Google launched its May 2026 core update on 21 May, and it’s still rolling out. If your rankings or traffic have moved in the last few days, this is probably why.
Here’s what we know, what’s unusual about this one, and what you should (and shouldn’t) do while it’s in progress.
What Google actually said
Not much. Google updated its Search Status Dashboard with: “Released the May 2026 core update. The rollout may take up to 2 weeks to complete.”
On LinkedIn, they added a line about it being designed to “better surface relevant, satisfying content for searchers from all types of sites.” That’s been the standard language for every core update for years. No companion blog post, no guidance specific to this update, and no new ranking systems announced.
That matches how Google handled the March 2026 core update too.
Why this one is getting more attention than usual
A few things make this update stand out from a typical rollout.
The timing. This core update kicked off the day after Google I/O wrapped up, where Google announced what it called “the biggest upgrade to our Search box in over 25 years,” along with a wave of AI-powered search features. At I/O, Google unveiled an AI-powered overhaul of Search built around an “intelligent search box” and interactive, agent-driven experiences that de-emphasise traditional blue links in favour of AI-generated interfaces and autonomous information agents.
Google I/O and the May core update are technically separate initiatives with different purposes, and Google has been clear on that. The core update is an under-the-hood adjustment to ranking algorithms designed to filter out low-quality web content. But they overlap conceptually: quality signals that feed traditional search also feed the AI products Google is building.
The gap between updates. This rollout arrives just six weeks after the March 2026 core update wrapped up, following weeks of ranking volatility. Many SEOs had expected more frequent updates in 2026, but the pace didn’t accelerate the way some predicted. The six-week window also means that any ranking movement you’ve seen since early April may be residual March effects, not early signs of this update.
The AI Mode complication. This update landed the same week Google deployed Gemini 3.5 Flash as the default model in AI Mode globally and confirmed AI Mode passed one billion monthly users at Google I/O 2026. That creates an unusual situation: a site might hold its rankings through the core update and still see click declines because AI Mode and AI Overviews are absorbing more of the traffic on those queries. Separating core update impact from AI feature impact is going to be harder than usual this cycle.
What we’re seeing so far
Early data suggests movement across news, e-commerce, and informational queries. Some sites moved up. Many dropped without warning. That part is normal for core updates.
The December 2025 core update ran for 18 days and produced some of the most severe ranking disruption documented in recent years, with publishers reporting traffic losses between 70% and 85%. Recovery from that update, for many affected sites, has not fully materialised across the March cycle. Sites still carrying damage from December or March are now in an especially difficult position.
Scaled AI content: the crackdown is getting sharper
One of the clearer patterns from this update, and from the March cycle before it, is that Google’s enforcement on scaled content abuse is tightening.
Companies using large AI-generated blog subfolders now appear to be hit hard right as the update landed. One group has seen it’s English blog presence effectively disappeared from search results while the broader domain remained visible. That pattern lines up with either enforcement actions or an algorithmic demotion aimed squarely at low-value, bulk-published content.
What made that kind of content vulnerable comes down to a few consistent characteristics: dense, repetitive AI text with low usefulness and poor readability; no original media or supporting assets; broad topic sprawl with no clear expertise or focus; and pages that look designed for search volume capture rather than actual readers.
The important nuance here is that Google does not penalise AI-generated content as a category. What it penalises is unhelpful content, and a lot of AI-generated content happens to be unhelpful: thin, derivative, and lacking any original insight. If you use AI in your content workflow, the question to ask is not “is this AI-written?” but “does this genuinely help the reader in a way they couldn’t get from three other sites?” If the honest answer is no, that’s where the vulnerability sits, not in the tool itself. Now we also would say that there is AI fatigue where people don’t like to watch clearly written AI material but if it’s useful it can still rank.
A pattern that keeps showing up in penalty post-mortems is the same each time: a site decides to scale content using AI, publishes dozens or hundreds of articles without meaningful human review, the traffic looks promising for a few weeks, and then a core update rolls through and the entire domain takes a hit. Not just the AI articles. Everything. The site owner assumes Google detected the AI content, but the actual problem was quality. Google reassessed the domain’s overall content signals and found them lacking.
Local SEO: directories losing ground on “near me” searches
Early reports from this update suggest that directory and aggregator sites that previously dominated local-intent queries are dropping from top positions for a wide range of “near me” searches. One example is a well-known directory site that had held positions one to three for essentially every “near me” keyword has noticed a considerable drop noticeably within the first 48 hours of the rollout.
The reason this makes sense isn’t complicated. Google already has Maps and Business Profiles to satisfy local intent quickly. Business Profiles include reviews, photos, hours, and direct contact details. A directory page often adds little beyond a list of providers and generic text. When the intent is explicitly local and service-based, Google may increasingly favour the actual provider over the middleman.
