Research brief · Search · October 2026 · 16+ forecasts & datasets

AI search
vs Google

AI answers are eating the click. Google still runs five trillion searches a year and just absorbed AI into its own results. We read the forecasts, from Gartner to Bain to Alphabet's own filings, and projected the next ten years, 2026 to 2035, with a call on what prevails.

By Christos Papadimitriou · House of 47 Research
Last updated:

See the decade
~90%Google's search share, 2025, first dip of the AI era
2.5BChatGPT prompts/day vs Google's ~14B searches
68%US searches with zero clicks, early 2026, from 60% in 2024
8%Click rate when an AI summary shows, from 15%
A digital search interface
The projection · 2026 → 2035 · global search

The next decade

The disruption is real where it counts for the open web, the click and the referral economy. But at the level of who controls distribution and revenue, the incumbent is winning for now. Here is how the ten years most likely run, and the two curves underneath it.

2026 → 2028 · Hybrid

AI answers go default

60%+

AI Overviews and AI Mode become standard on most Google queries. Zero-click climbs past 60%. Publisher referral traffic keeps falling. GEO enters marketing budgets.

2028 → 2031 · Fragmentation

Search splits by task

Crossover

By converting-traffic value, AI search rivals organic in several verticals. Google defends total share through distribution and query growth. Ad revenue grows, but slower and more expensive.

2031 → 2035 · Equilibrium

Consolidation, not a killer

Both

Google stays the largest single gateway, no longer the near-monopoly funnel. Answer engines own a real slice of research and shopping intent. SEO becomes GEO.

The wildcard

Regulation

?

Antitrust remedies on default-search deals and data-sharing could redistribute search share faster than any product shift. The moat, not the model, is the real risk to Google.

The two curves that decide it

The click leaves the page

Cross-check · Gartner, Bain, Pew, eMarketer, Alphabet filings

Where they agree

  • All
    The click is collapsing

    Pew: users click a link 8% of the time when an AI summary shows, versus 15% without. Bain: about 60% of searches now end without a click. The open-web referral is the casualty.

  • All
    Google still owns the scale

    ~90% share, five trillion searches a year, roughly 14 billion a day. ChatGPT's 2.5 billion daily prompts are the challenger, but still a fraction of that.

  • All
    AI search is growing far faster

    AI referral visits up ~357% year on year, ChatGPT queries up ~150%, Gemini from 90M to 750M monthly users in about 15 months. The threat is the trajectory, not the share.

  • All
    Google is absorbing AI, not being replaced

    AI Overviews reach 2 billion monthly users, and AI Mode passed 1 billion monthly users by I/O 2026. Google says AI features drive ~10% more queries where they appear. It is defending from inside the result.

  • Most
    The business model is the real fault line

    Google's $198B search-ad engine monetizes clicks. An AI answer shows fewer links and fewer ads. Answer engines lean on subscriptions and commerce, not display.

  • Most
    SEO is becoming GEO

    Getting found shifts from ranking links to being cited inside AI answers. The Princeton GEO study shows the right tactics lift visibility in AI responses up to 40%.

Where they disagree

What happens to Google by 2035?

From a replaced incumbent to an unshaken monopoly. Each camp placed by how much it thinks Google loses.

Google displacedGoogle holds

When does AI search overtake traditional organic?

By value of converting traffic, not raw volume. Placed by forecast year, 2026 to 2035.

SoonLate or never

Who wins the money, not just the query?

Ads on the page versus commerce and subscriptions inside the answer.

AI monetizesGoogle's ads win

Gartner 2024 (search volume -25% by 2026) · Bain 2025 (~60% zero-click, 15–25% traffic loss) · Pew 2025 (8% vs 15% click rate) · Ahrefs 2025 (up to -58% CTR with AI Overviews) · Semrush 2025 (LLM traffic overtakes organic ~2027–28) · eMarketer 2025 (~163M US GenAI users by 2029) · Morgan Stanley 2025 (Overweight, Google keeps the base) · Princeton GEO 2024 (up to +40% visibility in AI answers) · SparkToro 2026 (zero-click 68% in early 2026; AI Mode past 1B monthly users).

