Acquiring into Japanese Academic Publishing

2Why Japan: Five Structural Reasons

1. A rare market where "content and monetization are separated"

The structure of Japanese academic publishing differs from that of other countries. In Europe and the United States, learned societies typically hand their journals to commercial publishers and share the revenue, and the publishers themselves are content owners; in China, the right to sponsor a journal and the publishing license are tightly bound to state-owned entities. In Japan, by contrast: learned societies own journal copyright and editorial control, the state provides a free platform, and international publishers take global sales and indexing.

This "three-layer separation" creates a distinctive deal space. Learned societies do not sell themselves, but they have full discretion over their contract publishing relationships; J-STAGE is a national public good, but its service procurement and technical cooperation are open; international publishers control sales, but their contract publishing agreements have terms and renewal windows. In other words, the Japanese market is not one where "there is nothing to buy," but one where what is buyable is not on the conventional M&A checklist.

The platform side presents a different picture: J-STAGE expanded from about 3,000 titles to 4,396, and participating organizations grew from 1,687 to 2,728—the state has done infrastructure very well, but it has not solved the question of who will sell the content to the world.

2. Structural contraction in the domestic market is creating sellers

The Japanese publishing market has declined for four consecutive years, and in 2025 print media fell below 1 trillion yen for the first time (964.7 billion yen, down 4.1% year on year). The professional books and journals segment in which academic publishing sits was hit even earlier: core subsidies for university presses fell from nearly 700 million yen in 2006 to 268 million yen in 2024; the 24 university presses publish only 535 titles a year, and among them Chuo University Press, the only one that discloses its profit and loss, derives 87.4% of revenue from contract publishing and has run a deficit for two consecutive years.

The Physical Society of Japan's PTEP fell into a structural deficit after JSPS funding ended, and APC waivers must be reviewed case by case; learned-society membership is shrinking and competition from commercial publishers is intensifying; mid-sized science and engineering publishers broadly face generational handover. These are not short-term fluctuations but a sustainability crisis—one that has turned institutions that would "never sell" into parties "willing to talk." For a buyer, this is the most favorable condition in valuation negotiations.

The position of academic publishing is further compounded by a "quantity-quality divergence": Japan's share of papers still ranks fifth globally, but its share of highly cited papers has fallen out of the top ten. The declining international influence of journals further weakens learned societies' willingness and ability to publish on their own, making "letting someone else publish the journal" an easier option to accept.

This contraction is equally clear in DOI registration data: annual DOI registrations by Japanese institutions peaked in 2007 and fell from about 59,700 to 45,400 in 2024.

3. Money is flowing out: Japan has already proven the value of this market

Japanese universities pay 43.03 billion yen a year for foreign journal subscriptions and only 3.45 billion yen for domestic journal subscriptions; Japanese authors paid 12.94 billion yen in APCs in 2023, roughly 18 times the national journal publishing subsidy (711.7 million yen).

The implication of these figures is clear: the Japanese academic community is willing to pay for "better publishing," and most of that money goes to foreign organizations. If a Chinese publisher can improve the editorial quality, international visibility, and open access of Japanese domestic journals, it will face a market that has already been validated and that has stable willingness to pay. This is not new demand that must be cultivated but existing demand that requires replacing the supplier.

4. A policy window is opening: mandatory immediate OA in 2024 and mandatory full-text XML in 2029

In February 2024, Japan's Cabinet Office adopted the "Basic Policy on Immediate Open Access to Academic Papers," requiring immediate green OA for outputs of designated competitive funding, effective from FY2025; in 2026 JST announced that by FY2029 current J-STAGE journals must submit full-text JATS XML, with PDF relegated to a secondary role.

The combination of these two policies creates a clear technical imperative: of Japan's 3,705 current journals, 93.4% currently lack full-text XML capability. Whoever can provide low-cost, high-quality Japanese JATS XML conversion and production services will control the entry point to Japanese academic publishing over the next three years. For a Chinese publisher, this is a faster, cheaper, and harder-to-replace entry point than buying journals.

5. The yen cycle and valuation anchors: this is a buyer's market

China Science Publishing's 2019 acquisition of France's EDP Sciences (75 journals) had an enterprise value of only EUR 12 million, officially described as "the first time a Chinese science and technology publisher has completed an acquisition of a Western publishing institution in a truly meaningful sense." By comparison, Wiley paid US$298 million for Hindawi, and KKR paid US$1.62 billion for Simon & Schuster.

