The September 2026 spam update is still rolling out. If your GEO pages were built at scale, read this first.

Google's Search Status Dashboard lists the September 2026 spam update with a start date of 24 September 2026.

Spam updates usually finish in a couple of days. This one will not.

Search Engine Roundtable's Barry Schwartz reported on 1 October that Google told him this rollout would take "around two weeks, versus the typical two day roll out." He also reported a second visible wave of volatility from 30 September, and publishers describing swings in Discover and News traffic.

Two weeks from 24 September lands around 8 October. So if you are reading this before then, the update may still be moving your numbers.

Here is the complication for anyone buying GEO. A lot of GEO programmes are built on volume: a page per city, a page per competitor comparison, a page per question. Some of those are excellent. Some are the exact pattern Google's spam policy describes. And a spam update does not stay in classic search.

Step 1: Why a spam update reaches AI answers

Google's AI features page, last updated 10 December 2025, sets one requirement for a page to appear as a supporting link in AI Overviews or AI Mode.

The page "must be indexed and eligible to be shown in Google Search with a snippet". Then: "There are no additional technical requirements."

Read that in the other direction. If a spam system pushes a page out of ordinary results, that page is out of the pool AI Overviews and AI Mode draw links from.

PPC Land's 3 October report of an attendee recap by John Campbell of the agency ROAST says Google's Gary Illyes closed a three day Search Central event with the line "AI on Google is just SEO". That is a recap of a talk, so I would not hang a strategy on the wording. But it matches the documentation, and it matches what NovaTechRay found when it looked at how much of AI citation is just ranking.

So there is no separate AI index you can be safe in while classic rankings fall.

Step 2: Read the policy, not the summaries

Google's spam policies page, last updated 28 August 2026, defines scaled content abuse as "when many pages are generated for the primary purpose of manipulating search rankings and not helping users."

It then lists examples. Three of them describe common GEO tactics almost word for word:

• "Using generative AI tools or other similar tools to generate many pages without adding value for users"
• "Stitching or combining content from different web pages without adding value"
• "Creating many pages where the content makes little or no sense to a reader but contains search keywords"

Notice what is not on the list. AI-written pages are not banned. Many pages are not banned. Pages "without adding value" are the problem.

Most GEO vendors will tell you their pages add value. Ask them how they know.

Step 3: What Google has not said

This is where the reporting gets ahead of the evidence, so let me be plain about the gaps.

Google has not said what this specific update targets. The dashboard entry is a name and a date.

Google has not said why it takes two weeks this time. Anything you read about the reason is a guess.

Search Engine Roundtable also reported AI-translated Reddit pages surging again in several non-English results, and quoted Google as saying AI translation is fine "IF the content is high quality". That is one platform's pages in some markets. It tells you nothing about your local service pages.

I do not know whether this update hit GEO-style pages in particular. Nobody outside Google does yet.

Step 4: Date your drop before you blame anything

Most traffic drops in a two week window get blamed on whatever update is running. Some deserve it.

Here is the order I check things in:

1. Pull daily data, not weekly. In Search Console, set a date range from mid September to today and look at daily clicks and impressions.
2. Match the dates. A drop that starts on 25 September or 30 September is a candidate. A drop that started on 15 September is not this update.
3. Split by page type. Filter by URL pattern. If your templated city or comparison pages fell and your core service pages held, that is a pattern worth taking seriously.
4. Check Manual actions separately. Search Console has its own Manual actions report. Google's spam updates page describes automated systems, so an empty Manual actions report does not clear you.
5. Check AI visibility on the same pages. If you track AI Overviews or AI Mode citations, see whether the same URLs stopped appearing. Given the eligibility rule above, they should move together.

If the dates do not line up, look elsewhere first: a site change, a tracking break, or seasonality.

Step 5: If the templated pages fell

Google's spam updates page says "making changes may help a site improve if our automated systems learn over a period of months that the site complies."

Months, not days. Plan for that.

My view: cut harder than feels comfortable.

A templated page that only swaps a place name or a competitor name into the same paragraphs is the pattern the policy lists. Rewriting it with a different AI prompt does not change the pattern. Either the page gets something only you can say, like real prices, real photos, a real answer to a question your customers ask, or it goes.

The usual mistake is defending a large set of thin pages to protect a handful of good ones. Delete or merge the thin ones, and the good ones stop sharing their neighbourhood.

Also check the policy line on "creating multiple sites with the intent of hiding the scaled nature of the content." Some programmes spread pages across microsites. If yours does, that is a conversation to have with your vendor this week.

Step 6: If nothing moved

Good. Do not read that as approval.

The rollout may not be finished. And a page that survives a spam update can still be invisible in AI answers for ordinary reasons: weak authority, no mentions elsewhere, a competitor with a better answer.

Use the quiet to audit. Ask every page one question, taken from the policy: does it add value for users, or does it exist to rank?

For the publishing side of this, Siddharth's notes on publishing a hundred posts in two weeks cover what volume costs when every page still has to be true.

The short version

• The September 2026 spam update started on 24 September and is expected to run about two weeks.
• A page must be eligible for ordinary Search to be an AI Overviews or AI Mode link. A spam hit removes both.
• Google's scaled content abuse policy targets pages without value, not AI-written pages.
• Google has not said what this update targets. Do not let anyone tell you otherwise.
• Match your drop to the dates before blaming the update.
• If templated pages fell, cut and merge. Recovery takes months.
• If nothing moved, audit anyway.

NovaTechRay works on getting businesses named and cited in AI answers, page by page, at novatechray.com.