Probability of outcome · House of 47 synthesis

The odds

Synthesis estimates across the forecasts and datasets below, scenario estimates, not market forecasts. The through-line: the open-web click is structurally impaired, while Google's distribution and revenue prove more durable than the disruption headlines imply.

Intent by surface · who owns each kind of search in 2035

Who owns which query

Search does not die, it fragments. Different intents settle on different surfaces. Google holds the fast, navigational and local queries; answer engines take open-ended research and, increasingly, shopping.

Query typeGoogleAI engines2035 lean

"Google" = classic SERP + AI Overviews + AI Mode. "AI engines" = ChatGPT, Perplexity, Gemini, Copilot. Lean is the House of 47 call on where each intent mostly settles by 2035. Directional, synthesised from the forecasts above.

House of 47 read · What to do about it

Get cited, not just ranked

Optimize for the answer

The discipline is shifting from ranking a link to being the source an AI quotes. Statistics, cited authority and expert quotes are what lift you into the answer. This move from SEO to GEO is the core work of Resocial.

Build the brain AI reads

Structured, factual, well-sourced content is the raw material answer engines cite. The sites that win are the ones the models can trust and lift cleanly. Depth and proof beat keyword stuffing.

Own the audience, not the ranking

If the click is leaving, the direct relationship is the moat. Email, community and brand search compound while referral traffic erodes. Be the name people type, not the link they skip.

Measure across surfaces

One ranking is no longer the scoreboard. Track citations and visibility across Google, ChatGPT, Perplexity and Gemini. Search is everywhere now, so measurement has to be too.

Base case

Plan for consolidation with fragmentation. Google stays the biggest gateway but loses its monopoly on intent. AI answers become the primary interface of search rather than a replacement for Google. Optimize to be cited everywhere, not ranked in one place.

Tail

Stress-test for a faster unbundling. If antitrust breaks the default deals, or a native AI assistant wins the phone, share could move quicker than any product trend. Cheap to prepare for, expensive to be caught flat.

Google will keep the search box. It is losing the click underneath it.

Sources · figures checked against primary and reputable reporting

Sources

SparkToro 2026, zero-click 68% early 2026, AI Mode past 1B users, via Search Engine Land

Gartner 2024, search volume -25% by 2026, prediction

Pew Research 2025, 8% vs 15% click rate with AI summaries, study

Bain & Company 2025, ~60% zero-click, 15–25% traffic loss, insight

SparkToro 2024, zero-click study, data

Ahrefs 2025, AI Overviews cut CTR up to 58%, study

StatCounter 2025, Google ~90% share, market share

Google 2025, 5 trillion searches a year, via Search Engine Land

Alphabet Q3 2025, AI Mode 75M+ DAU, query growth, earnings

Alphabet FY2024 10-K, $198B search revenue, SEC filing

TechCrunch 2025, AI Overviews 2B users / AI Mode 100M, report

OpenAI 2025, ChatGPT 800M weekly users, via TechCrunch

ChatGPT 2025, 2.5B prompts a day, report

Perplexity 2025, $18B valuation, 780M queries/mo, report

Gemini Q3 2025, 650M monthly users, via 9to5Google

Similarweb 2025, AI referrals +357% YoY, via Digiday

Adobe Analytics 2025, GenAI retail traffic surge, report

eMarketer 2025, ~163M US GenAI users by 2029, forecast

Morgan Stanley 2025, Overweight, Google keeps the base, via TipRanks

US v. Google 2025, antitrust remedies, ruling

Princeton GEO 2024, +40% visibility in AI answers, paper

Changelog · what changed in this brief, newest first

Changelog