Google's method is known: wait, negotiate with the biggest names one at a time, let the coalition crack. For this to move publisher economics you need the Autorité to impose interim measures before Google appeals, and that is rare. Worse for the publishers' own argument: AI Overviews eat soft informational traffic, the "what year did X happen" sessions that were already monetizing at basement CPMs and never converting to subscriptions. Losing sessions you barely earned on is not the catastrophe the coalition is selling.

The Operator. For a generalist: the people who actually run ad sales at these sites have to plan for fewer visitors starting now. Model a 15 to 25 percent referral haircut over the next two quarters as AI Overviews become a habit for French users. US data on Google's AI search already shows click-through decay on informational queries, so this is not a thought experiment. Fewer landing-page visits means fewer impressions, which means your programmatic revenue falls even if CPMs hold. First contracts to trip are audience-extension deals tied to site traffic thresholds. Revenue ops should stress-test yield against lower session volume this week, and the direct-deal and subscription teams should be filling pipeline before the traffic curve confirms the slide, not after.

The Customer / End User. Two customers here, and they pull apart. The French reader loves AI Overviews. They get the answer without clicking, which is the entire point, and no complaint from 300 newspapers changes that behavior. The advertiser buying open-web inventory is the one who quietly loses: the contextual and content-adjacency products that need a reader on a publisher page have less page to sit next to. Nobody in this story is asking for the open web to shrink. It shrinks anyway, because the product is better for the person doing the searching.

The CFO. Real cost, not the press-release version. For a French publisher, the compensation from any eventual deal is a rounding error against the referral revenue at risk, and it arrives years late after appeals. The expensive mistake is treating a possible regulatory win as a reason to defer the hard work. If you plan your 2026 revenue around a settlement check, you have bet the P&L on the Autorité doing something it rarely does quickly. The disciplined move is to write the traffic decline into the forecast now and treat any compensation as upside, not baseline.

Where the council splits:

1. Does the lost traffic matter? The Skeptic says the vanishing sessions were low-value informational clicks that never paid. The Operator says a 15 to 25 percent volume cut still strips impressions off the top line regardless of quality. Both can be right: revenue falls, but by less than the raw traffic number suggests.

2. Does the complaint produce durable money? The Market Analyst and Skeptic see 2019 repeating: symbolic deals, fragmented coalition, no structural change. The strategic read is that the prize isn't the check but a collective-bargaining precedent that spreads across the EU. The hinge is one regulatory decision, not the complaint itself.

3. Who actually benefits? Every persona lands in the same uncomfortable place: the reader wins, walled gardens and retail media win, and the open-web publisher and its advertisers absorb the cost.

What it hinges on: whether the Autorité grants interim measures before Google appeals. Grant them, and Google faces a real cost to running AI Overviews in Europe, which either slows the rollout or pushes Google toward licensing content directly at scale. Refuse them, and this is 2019 again: years of process, individual side deals, symbolic money, and publishers left to solve their own traffic problem.

The council leans toward the Skeptic on the money and the Strategist on the stakes. The complaint probably won't put meaningful revenue in French publishers' accounts on any timeline they can plan around. But it plants the precedent that Google most wants to avoid, which is why Google will fight the collective-bargaining mechanism far harder than it fights any dollar figure.

What to verify before acting: watch for whether the Autorité signals interim measures. That single procedural move separates "leverage" from "symbolic," and it will show up long before any compensation number does.

Prediction: France's Autorité de la concurrence will NOT impose binding interim measures forcing Google to pay French publishers for AI Overviews before the end of Q2 2026; any relief will be procedural or deferred into Google's appeal process.

Confidence: Medium. The 2019 neighboring-rights fight is the template, and it shows how slowly the regulator actually moves against Google.

Why: The one thing that would turn this complaint into real, plannable revenue is a provisional order that binds Google while the case runs, and French regulators have historically not moved that fast against Google. In the 2019 neighboring-rights matter the process stretched for years and resolved into individual deals most observers called symbolic, which is the pattern Google engineers by negotiating with the largest players one at a time and letting the coalition fragment. For the opposite to happen, the Autorité would have to break its own precedent and grant emergency relief on a feature that launched only last month, before any French traffic data proves harm, which regulators are reluctant to do without a demonstrated, quantified injury.

Revisit by 2026-06-30: We're right if no binding interim measures have been imposed and the matter is in normal review or appeal. We're wrong if the Autorité issues a provisional order requiring Google to compensate French publishers or to pause AI Overviews before then.