The valuation anchor for academic journal assets is more than an order of magnitude lower than that for trade publishing. Japanese targets are generally smaller and have weaker international influence, so their reasonable valuations should fall in an even lower range. Combined with the yen's long-term weakness, Chinese buyers enjoy a clear cost advantage.

3Evaluation Framework: A Six-Dimension Scoring Model

This report does not rely on the intuition that "a big name means a good buy"; instead it builds a reproducible six-dimension scoring model. Each dimension is scored from 0 to 5, weighted, converted to a 100-point scale, and a risk score is deducted separately.

It should be emphasized that D6 is a risk dimension and is subtracted in the composite score. We deliberately place "existing high-impact English-language journals" in the middle-to-lower ranks of their categories—because their contract publishing agreements are mostly locked up long term by OUP, Springer Nature, Wiley, and IOP, making them the most tempting on the surface and the hardest to obtain in practice. This is where the model differs most from intuition.

The radar chart makes the differences among the leading targets clearer: Kokusai Bunken-sha scores high on both transferability and corpus value; Ichushi earns a perfect score on corpus value but its transferability is constrained by its legal form; and the society journal portfolios are stronger on strategic leverage and weaker on transferability.

4The Target Pool at a Glance: Five-Layer Structure and Acquirability

Once Japanese academic publishing is broken into five layers, the distribution of "acquirability" becomes clear. The content-owning layer (learned societies and universities) is almost entirely non-acquirable; the services-and-platform layer is partly acquirable; and the monetization layer (international publishers) is not in Japan at all. This determines the feasible set for this transaction.

One boundary must be stated clearly: there is not a single "internationally scaled" publisher in Japanese academic publishing. The Japanese publishing organization with the highest OpenAlex coverage is JSME (about 316,000 records, after excluding sources with anomalous metadata), followed by the Japan Neurosurgical Society, the Japanese Circulation Society, the Pharmaceutical Society of Japan, and the Physical Society of Japan. Nowhere in Japan is there an entity approaching the scale of Elsevier, Springer Nature, or Wiley. This means acquiring Japanese targets cannot buy a global platform in one step, only a set of capability components—which is precisely why this report recommends a "portfolio transaction."

Another noteworthy phenomenon: the most internationalized Japanese journals are almost all in the hands of international publishers. The chart below plots "document volume × two-year average citations," and the high-impact journals in the upper right—Gastric Cancer, Cancer Science, Polymer Journal, and Hypertension Research—are overwhelmingly published under contract by Springer Nature, Wiley, and others.

5Category A: Society Journal Portfolios—You Cannot Buy the Entity, but You Can Buy Publishing Rights and Corpora

The copyright of Japanese academic journals belongs almost entirely to learned societies. Learned societies are public-interest corporations or incorporated associations and cannot be acquired like companies, but substantive control or use rights can be obtained in three ways: taking over contract publishing (the most direct), signing data and translation licenses (the lightest), and establishing a joint-venture publishing company (the heaviest but most stable).

A1 Japan Society of Mechanical Engineers (JSME): the largest existing stock of engineering corpora

JSME is the Japanese institution with the largest DOI registration volume—Crossref records 338,673 DOIs, more than four times the runner-up. It publishes Japanese-language proceedings, English-language journals, and a large number of conference proceedings, with about 316,000 records covered by OpenAlex.

Its value lies not in its impact factor (JSME International Journal has only average international visibility) but in the irreplaceability of Japanese-language engineering application literature: a great deal of process detail in machinery, materials, robotics, and precision machining exists only in Japanese, and China's engineering discipline has the largest absolute shortfall of any field (in 2023 Chinese authors produced 207,000 papers while domestic journals carried 12,700, a gap ratio of 16.4 times).

• Transaction form: contract publishing succession (English-language journals) + data licensing (Japanese-language proceedings)
• Valuation anchor: priced on corpus scale rather than journal revenue
• Main obstacle: JSME is large and financially relatively sound, giving it the lowest willingness to negotiate

A2 Physical Society of Japan: a rare window of willingness to talk

JPSJ is Japan's physics flagship (about 45,000 records, h-index 239), and PTEP is an open-access theoretical physics journal. After JSPS funding ended, PTEP fell into a structural deficit and APC waivers must be reviewed case by case—the clearest "financial distress signal" in the entire pool. For a buyer, this means the society has a genuine incentive to seek a new publishing partner.

