Google Confirms LLMS.txt Files Have Zero SEO Impact

Google put a clear stop sign on one of the more persistent SEO myths floating around in 2026. According to Daily Search Forum Recap: June 16, 2026, Google updated its official guide for optimizing for generative AI search to confirm two things: LLMS.txt files have no effect on search rankings, and AI files and Markdown files are not used in Google Search at all. If you've been spending time implementing or obsessing over LLMS.txt as part of your technical SEO workflow, this one's worth reading carefully.

The clarification came on the same day Google's *Off The Record* podcast — featuring John Mueller and Martin Splitt — tackled the Markdown vs. HTML debate directly. Their conclusion was plain: HTML is the standard for SEO. Markdown gives you no ranking benefit. Full stop.

What Google Actually Said

Google's help document update addressed LLMS.txt head-on. The guidance makes clear that these files won't help your rankings and won't hurt them either — Google simply ignores them. It's not a gray area or a "we're still evaluating" situation. Google's position is that HTML is the web standard for search, and anything outside of that lane doesn't factor into how your pages rank.

On the podcast side, Mueller and Splitt reinforced the same point from a different angle. Both acknowledged that Markdown has legitimate use cases in other contexts — documentation, developer workflows, internal tooling — but when it comes to SEO and Google Search specifically, HTML is what the crawler needs. Markdown doesn't give you any additional signal.

Also surfacing on June 16: John Mueller clarified that building a dedicated US subfolder structure for a multinational site carries no practical SEO advantage. It might make your analytics cleaner or your internal workflows easier to manage, but from a pure ranking standpoint, Google doesn't reward the folder structure itself.

And separately, researcher Ahrefs published a study finding that 97% of LLMS.txt files are never read — based on an analysis of 137,000 sites. That data aligns almost perfectly with what Google confirmed in its documentation update.

Why This Matters for Your SEO Strategy

Here's the honest reality: a lot of time has been spent in SEO communities this year debating whether LLMS.txt was the next big technical lever. Conferences have had sessions on it. Threads have gone long. And some agencies have added it to their technical checklists without much evidence it was doing anything.

Google just answered the question. I've seen this pattern play out before — a new file format or meta-layer gets traction because it *sounds* like it should matter for AI visibility. The logic seems reasonable on its face. But when Google's own engineers say they ignore it, the conversation is over. Your crawl budget, your HTML structure, your internal linking — those still move the needle. An obscure text file sitting in your root directory does not.

The broader signal here is worth paying attention to. Google is drawing a clean line between what helps in *Google Search* and what might matter in the broader AI agent and LLM ecosystem. LLMS.txt may have some future relevance for non-Google AI crawlers — the concept exists for a reason — but conflating that with Google Search rankings has been the mistake. These are different systems.

The US subfolder news follows a similar pattern. It's a reminder that structural decisions which look meaningful on paper often have no measurable SEO impact. Spend that energy on things Google actually processes: page content, links, structured data, and render path.

What to Do Now

1. Stop treating LLMS.txt as an SEO task. If you have one, it's not hurting you — but it's not helping you either. Remove it from your technical SEO checklist and reclaim that time for something that actually moves rankings.
2. Audit your HTML quality, not your file formats. Google confirmed HTML is the standard. That means clean, crawlable, well-structured markup matters far more than any supplementary file. Run a technical crawl and prioritize render issues, indexation gaps, and internal linking — the fundamentals still win.
3. Don't restructure US site folders for SEO reasons alone. If you're managing a multinational site and considering a folder-based US structure, make that call based on analytics workflow or CMS requirements — not rankings. Mueller was clear: there's no practical SEO difference.
4. Keep AI visibility strategy separate from traditional SEO. Getting cited in AI Overviews, ChatGPT, or Perplexity is a real goal worth pursuing — but the tactics are different from Google Search ranking. Strong E-E-A-T signals, clear entity definition, and being cited by authoritative sources matter more in that context than file-based signals.
5. Watch how Google defines its AI search guidance going forward. The help document update shows Google is actively clarifying what does and doesn't count in its generative AI search context. Bookmark Google Search Central and check it when new AI-adjacent SEO tactics start circulating.

Background and Context

LLMS.txt emerged as a concept from the AI and developer community as a way to tell large language models how to interpret or use a website's content — similar in spirit to how robots.txt communicates with crawlers. The idea had intuitive appeal, especially as AI search became a louder conversation in 2025 and early 2026.

But there's a big difference between a file that's *theoretically readable* by an AI system and one that actually influences rankings in Google Search. Google's documentation update draws that line explicitly. It also fits a broader pattern: Google has consistently resisted assigning ranking value to signals it can't reliably verify or that don't serve searchers. LLMS.txt, by its nature, is self-reported. That alone should have tempered expectations.

The Markdown vs. HTML debate has also been simmering for a while, particularly as more teams experiment with AI-assisted content workflows that natively produce Markdown. Mueller and Splitt's podcast timing was deliberate — Google clearly wanted to address both issues in the same news cycle. The answer is the same in both cases: build for HTML. Everything else is secondary.

Also mentioned in the June 16 recap: Google's image search has been mixing sponsored ads within organic results since October 2023, under what Google calls dynamic ad placement. This practice is now getting wider public attention after notable commentary on X. It's not a new policy, but it's a useful reminder that the line between paid and organic continues to blur across Google's properties.

If you're tracking your AI search visibility alongside traditional rankings, Aergos has a free tool to show you where you stand across AI platforms — worth a look as you separate your Google Search strategy from your broader AI visibility work.