What this means in practice depends on where you sit:
If you operate a local service business, you may see improved visibility as directories slide, especially if your Google Business Profile is well maintained and your site has genuine local authority signals.
If you run a directory or aggregator, you may need clearer differentiation: better information, more specific content, and a reason to exist beyond collecting listings. A list of providers with thin descriptions doesn’t justify a position one ranking when the user could tap Maps instead.
If you do both, your strategy may need to split between brand-level authority and high-intent, provider-specific pages. The two audiences have different needs and Google is starting to treat them differently.
Google Search Console: the links report broke at the worst possible time
On the same day the May core update launched, a separate issue started surfacing across the SEO community: Google Search Console’s links report went haywire.
Some SEOs reported seeing their link counts drop by 80% or more overnight. Some accounts showed zero links entirely. Screenshots spread quickly across X and Bluesky, with practitioners checking client profiles and finding the same thing across multiple accounts. Within PrimaryRush, all our client accounts are reporting as standard and have not noticed this happening.
The response from Google came the same day. John Mueller acknowledged the reports and confirmed the team was investigating. This appeared to be a reporting bug, not evidence that your backlinks disappeared.
The link report and the core update have nothing to do with each other. The reports in Search Console are typically days delayed anyway, so any link changes from an algorithmic action wouldn’t show up instantly. If your link counts look wrong right now, it’s almost certainly the reporting, not your site’s actual backlink profile.
This matters because the timing was genuinely terrible. SEOs trying to diagnose core update impact were simultaneously looking at broken link data, which made an already noisy situation worse.
What the community is saying
The reaction across LinkedIn, X, Reddit, and forums has been predictably split.
Some people are relieved. Sites with strong editorial standards, original content, and a clear topical focus have reported ranking improvements. The sentiment in those cases is “finally.” Publishers who’ve been doing the right things and watching thin AI content outrank them for years are seeing that change.
Others are panicking. Traffic graphs pointing down, Search Console data looking broken, and an update that’s still actively rolling out make for a stressful combination. These are the people who use the line that “SEO is dead” when it clearly isn’t. Their strategies have not worked. The advice from experienced SEOs is consistent: wait for the rollout to complete before drawing conclusions or making changes.
The people doing well long-term in organic search tend to be the ones who stay adaptable rather than getting attached to one set of tactics. That’s not particularly new advice, but the pace of updates in 2026 is making it harder to ignore.
What to do (and not do) right now
Don’t make major content changes during the rollout. Core update volatility in the first week is high and unrepresentative. Rankings can swing significantly before settling. Any changes made during an active rollout will be evaluated under an algorithm that’s still shifting, making it almost impossible to know whether those changes helped or hurt.
Wait before drawing conclusions. Google recommends waiting at least one full week after a core update finishes before reviewing your Search Console data. Your baseline should be the weeks before 21 May, compared against performance after the rollout completes.
Don’t panic-publish. Adding more content while the update is live is unlikely to help. The pages Google is reassessing already exist. New pages won’t influence how existing ones are being evaluated.
Ignore your links report for now. Given the reporting bug that surfaced on day one, your link counts in Search Console are not reliable at the moment. Don’t use them to diagnose anything until the issue is confirmed resolved.
Do keep an eye on your tracking tools. Monitor organic traffic and keyword visibility trends, but hold off on acting until the rollout is confirmed complete.
What this update rewards (and what it doesn’t)
Based on patterns from this update and recent ones, here’s what tends to gain and lose ground.
Sites doing well tend to have original research or first-hand experience in their content, a clear focus on one topic area rather than publishing across every category, and a track record of accuracy and editorial standards. Google increasingly favours content that demonstrates real expertise and practical understanding. Articles that include real examples, case studies, expert insights, or first-hand experience are becoming more competitive than generic rewritten summaries.
Sites taking hits tend to be thin aggregators, pages that repackage information from stronger sources without adding anything new, and content that exists primarily to rank rather than to help someone.
One thing worth remembering: sites that haven’t changed at all can still be affected. If your competitors have published significantly better content since the last core update and gained great backlinks, rankings may fall even if nothing is wrong on your end. The ranking systems are always relative.
When the dust settles
Once the rollout completes (Google’s estimate is within two weeks of 21 May), do a proper before/after comparison in Search Console. Look at which pages moved, whether gains and losses cluster around specific content types or topics, and whether traffic changes track with ranking changes or show a disconnect that might point to AI Overviews absorbing clicks.
If pages dropped, the question to ask is whether they genuinely deserved to rank where they were. A drop doesn’t mean something is broken; it may mean Google found a better answer elsewhere. That’s worth knowing too.
The update is still live. More will follow.