But physics happens to be a discipline China is relatively less short of (a gap ratio of only 8.9 times, among the lowest of all disciplines). The model therefore ranks the physics target mid-pack: it is a good "relationship entry point" and "credibility vehicle," but not the most urgently needed discipline-fill target.

A3 Japan Society of Applied Physics (JJAP/APEX): closely aligned with industrial weak spots

JJAP and APEX are currently published under contract by IOP Publishing, which has had a Tokyo office since 2002. Applied physics, semiconductors, and materials are precisely the fields where China's "chokepoint" problems and weak domestic journals coexist. The contract with IOP is a ready-made negotiation interface—contract expiry or renewal is the window for a new tender.

A4 Chemical Society of Japan: the most valuable English-language asset, but already spoken for

BCSJ (about 38,700 records, two-year average citations 3.65, h-index 169) and Chemistry Letters (about 30,300 records) are in the first tier of Japanese domestic journals. But as of January 2024, the Chemical Society of Japan has transferred the publishing operations of these two journals to Oxford University Press.

This case is highly instructive: it proves that "society-international publisher" contract relationships do change, and that change windows genuinely exist. But it also shows that for a Chinese buyer to take over such an asset it must offer terms better than OUP's—a matter not just of price but of brand and trust. This is why the model gives the chemistry target a D3 transferability score of only 2.6.

A5 Pharmaceutical sciences and clinical medicine: the content assets with the highest discipline-fill value

The Pharmaceutical Society of Japan (CPB 29,200 records, Yakugaku Zasshi 25,400 records, and BPB 11,400 records, with BPB now in DOAJ) and the journal portfolios of the various clinical medicine societies are the content assets with the highest discipline-fill value:

• China's medicine shortfall is the largest in absolute terms (121,700 papers produced in 2023, 4,588 carried domestically, a gap ratio of 26.5 times)
• The gap ratio for pharmacology and pharmaceutics is 21.8 times, and China's domestic capacity rate is extremely low
• Japan's clinical medicine output (45,000 papers) far exceeds the capacity of China's domestic medical journals, and Japanese-language clinical experience literature is almost irreplaceable in its unique value for medical AI

• Transaction form: data licensing for Japanese-language journals + contract publishing succession for English-language journals
• Valuation anchor: corpus value + APC revenue sharing rather than subscription revenue
• Main obstacle: a large number of societies with dispersed rights, requiring negotiations one by one and entailing high due diligence costs

6Category B: Publishing Services and Production Companies—the Most Tradable Assets in the Pool

If only one transaction can be made, this report recommends directing it at this layer. The reason: they do not hold journal copyright and are therefore not bound by society ownership; they are corporatized and therefore acquirable; and what they possess are processes and relationships, which are extremely hard to replicate.

B1 Kokusai Bunken-sha (Bunken): the "operating system" of Japanese academic publishing

Since 1995, Kokusai Bunken-sha has taken on society-secretariat outsourcing and has supported J-STAGE since 2002, covering submission systems, peer-review workflows, typesetting and production, full-text XML conversion, and DOI and PMC registration—essentially the entire workflow of a journal from submission to launch. It serves a large number of learned societies, which amounts to controlling the operational pipeline and customer network of Japanese academic publishing.

Why this ranks first in the pool (composite score 72.9):

• D4 AI corpus value 4.6: it already processes full-text and structured data, so corpus acquisition does not require building a new pipeline
• D3 Transferability 4.6: a privately held stock company with tradable equity, and as a services company it is not subject to publishing license restrictions
• D1 Strategic leverage 3.6: buying it is equivalent to gaining an entry point for taking on more society journal business
• D5 Valuation feasibility 3.8: labor-intensive with limited gross margins, so the valuation will not be too high

Key risk: customers (learned societies) are extremely sensitive to changes in control. The transaction must therefore be designed to "retain the existing management, the existing brand, and the existing service commitments," and the announcement must explicitly state that "service terms for Japanese learned societies will not change." This is not rhetoric but a precondition for the deal to succeed.

B2 Nakanishi Printing: an alternative in the same lane and a price benchmark

Founded in 1865, Nakanishi Printing likewise handles society editorial office operations and academic typesetting, and has partnered with Oxford University Press since 1999. Together with Kokusai Bunken-sha it forms a comparable in the same lane: smaller in scale, longer in history, and more privately held.

