Piperic
AI Access Report
AI Report
🌍 The AI Access Report · 2026-07-13

🇨🇳 China's web is the world's most silent on AI — 98% have declared nothing

Of 275,688 live Chinese websites, only 2,984 (1.08%) block an AI crawler — 2nd most open of 43 countries. And 98.4% neither block nor declare a policy, the most passive web in the report.

Analysis of 41,818,678 live, content-validated websites worldwide · snapshot 2026-07-13 · by Piperic

Share:XLinkedIn

98.9%open to AI crawlers

Measured purely on the robots.txt signals this report tracks, the Chinese web is strikingly open. Of 275,688 live .cn sites, only 1.08% block a Western AI crawler — 2nd most open of 43 countries, behind only Japan. (Note: China's own regulatory and network controls sit outside these public web signals; this measures declared crawler policy, not access in practice.)

The defining Chinese number is silence: 98.41% of sites are 'open by silence', the highest in the entire report. Conscious opt-in (0.57%) and llms.txt adoption (0.58%) are the lowest of any country measured — the emerging Western conventions for signalling AI policy have essentially no footprint here. Notably, in our sample zero Chinese news sites block via these tokens.

Where the little blocking exists, it's imprecise: 13.2% of GPTBot-blockers also lost OAI-SearchBot, one of the higher paradox rates — but the base is tiny. China's profile in this dataset is an internet almost entirely untouched by the Western robots.txt-and-llms.txt conversation.

The three numbers that matter

98.99%
AI Training Exposure
share of sites NOT blocking any AI training bot (GPTBot, ClaudeBot, CCBot…) — their content can be used for model training.
99.85%
AI Search Visibility
share of sites reachable by AI search crawlers (OAI-SearchBot, PerplexityBot…) — they can appear in AI answers.
1.75%
AI Policy Maturity
share of sites with ANY conscious, machine-readable AI signal (llms.txt, ai.txt, robots AI rules, ToS clause). Low = the web hasn't decided yet.

Where China stands in the world

Share of sites blocking at least one AI crawler — lower = more open. China ranks #2 of 44 countries analysed.

#1 🇯🇵 Japan1.0%#2 🇨🇳 China1.08%#3 🇩🇪 Germany1.58%#4 🇦🇹 Austria1.82%#5 🇷🇺 Russia1.85%#6 🇫🇷 France2.14%#7 🇧🇪 Belgium2.23%#8 🇪🇪 Estonia2.66%

Neighbours shown for context. Full ranking in the data download.

💬 2nd of 43 on declared crawler openness — behind only Japan. A measure of robots.txt signals, not of China's broader network controls.

The four AI postures of China's web

Every website falls into one of four groups, based on what it blocks and what it declares:

98.41%
Open by silence
doesn't block AI and declares nothing — no decision has been made. This is the real story: the vast majority.
0.57%
Consciously open
doesn't block AI and publishes an explicit policy file (llms.txt / ai.txt) — a deliberate yes.
0.87%
Controlled access
blocks training bots but stays visible to AI search — the technically mature strategy.
0.15%
AI-dark
blocks AI search crawlers too — protected, but disappearing from AI answers.
Open by silence: 98.41%Consciously open: 0.57%Controlled access: 0.87%AI-dark: 0.15%98.41%Open by silence
💬 98.41% open by silence — the report's highest. Conscious AI signalling (0.57%) and llms.txt (0.58%) are the lowest anywhere.

The ChatGPT-invisibility paradox

1,937GPTBot ⛔
256+ OAI-SearchBot ⛔
13.2%ChatGPT Search ✕

1,937 China websites block GPTBot to keep their content out of AI training. But 256 of them — 13.2% — ALSO block OAI-SearchBot, most likely by accident. Those sites have removed themselves from ChatGPT Search results entirely: protected from training, but invisible where their customers now search.

GPTBot (training) and OAI-SearchBot (ChatGPT Search) are independent robots.txt tokens — blocking one does not block the other.

💬 13.2% accidental self-exclusion among a very small set of blockers — high rate, tiny base.

Which AI bots does China block?

Share of China sites blocking each crawler in robots.txt.

