The Extra Hop That Could Make AI Search More Expensive

Key takeaways

• Opaque search links can make destination addresses harder to extract.
• Resolving each link separately can add requests, delays, and retries.
• AI services that collect URLs from Google results could face additional costs.
• Assessing the competitive impact requires examining access rules and alternatives.

An extra step in a search link can be almost invisible to the person clicking it. For an AI search service collecting web sources at scale, that same step could add requests, delays, and failures to handle.

When a link becomes a claim ticket

A destination URL often gives you useful clues before you click. The domain identifies the website. The path might contain an article title or a document number. Software can read that address directly and add it to a list.

An opaque URL makes that harder. If the original address is replaced by an identifier that reveals nothing about the destination, a program may need to contact an intermediary server to find out where the link leads. Think of a coat-check ticket: it identifies something, but you need the desk to retrieve it.

Suppose a Google search result routes through an address such as google.com/goto. If that link conceals the destination, collecting the original URL could require an extra step.

The path name alone doesn’t establish that behavior. The destination might still be recoverable from information in the link. What matters is how the address is constructed and what happens when a client requests it.

Nor does a redirect automatically qualify as a scraping defense. Redirects are a routine part of the web. Their effect on collection depends on the implementation.

Small requests can become a large bill

Scraping means using software to extract information from web pages. For search results, that information can include titles, snippets, and destination URLs.

When the original URLs appear directly in the results page, a program can read them from that page. If each destination must be resolved separately, collecting the same information takes more requests.

Consider 100 search results. Assuming each needs one additional request to reveal its destination, that’s 100 extra requests just to assemble the URL list. Fetching the underlying documents is another job.

As collection grows, those additional requests can mean more waiting and more failed attempts to retry. Rate limits or authentication requirements could increase the burden further.

Opacity alone doesn’t stop scraping. Software can follow redirects too. But a design doesn’t have to make collection impossible to make it more expensive.

For a search operator, these mechanisms can help manage heavy traffic. Whether a particular implementation serves that purpose—and how effectively—depends on its operating rules.

AI search feels the cost unevenly

An AI search service needs to find relevant documents before it can build an answer from them. Where it gets those documents determines its exposure.

A service extracting destination URLs from Google results could have more work to do. Resolving links might slow retrieval. Failed resolutions could leave fewer sources available for an answer.

A service using its own web crawler or another search index would be less directly exposed to that particular change. An index is the organized collection of documents a search engine uses to find results.

That creates a potential cost asymmetry. An outside service may have to resolve links one by one, while the search operator already knows their destinations.

This could affect competition among AI search services. It does not, by itself, establish that a URL design targets a particular rival. The technical burden and the operator’s intent are separate questions.

The access rules matter more than the URL

To assess the competitive effect, start with who faces restrictions and when.

A limit triggered by excessive request volume has different implications from special conditions applied to a particular service. The availability of other access routes matters too. An official way to obtain results could change the picture, depending on its price and terms. Technically available access can still be commercially impractical.

There is also a user-facing issue. If a link obscures its destination, people may have less information when deciding whether to visit a site. Whether the search interface or browser displays that destination elsewhere becomes important.

The broader question is who sets the cost of using search results. A search operator’s defensive choices can change the economics for services built around those results. A legitimate traffic-management purpose doesn’t settle the separate question of competitive impact.

Opaque links could raise collection costs, but their significance depends on whom they restrict and what alternatives remain. A small decision about how a link works can help shape who can afford to build the next search service.