Generational handover is the typical sale trigger for family-run printing businesses of this kind. For a buyer, Nakanishi Printing's value lies in providing a price reference and an alternative, and it may transact at a lower valuation. The model gives it a composite score of 66.8 and a D5 valuation feasibility of 4.2 (higher than Bunken), reflecting the fact that it is "cheaper but somewhat narrower in capability."

B3 Center for Academic Publications Japan (CAPJ): not acquirable, but a joint venture is possible

CAPJ was established by the University of Tokyo Press in 1971 as a 20th-anniversary project and is a nonprofit incorporated foundation. It cannot be acquired, but it can serve as a joint-venture partner or business successor, and its nonprofit status can serve as a buffer for public opinion and compliance.

Recommended use: when a controlling transaction is sensitive, use CAPJ as the nominal entity to take on sensitive business, with the Chinese side providing technology and capital. This can significantly reduce the narrative risk of "Chinese capital controlling Japanese academic publishing."

7Category C: Secondary Indexes and Data Assets—the Highest AI Corpus Value Density

The distinguishing feature of this category is that its value lies not in full texts but in structured indexes. For AI training, secondary literature with controlled subject terms, abstracts, and classification codes is often more efficient than raw full text—because annotation density is higher and noise is lower.

C1 Japan Medical Abstracts Society (Ichushi-Web / Ichushi): the most undervalued asset

Ichushi is the only systematic secondary index of Japanese clinical and nursing literature, containing tens of millions of Japanese-language medical records and abstracts. Its irreplaceability rests on three points:

1. Language exclusivity: Japanese-language clinical literature has no equivalent substitute source anywhere in the world
2. Structured conventions: medical subject terms and classification codes make it naturally suited for use as training and retrieval corpora
3. Discipline match: medicine is the discipline with China's largest absolute shortfall, and the gap ratios for nursing and health professions are as high as 36 times and 27 times

The model gives it D4 = 5.0 (a perfect score): among all 25 targets, it delivers the highest AI corpus value density per unit of investment.

Main obstacle: the Japan Medical Abstracts Society is a nonprofit corporation and cannot be acquired through equity. The realistic path is data licensing, joint development, or joint-venture operation. In addition, the Japanese side is highly sensitive to cross-border flows of medical data—"data stays in Japan, the model is trained domestically, and only model capability is delivered" may be the only feasible structure.

C2 Japan Association for International Chemical Information (JAICI)

JAICI is the Japanese agent for CAS and STN, providing chemical structure, literature, and patent search services. Chemical and patent data has direct value for materials and pharmaceutical AI, but JAICI does not hold copyright in original content and is easily substituted (CAS and Reaxys internationally already cover similar needs). The model therefore gives it a composite score of only 54.7, placing it second in Category C.

C3 JST Current Bibliography on Science and Technology / J-GLOBAL: explicitly non-tradable

JST is a national research and development agency, and its Current Bibliography and J-GLOBAL are national infrastructure and cannot be acquired. This report includes it in the target pool to draw a clear "non-tradable boundary": the state-dominated portion of Japan's academic information system (J-STAGE, JaLC, CiNii, and J-GLOBAL) can only be partnered with, not acquired. Any valuation model that counts it as an acquisition target is wrong.

8Category D: Commercial Publishers—Genuinely Acquirable, but They Require a Bundling Logic

Commercial publishers are the only category that can be handled under conventional M&A logic: corporatized, with clear equity, financial statements, and negotiability. But individually they are all small, and their content is mainly textbooks and professional books. The core recommendation of this report is therefore "bundling" rather than "picking one star publisher."

D1 Igaku-Shoin: Japan's largest commercialization vehicle for medical content

Igaku-Shoin has revenue of about 10.69 billion yen (FY2021) and owns roughly 30 clinical monthlies plus a large number of medical and nursing textbooks, making it Japan's largest medical publisher. Its value has three layers:

• Content layer: Japanese-language clinical textbooks and journals, directly matching China's enormous shortfall in medicine and nursing
• Channel layer: long-standing subscription relationships with Japanese clinicians and nursing staff
• Brand layer: a recognized authority in Japanese medical education

The model gives it a composite score of 68.2 and a D2 discipline fit of 4.6: it is the most complete target, combining content, channels, and brand in one. Generational handover at Japanese family businesses is the typical sale trigger for companies of this kind.

