UDX Original Research · July 2026
3,709 Listed Japanese Companies: How Ready Is Japan for AI Search?
LLMO (Large Language Model Optimization) is the set of technical and content practices that determine whether ChatGPT, Claude, and Perplexity cite or recommend a company when someone asks about its category. UDX surveyed every domain-resolved listed company in Japan (3,709 companies across the Prime, Standard, and Growth markets) as of July 20, 2026. The headline finding: only 4.5% (167 companies) have adopted llms.txt, the emerging standard for making a site legible to AI crawlers.
| Companies surveyed | 3,709 |
| llms.txt adoption | 4.5% (167 companies) |
| Structured data (schema.org) on homepage | 29.0% |
| hreflang (multilingual signal) implemented | 12.5% |
| Companies explicitly blocking major AI crawlers | 46 |
Source: UDX, “Japan Listed Companies AI-Search Readiness (LLMO Readiness) Survey,” as of July 20, 2026, n=3,709. Updated periodically at the UDX Digital Intelligence Hub (Japanese).
Why this matters for companies entering Japan
When a Japanese buyer asks an AI assistant “which companies in [category] are reliable,” the assistant answers from whatever it can read cleanly — structured data, llms.txt summaries, and third-party mentions. With 95.5% of listed companies lacking llms.txt, the category is wide open: a foreign entrant with a well-structured Japanese-language presence can outrank domestic incumbents in AI answers, not just traditional search.
Readiness by industry
Among industries with 20+ surveyed companies, adoption is concentrated in a handful of sectors and near-zero in most others.
| Industry | n | llms.txt |
|---|---|---|
| Securities & commodity futures | 37 | 10.8% |
| Information & communications | 602 | 10.0% |
| Services | 534 | 8.4% |
| Retail | 322 | 3.4% |
| Land transportation, precision instruments, textiles, electric & gas, non-ferrous metals, banking, warehousing | 20–79 each | 0.0% |
Methodology
- Universe: All domain-resolved companies listed on the Prime, Standard, and Growth markets of the Japan Exchange Group (JPX), n=3,709.
- Domain source of truth: Kabutan “company site” field, cross-checked against Wikidata where mismatched.
- Signals measured: robots.txt permissions for major AI crawlers (GPTBot, ClaudeBot, PerplexityBot, etc.), presence and content of llms.txt, homepage schema.org JSON-LD, and hreflang implementation.
- Collection: Machine-fetched from each company’s official domain; llms.txt adoption requires a verified, non-empty file (not just a 200 response).
- Survey date: July 20, 2026. Re-surveyed on a rolling weekly basis at the UDX Digital Intelligence Hub.
Corrections
This is a factual, machine-collected technical survey of publicly available site configuration — it is not an evaluation of company quality. Domains and site configurations change; if a listing looks inaccurate, contact info@udx.co.jp and we will verify and correct it promptly.
Want to know where your brand stands in Japanese AI search?
The methodology behind this study is the same one we use in the Japanese AI Visibility Audit and the AI Visibility Snapshot.
See the USD 299 Snapshot →