How much AI is Japan's public sector actually buying? We pulled every procurement notice with “AI” in the title from the Public Procurement Information Portal (官公需情報ポータルサイト, kkj.go.jp), run by Japan's Small and Medium Enterprise Agency, removed unrelated hits, and counted the rest by buyer, prefecture, month and purpose. This page is the English edition of our Japanese report; the figures are generated from the same data.
Key figures
Japan's central government, local governments, universities and research institutes published 471 AI-related procurement notices from 2026-02-01 to 2026-10-10 (AI PICKS analysis of the government's public procurement portal).#
According to AI PICKS (471 AI-related procurement notices on Japan's public procurement portal, published 2026-02-01 to 2026-10-10), 52 notices (11%) name generative AI in the title.#
According to AI PICKS (471 AI-related procurement notices on Japan's public procurement portal, published 2026-02-01 to 2026-10-10), 48% (225 notices) are for introducing AI systems or services.#
According to AI PICKS (471 AI-related procurement notices on Japan's public procurement portal, published 2026-02-01 to 2026-10-10), 26 notices (6%) are for AI training of public-sector staff.#
According to AI PICKS (471 AI-related procurement notices on Japan's public procurement portal, published 2026-02-01 to 2026-10-10), the most common type of buyer is local governments, with 32% of notices (152).#
According to AI PICKS (471 AI-related procurement notices on Japan's public procurement portal, published 2026-02-01 to 2026-10-10), Tokyo has the most notices: 199 (42%).#
Summary
What the notices are for
| AI systems and services | AI software, chatbots, analytics and other systems225 (48%) |
|---|---|
| AI training | Generative AI and AI literacy training for staff26 (6%) |
| Other AI-related work | Studies, pilots, data preparation and similar220 (47%) |
Categories are assigned by our editors from the notice title only.
Notices by month
The last month is partial: it covers notices up to the data date (2026-10-10).
Who is buying
Buyer types are assigned mechanically from the organization name (for example, names containing 大学, university, or 省, ministry). This is not a legal classification of the organizations.
Top 10 prefectures
Tokyo ranks high largely because central ministries and national agencies are based there. 10 notices do not state a prefecture and are left out of this chart.
Method and limitations
- Source: tender and procurement notices published through the search API of the Public Procurement Information Portal (kkj.go.jp), run by Japan's Small and Medium Enterprise Agency.
- We extract notices whose title contains “AI” in either Latin or full-width Japanese characters (AI / ＡＩ). Our editors then remove unrelated hits, such as titles where the letters are part of another word or code.
- “Generative AI”, “AI training” and “AI systems and services” are editorial classifications based on the title. Duplicate notices are removed by the portal's notice key.
- We count notices, not money. Many notices do not disclose the planned price or the winning bid at the time of publication.
- The data is refreshed regularly, so the figures on this page change over time. Cite them with the data date shown above.
How to cite this data
You may quote the figures and charts on this page in articles, papers, reports and social posts under CC BY 4.0. Please credit AI PICKS and link to this page.
Source: AI PICKS, "AI Procurement by Japan's Public Sector 2026", data as of 2026-10-10. https://aipicks.jp/en/research/gov-ai
Questions about the method or the underlying data, and requests for custom cuts, are welcome through our contact form (English is fine).