D2 Ohmsha: both a publisher and an interface to society journals

Ohmsha's contract printing business covers the journals of IEEJ, IEICE, and the Information Processing Society of Japan (IPSJ), while it also publishes a large number of science and engineering textbooks. Its strategic significance lies in its dual identity: as a publisher it is acquirable; as a contract printer for society journals it is a springboard for accessing Category A society assets.

The model gives it a composite score of 66.9, including a D3 transferability of 4.0—against the backdrop of Category A assets being broadly non-tradable, "buying the interface" is a pragmatic alternative.

D3 A bundle of mid-sized science, engineering, and natural science publishers

Mid-sized publishers such as Asakura Shoten, Shokabo, Kyoritsu Shuppan, Tokyo Kagaku Dojin, and Uchida Rokakuho all face the twin pressures of market contraction and generational handover. Individually they are all small, but bundled together they can form a complete corpus of Japanese-language STM textbooks and professional books, covering foundational disciplines such as mathematics, physics, chemistry, biology, and engineering.

The model gives it a D5 valuation feasibility of 4.4 (among the highest in the pool), because the transaction prices of such assets are likely to be extremely low. Its drawback, however, is a D1 strategic leverage of only 2.4: fragmented brands, limited synergy, and high due diligence costs.

9Category E: University Presses—Low Political Cost, Limited Economic Return

Japan has 24 university presses (members of the Association of Japanese University Presses), which together published 535 titles in 2025. Their legal forms vary: foundations, associations, stock companies, and even units within universities. Only a very few of them, in corporate form, offer any possibility of an equity transaction.

E1 University of Tokyo Press: the best brand, the hardest path

University of Tokyo Press is the leading brand among Japanese university presses, releasing 123 new titles in 2025 (23% of the association total) and more than 8,300 titles cumulatively, with English-language books distributed through Columbia University Press. The quality of its Japanese studies and humanities content is extremely high.

But it is an incorporated foundation, deeply tied to the University of Tokyo, and there is almost no possibility of acquisition. The model gives it a D3 transferability of only 1.6. The realistic path is co-editions, translation licensing, and digital distribution partnerships.

E2 Keio University Press: the most negotiable in structural terms

Keio University Press released 78 new titles in 2025, second in the association, and is organized as a stock company. Among AJUP members, its ownership structure comes closest to being tradable. Its content spans the humanities and social sciences, economics, law, and medicine.

The model gives it D3 = 2.8 (the highest in Category E). Note, however: even if it is legally tradable, as one of Japan's top private universities Keio would still present substantial public-opinion and governance resistance.

E3 Chuo University Press: the pilot target most likely to be willing to talk

Chuo University Press is very small (21 titles in 2025), but it is the only university press in Japan that discloses complete profit and loss figures: contract publishing accounts for 87.4% of revenue, it has run a deficit for two consecutive years, and it has publicly commented on the contraction in professional book retailing.

For a buyer seeking "low-cost exposure to Japanese publishing operations," it is an ideal small-scale pilot. The model gives it D3 = 3.4 and D5 = 3.6. It is recommended as the first target for a "joint venture plus debt-restructuring-style cooperation," rather than as a source of economic return.

10Category F: Long-Tail Japanese-Language Corpora—the Lowest Unit-Cost Strategic Asset

This is the category this report most wants to emphasize, and the one most easily overlooked from a conventional M&A perspective.

J-STAGE hosts 3,705 current journals and 6 million accumulated documents from 2,728 publishing organizations. Of these, purely Japanese-language journals number about 818 and Japanese-Chinese mixed journals about 868. The individual impact factors of these journals are negligible and international citations are extremely low, making them worthless from the standpoint of "buying a journal brand."

But from an AI corpus perspective, they are the world's largest, most disciplinarily complete, and most exclusive collection of Japanese academic texts, and moreover:

• Extremely low unit cost: the licensing cost per article is far below that of acquiring any publisher
• Extremely high language scarcity: the global supply of high-quality Japanese academic corpora is limited
• Strong disciplinary diversity: full coverage of medicine, engineering, agriculture, the humanities, and the social sciences, precisely compensating for the narrow register of Chinese-language corpora
• Stepwise approach possible: sign a 100,000-article pilot first and expand after validation, keeping risk controllable

The model gives Category F targets D4 = 5.0 and D5 = 4.6, for a composite score of 68.2 and a joint second-place ranking.