Bytespider · TikTok0.87%GPTBot0.7%ClaudeBot0.66%Amazonbot0.66%CCBot0.65%Google-Extended0.63%Meta-ExternalAgent0.6%Applebot-Extended0.53%ChatGPT-User0.11%

Bytespider is ByteDance's (TikTok) crawler — it ignores robots.txt more often than the majors, yet it is among the most-blocked.

Industry by industry: who blocks, who doesn't

AI-blocking and llms.txt adoption across China's biggest website categories.

SectorSitesBlocks AIllms.txt
📰 News & media2670.0%0.37%
entertainment52,5960.41%0.0%
sensitive topics52,1270.12%0.0%
business and finance50,6351.01%0.84%
technology and computing30,5481.8%1.19%
personal finance12,9140.99%0.13%
sports10,4513.84%1.31%
home and garden6,1820.58%0.94%
video gaming5,5301.59%0.13%
real estate5,4340.28%0.13%
automotive4,6420.58%1.18%
education3,9351.07%0.79%

China news sites block AI 0.0× more often than the national average (0.0% vs 1.08%) — publishers are the most defensive segment of the web.

💬 Uniformly low blocking; in our sample no Chinese news sites use these AI-blocking tokens at all. The Western signalling conventions are effectively absent.

The e-commerce exposure

4,470 China online shops are open to AI crawlers (97.3% of 4,596). Their product content can already be read, compared and recommended by AI assistants — whether they know it or not.

Platform matters

Share of sites blocking at least one AI bot, by platform:

3.97%WordPress0.88%Shopify1.08%country average

The trend starts here

This is the baseline edition (2026-07-13). Piperic re-measures continuously; the next edition will show how fast China's web is closing — or opening.

💬 China begins as the most silent web in the report on Western AI signals. The next edition tracks whether any of these conventions take hold.

Check your own website

Free, no signup — see your site the way AI crawlers do:

For journalists — press kit

“China is the second most open web in the world by our measure and by far the most silent: 98 percent of sites have made no AI declaration at all. Formal AI policies like llms.txt are almost non-existent — the Western debate over robots.txt simply hasn't reached this web,”
— said Attila Rácz-Akácosi, founder of Piperic.

How to cite: “According to Piperic's AI Access Report (2026-07-13), based on 275,688 live China websites…” — link to this page.

Download the country dataset (JSON) Global rankings dataset

Methodology

Sample: 275,688 live, content-validated websites (China, ccTLD-based assignment), out of 41,818,678 live domains analysed worldwide. “Live” = HTTP 200 and not parked. “Blocks AI” = the site's robots.txt disallows at least one of 14 known AI crawlers. llms.txt/ai.txt = the file exists at the site root. Snapshot: 2026-07-13 (frozen — numbers do not move with the live crawl). Full-web note: these figures cover the ENTIRE live web including the long tail; industry headlines like “25% of top sites block GPTBot” measure only large publishers — both are true, they measure different things. Country assignment via ccTLD under-counts .com/.org sites; treat US figures as indicative.

Press contact: press@piperic.com · Data: Piperic Business Intelligence

Every number on this page is reproducible from the public dataset above.

Country reports

🌍 Worldwide🇩🇪 Germany🇷🇺 Russia🇬🇧 United Kingdom🇳🇱 Netherlands🇫🇷 France🇧🇷 Brazil🇨🇭 Switzerland🇯🇵 Japan🇵🇱 Poland🇦🇺 Australia🇸🇪 Sweden🇨🇳 China🇮🇹 Italy🇨🇦 Canada🇨🇿 Czechia🇧🇪 Belgium🇦🇹 Austria🇪🇸 Spain🇮🇳 India🇨🇴 Colombia🇸🇰 Slovakia🇩🇰 Denmark🇭🇺 Hungary🇲🇽 Mexico🇳🇴 Norway🇰🇷 South Korea🇦🇷 Argentina🇳🇿 New Zealand🇫🇮 Finland🇬🇷 Greece🇺🇦 Ukraine🇨🇱 Chile🇿🇦 South Africa🇷🇴 Romania🇻🇳 Vietnam🇹🇷 Turkey🇪🇪 Estonia🇵🇹 Portugal🇹🇼 Taiwan🇮🇪 Ireland🇮🇱 Israel🇮🇩 Indonesia🇱🇹 Lithuania