Google is quietly introducing a technical change to its search engine that could have significant consequences for the SEO industry, data analysis and companies that use information extracted from its results pages.
The company has started rolling out links in the google.com/goto format in organic results. Instead of linking directly to the page that appears on Google, some results first go through an address controlled by the search engine itself, which then redirects the user to the corresponding site.
The change started being detected during tests in mid-2026, and Google confirmed this week that the system is being rolled out. Asked by trade media, the company explained that it has a long track record of implementing technical measures against new forms of abuse and that it regularly adopts mechanisms designed to protect its services and users. (DemandSphere)
For a person doing a normal search, probably nothing changes: they click a result and land on the expected page. For machines, however, the story is different.
A new barrier to scraping Google
Traditionally, tools that analyze results pages (known as SERPs) could directly identify the URLs of the sites ranking on Google.
The new mechanism introduces an intermediate layer.
Previously, a scraper could find in the search engine’s HTML an address like:
https://example.com/product
Now it may find something like:
https://google.com/goto?url=CAESBQgAEP8B...
The parameter used by /goto doesn’t necessarily contain the final URL in directly readable form. As a result, an automated tool may need to make an additional request and resolve the redirect before knowing the real destination. (Autom)
The difference seems small, but at scale it can be considerable.
Platforms that monitor millions of keywords daily could need to make more requests, maintain additional infrastructure and develop new mechanisms for resolving the URLs.
The potential result: more processing, more infrastructure and a higher cost for each SERP analyzed.
SEO specialist Alex Amigo warned about precisely this effect: if the system takes hold, tools that depend on scraping Google could face higher operating costs. (LinkedIn)
Google had already done something similar
The strategy isn’t entirely new.
For years Google has used redirect mechanisms such as google.com/url?q= to log certain clicks made from its results.
The technical difference is important: in those systems the destination URL could be found within the link’s parameters. With /goto, the identifier can behave like an opaque reference that has to be resolved by Google before the final destination is known. (Autom)
The change therefore gives Google greater control over the path between the SERP and the destination site.
And it comes at a particularly sensitive time for those who obtain data programmatically from the search engine.
Google will also shut down one of its search APIs
There’s a second move that makes this change especially relevant.
Google confirmed that its Custom Search JSON API is closed to new customers and will stop being available to existing customers as of January 1, 2027.
This tool lets you obtain search results through REST requests and receive them structured directly in JSON format, precisely avoiding the need to extract information from a SERP’s HTML. (Google for Developers)
Currently, existing customers can make 100 free queries a day and pay $5 for every 1,000 additional requests, with a limit of 10,000 queries a day. Those terms will remain in place only during the transition period. (Google for Developers)
As an alternative, Google recommends Vertex AI Search for systems that need to search across up to 50 domains. For applications that require searching the whole web, the company is inviting developers to register their interest in a new full search solution. (Google for Developers)
The possible impact on Semrush, Ahrefs and rank trackers
The scenario raises a particularly relevant question for the SEO industry.
Much of the ecosystem of ranking tools needs to obtain, directly or indirectly, enormous amounts of information about Google’s results: positions, domains, URLs, snippets and special SERP features.
That doesn’t mean platforms like Semrush, Ahrefs or other providers will necessarily stop working. These companies have their own infrastructure, crawlers, databases and methodologies.
But the change may raise the technical complexity of obtaining certain data from Google.
If each result needs additional steps to identify its destination, the marginal cost of monitoring millions or billions of results can increase.
In that scenario, the effects could end up reaching customers through higher prices, stricter query limits or changes in how often rankings are updated.
There’s another interested party: artificial intelligence
The move doesn’t happen in a vacuum either.
Search engine results have become an extremely valuable source for AI agents, automated research systems, answer engines and models that need to look up up-to-date information on the internet.
That’s why /goto can be read as part of a broader trend: automated access to the web’s information is becoming an increasingly controlled and economically valuable resource.
Google hasn’t said the system was created specifically to target artificial intelligence companies. Its official explanation talks about measures against evolving forms of abuse. Attributing the change exclusively to AI would therefore be speculative. (DemandSphere)
Technically, however, the mechanism can affect precisely the systems that try to process search results automatically.
The battle is no longer just about appearing on Google
For decades, the discussion around the search engine was focused on one question: who managed to appear first.
The new discussion could be different: who can access the data behind those rankings, under what conditions and at what price.
Google is introducing an additional layer between its results and those who try to analyze them automatically. At the same time, it will retire an API that for years made it possible to access structured results programmatically.
Separately, both changes may look like technical decisions.
Together they show something more interesting: on an internet where search engines, SEO platforms and artificial intelligence systems need to consume ever larger amounts of information, SERP data is no longer just search results; it’s also becoming strategic infrastructure.
And controlling that infrastructure may end up being as important as controlling the search engine.