But one key constraint must be faced squarely: free does not mean it can be monopolized

More than 80% of J-STAGE content can be read for free, but this does not amount to an open license. Data from 2022 show that of 3,632 J-STAGE journals, only 214 (about 6%) are genuinely open access (with explicit reuse licenses), while 2,943 (about 81%) are merely "free to read." With content that is free to read, you cannot stop competitors from reading it too.

The value of long-tail corpora must therefore be built on processing and exclusivity, not on raw accessibility:

1. Processing barriers: convert PDF/HTML into high-quality JATS XML, perform OCR error correction, and add terminology and entity annotations
2. Exclusivity clauses: secure an exclusivity period or a right of first renewal in the license agreement
3. Structured value-add: organize scattered journals into a searchable, traceable dataset with a citation network

These three things are precisely the capabilities of Category B publishing services companies. Hence this report repeatedly stresses: Category F corpora and the Category B pipeline are two ends of the same business and should be pursued together.

11Discipline-Fill Map: China vs. Japan

The value of an acquisition must land on specific disciplines. Using a consistent methodology, we calculated the "domestic journal capacity rate" for China and Japan alike (domestic journal output ÷ domestic author output, 2023, OpenAlex basis), producing a map that can directly guide target screening.

The most critical finding is that Japan is not strong in every discipline, and the distribution of China's shortfalls is partly mismatched with Japan's strengths. This requires distinguishing three categories of disciplines:

Disciplines with the largest gaps (2023)

One mismatch must be pointed out honestly: the frontier fields that Chinese official assessments identify as needing their "gaps filled"—quantum science and technology, medical informatics, computer science-cybernetics, developmental biology, substance abuse, and biodiversity conservation—are precisely the fields in which Japan is also weak. Buying Japan cannot fill China's most frontier gaps. What Japan can fill is clinical medicine, engineering applications, materials processing, chemistry, and agriculture—fields that are "mature in system but lagging at the frontier." Clarifying this mismatch is more valuable than saying in general terms that "Japanese journals are worth buying."

12Quantifying AI Corpus Value

Market size: an enormous existing stock, with a very low usable share

Assuming an average of 8,000–12,000 tokens per article, 6 million documents correspond to a raw token volume of between 48 billion and 72 billion. Even if only 20% of that is structurally processed, it amounts to high-quality Japanese academic corpora on the order of 10 billion tokens—a considerable increment for multilingual large models.

Another signal favorable to the corpus case: China is already J-STAGE's third-largest overseas readership. In FY2025, China downloaded from J-STAGE 81.9 million times, accounting for 12.0% of all downloads. This means Chinese researchers have genuine demand for Japanese-language content itself, and the downstream application scenarios for corpus building already exist.

But "accessible" is by no means the same as "usable." The comparison in the chart below is critical: Japan's green OA rate is only 3.2%, far below that of the United States, the United Kingdom, Germany, and France; at the same time, only 95 Japanese journals have genuinely obtained open-license certification.

Corpus value assessment: six leading targets

Legal environment: Japan is more permissive than Europe and the United States, but China's requirements are stricter

Article 30-4 of Japan's Copyright Act provides a relatively broad exception for text and data mining, not resting solely on a "non-enjoyment purpose" requirement, but it retains a proviso for "unfairly prejudicing the interests of the rights holder"; at the same time, it does not apply where the use is for "enjoyment purposes." This means the legal risk of conducting model training within Japan is significantly lower than in Europe and the United States.

But the requirements on the Chinese side are stricter: the Interim Measures for the Administration of Generative AI Services require training data to be lawfully sourced and not infringe intellectual property. Therefore, when obtaining corpora from Japan, an explicit authorization permitting cross-border use for model training must be secured in the contract; it is not enough to rely on Japan's TDM exception alone.

One reality must be faced squarely: as of now, there is no publicly disclosed case of a Chinese AI company purchasing Japanese-language academic corpora. This is not because demand does not exist, but because the path has not yet been blazed—for a first mover, this is both a risk and an opportunity.

13Valuation and Comparable Transactions

A limitation that must be acknowledged: no reliable EV/EBITDA can be given

Among Japanese academic publishing targets, aside from a few listed companies (such as Maruzen CHI Holdings) and organizations that disclose segment profit and loss (such as Chuo University Press), the overwhelming majority do not disclose financial data. Japanese local research institutions likewise cannot produce a reliable industry valuation multiple—because the targets are unlisted and comparable transactions are lacking. Any claim that "Japanese academic publishers are valued at an average of 8 times EBITDA" is fabricated.

