By Olivia Hart

The ranked results page stops being Google's default first response for most U.S. queries between 2028 and 2030. That is a forecast, not a finding, and it comes from search industry analyst Duane Forrester.

Forrester places the U.S. first. The UK and EU follow a year or more behind. The flip is not about deleting the SERP. It is about demoting it one click away, still reachable on purpose but no longer the page you land on by default.

What did Google actually change at I/O 2026?

A widely repeated claim says Google made AI Mode the global default at I/O 2026. Google's own announcement says something narrower. Gemini 3.5 Flash became the new default model in AI Mode. The search box got its biggest rebuild in 25 years. Google also said you will continue to get a range of results from Search, just like you do today.

Two weeks later, a Chrome Canary flag briefly sent address-bar searches straight into AI Mode. Google's VP of Search Engineering Rajan Patel called it an error. "We're not planning to make AI Mode the default for Chrome searches," he wrote. The default has not flipped yet, and Google has not committed to flipping it any time soon, but the room where the ranked list would move into was already built when Google quietly added a simple Web filter in May 2024.

How big is the answer layer already?

Two measurements tell the story. In March 2025, a Pew Research Center study of 900 U.S. adults found 18% of Google searches produced an AI summary. Users clicked a traditional result 8% of the time with a summary, against 15% without one. By May 2026, Similarweb's Generative AI Landscape put AI Overviews on 43% of U.S. searches. That is up from 15% a year earlier.

The two numbers use different methods. They should not be stitched into one line. But they point in the same direction at roughly the same speed, from under a fifth of U.S. searches to over two-fifths in about 14 months. Takeaway: the answer layer is growing fast, but the click-through rate is roughly half what it used to be.

Then the conversational layer. Google says AI Mode passed 1 billion monthly users within a year of launch. Queries more than doubled every quarter, per Google. Set that beside Similarweb panel data, which found AI Mode accounted for 0.34% of Google searches from January through April 2026. A billion people have tried the answer-first surface. Almost none have made it where they start.

Similarweb also measured AI Mode generating outbound referrals on 1.6% to 2.5% of queries. Traditional search sent referrals on 17% to 19% of queries. That gap is the clearest single picture of what a default flip would mean for any site that depends on the click.

Behavior is drifting in the expected direction. Similarweb recorded a 5.4% increase in average Google query length after AI Mode arrived, which matters because longer queries are the ones most likely to trigger a generated answer.

When does money force the flip?

Google has every capability it needs to flip the default tomorrow. It has not. The reason is money. The ranked page still carries the most proven ad inventory in advertising history. The answer layer's ad formats are months old.

At Google Marketing Live in May 2026, Google introduced ads inside AI Mode responses. Gemini can build the ad around the user's question. Google's own recap read: "We aren't just bringing ads to AI Search experiences; we are reinventing what an ad is in that context."

In Q2 2026 earnings remarks, Sundar Pichai said Google had "reduced the cost of AI Mode responses to its lowest level since launch" and "brought together AI Overviews and AI Mode into one seamless Search experience." Network advertising, the revenue Google earns on other people's pages, declined 1% that quarter. The part of its business that depends on sending people to the open web is already the part that is shrinking.

The trigger becomes visible. Google flips the default when ad yield per query in the answer layer reaches rough parity with the ranked page. Before that point, flipping costs Google money. After it, keeping the old default costs Google money. That single crossover turns this whole transition from a forecast about direction into a calendar event about timing.

What slows the transition down?

Three regulatory processes now bear directly on whether a ranked layer stays visible. They run on different clocks.

• UK: The Competition and Markets Authority issued binding conduct requirements on June 3, 2026. Publishers can opt out of AI Overviews and AI Mode without losing rank in standard search. Google has nine months to comply, putting the deadline around March 2027.
• EU: The European Commission opened a formal investigation in December 2025 into whether AI Overviews and AI Mode use publisher content without compensation. The case is still open.
• U.S.: The Justice Department and plaintiff states filed a cross-appeal in July 2026. They argue Judge Mehta should have considered banning Google's payments to distributors and warned that "a competitive window may close."

A ranked layer in the UK is no longer just a product choice. It is where opted-out publishers are entitled to remain visible. Burying it would invite exactly the scrutiny the rule was written to create.

A U.S. flip while UK and EU rules are still settling produces a staggered rollout by market, which is why the forecast splits dates by region rather than giving one global number.

What would prove this wrong?

The forecast comes earlier if Google reverses its position on the Chrome or Android address bar, if AI Mode's measured share passes roughly 10% in independent panel data, or if Google begins reporting answer-layer ad revenue as its own line.

It comes later if AI Mode's share plateaus in the low single digits through 2027, if advertisers report weak returns from AI Mode placements, or if the appeals court extends remedies to generative AI distribution in ways that constrain default placement.

Forrester's math floor is striking. If AI Mode's 0.34% share kept doubling every quarter, it would pass 40% in about seven quarters, landing in late 2027. Nothing doubles forever. That number is a floor, not a prediction. Its main use is showing the front edge of his range is not conservative for its own sake.

Watch the ad-yield crossover in the answer layer and the next UK compliance milestones around March 2027. Either signal moves the date.