In short: small websites are not doomed by AI search, but the model of publishing lots of useful articles, ranking in Google and living on referrals is becoming much riskier. The weakest position belongs to sites that depend almost entirely on search traffic and publish information AI can summarize with little loss. Better positioned are specialist sites with proprietary data, tools, products, local expertise, recognizable authors or a returning audience.
Large sites once enjoyed an almost automatic advantage through scale. Generative search reduces the value of scale when that scale consists mostly of interchangeable information, while increasing the relative value of things that are difficult to copy or compress.
Search traffic is under real pressure
A Pew Research Center study found that users clicked traditional Google results in about 8% of searches with an AI summary, compared with 15% without one. Links inside the AI summary itself received about 1% of clicks.
The Reuters Institute reports that surveyed media leaders expect search traffic to decline by roughly 43% over the next three years. Utility and lifestyle content appears especially exposed because users often need only one quick answer.
Small does not mean weak
AI search changes the traditional advantage of scale. Google says AI Overviews and AI Mode can use query fan-out, running several related searches and surfacing a wider set of sources. That does not guarantee traffic, but it means a small site is not excluded simply because it has fewer pages.
Google’s 2026 guidance does not require special AEO markup. It emphasizes useful, trustworthy and unique, non-commodity content. This can favour focused specialist resources that know one subject more deeply than a general portal.
| Small-site model | Google dependence | Can AI replace the core value? | Resilience |
|---|---|---|---|
| Generic SEO article site | High | Very easily | Low |
| Niche expert blog | Medium | Partly | Medium/high |
| Local service site | Medium | Can replace the answer, not the service | High |
| E-commerce store | Medium | Can replace information, not the purchase | Medium/high |
| Proprietary database | Low/medium | Difficult | High |
| Calculator or online tool | Low/medium | Difficult | High |
| Forum or community | Medium | Difficult | High |
| Author-led project | Low/medium | Partly | High |
The critical metric is dependence on someone else’s audience
A site can have 50,000 monthly visits and still be fragile if almost all of them come from Google. Another site can be much smaller but have a meaningful share of direct, newsletter, bookmark and branded-search traffic. From a risk perspective, the second site is often stronger.
Reuters Institute finds that subscription businesses with strong direct traffic see a clearer path to sustainability. For a small publisher, the goal is increasingly not only to rank once, but to become a source users remember by name.
AI can also strengthen small sites
Generative AI lowers the cost of work that once required a bigger team. A single author can analyze documents faster, translate drafts, write simple tools, process datasets, create charts and automate routine tasks.
The same AI that threatens generic SEO sites can help a specialist resource build databases, trackers, calculators or other products that would once have been too expensive.
Money may move differently as well
On September 30, 2026, Cloudflare launched the beta version of Pay Per Use. It is too early to assume that small sites will soon earn meaningful AI licensing revenue, but standardized payment infrastructure matters precisely because small publishers cannot negotiate bespoke contracts with every platform.
Which small sites are most exposed
The highest-risk model depends on one loop: find a keyword, publish an optimized page, get a click and monetize it through ads or affiliate links. Such sites may continue to work, but the structural advantage of the model is weakening.
The dangerous position is to be too small to win by scale and too generic to offer anything users cannot get elsewhere.
A survival test: what remains if Google referrals fall by half?
| Question | Stronger position | Weaker position |
|---|---|---|
| Do users search for the site by name? | Direct and branded traffic exists | Almost all users arrive through generic queries |
| Does the site own something unique? | Data, tool, research, database | Mostly repackages public information |
| Can AI deliver all useful value without a visit? | Only part of it | Almost all of it |
| Is there revenue without an ad pageview? | Product, service, subscription, licensing | Mostly advertising |
| Do users return? | There is a reason to come back | The relationship ends after one answer |
| Does the page gain value over time? | Data and history are updated | The text simply ages |
The small website of the future may look less like an article archive
Articles will remain important because they explain, attract and build trust. But strong small sites will increasingly add assets that accumulate value: live databases, trackers, maps, calculators, transparent rankings, archives, subscriptions or communities.
For an author-led site, the irreplaceable asset may be the author’s expertise and reputation. For a local resource, it may be detail national publishers cannot match. For a technical niche, it may be years of practical observations.
Not every small site will survive, but that is not the end of the small web
Some small information sites will close or move into minimal-maintenance mode if traffic and revenue fall below an economic threshold. Projects built mostly around interchangeable SEO content are especially exposed.
At the same time, AI lowers the cost of building sophisticated digital products. The dividing line will therefore not simply be between large and small sites. It will be between interchangeable text repositories and resources that users know, use and still have a reason to visit after they already know the short answer.