This report therefore adopts bottom-up order-of-magnitude estimates rather than a multiples approach.

Comparable transactions and contracts

Order-of-magnitude estimates

Based on the above anchors and the scale of the three target categories, this report offers the following order-of-magnitude ranges (all labeled as estimates, not quotations):

The exchange-rate window: now is a buyer's market

The US dollar traded at 144.4 yen in April 2025 and rose to 160.7 yen in June 2026; the renminbi rose against the yen from 20.58 in August 2025 to 23.56 in August 2026. The yen weakened by about 14% against the renminbi, meaning Japanese assets are becoming cheaper in renminbi terms. At the same time, note that Chinese state-owned entities face stricter notification requirements in Japan (see the next section), so the exchange-rate advantage and institutional constraints must be weighed together.

14Transaction Structure Design: Five Paths

Core principle: exchange the minimum control for the maximum right to use content and data.

Path One: corpus first—collective licensing plus data processing

• Applicable targets: J-STAGE long-tail journals (Category F), Ichushi (Category C)
• Structure: sign data licenses with JST, learned societies, or publishing service providers; use a "license fee + processing service fee" structure; seek an exclusivity period or right of first renewal
• Advantages: does not trigger FDI screening, does not arouse public-opinion risk, can be validated step by step, and cash outlay is controllable
• Risks: non-renewal at the end of the license term; without exclusivity it is hard to build a barrier; an incomplete copyright chain
• Recommendation level: high (the preferred starting path)

Path Two: pipeline acquisition—buy a controlling stake in a publishing services company

• Applicable targets: Kokusai Bunken-sha (Category B), Nakanishi Printing (Category B)
• Structure: acquire a controlling stake while retaining the existing management, brand, and service commitments; focus the public narrative on "enhancing service capability and open access"
• Advantages: gains the production pipeline, submission systems, and a network of relationships with about 200 learned societies; an entry point for taking on more business
• Risks: customers (learned societies) are sensitive to changes in control; labor-intensive with limited gross margins
• Recommendation level: relatively high (the best medium-term path)
• Key design: include performance earn-out and service commitment clauses, with the society renewal rate as the core earn-out metric.

Path Three: publishing-rights succession—take over when international publisher contracts expire

• Applicable targets: JJAP/APEX of the Japan Society of Applied Physics (currently published by IOP), PTEP of the Physical Society of Japan (in deficit), and the various English-language clinical medicine journals
• Structure: sign a new contract publishing agreement with the society, committing to no price increase, expanded dissemination in the Chinese-speaking world, and higher APC waiver allowances
• Advantages: obtains global distribution rights and editorial control without buying assets—the best value for money
• Risks: exclusivity clauses in existing contracts; learned societies are most politically sensitive to "Chinese control"
• Recommendation level: relatively high (high return but requires waiting for a window)
• Basis for feasibility: in 2024 the Chemical Society of Japan transferred BCSJ and Chemistry Letters to OUP, and in 2025 Primates moved back from Springer to the society—proving that such relationships can both flow outward and flow back

Path Four: university press joint ventures—a low-political-cost entry point

• Applicable targets: Chuo University Press (Category E, deficit for two consecutive years), Keio University Press (Category E, corporate form)
• Structure: focus on co-editions, translation licensing, and digital distribution joint ventures; offer debt-restructuring-style cooperation to the most distressed presses
• Advantages: low political resistance; builds experience and contacts in Japanese publishing operations
• Risks: limited economic return; slow university governance processes
• Recommendation level: medium (pilot and learning)

Path Five: platform coupling—deep integration with the J-STAGE ecosystem

• Applicable targets: the JST / J-STAGE ecosystem (state-owned, not an acquisition)
• Structure: exchange technology, translation, and AI retrieval services for data access; jointly build Japanese JATS XML conversion capability directly aligned with the FY2029 mandate
• Advantages: matches an industry-wide imperative and provides low-cost access to structured full texts
• Risks: procurement and security review by a state institution; geopolitical sensitivity
• Recommendation level: relatively high (strategic positioning)

15Risks, Compliance, and Governance

An important correction: FDI screening is not the biggest obstacle

This is the most important divergence between this report and common assumptions. After reviewing notifications from Japan's Ministry of Finance, the current positive list of "designated business sectors" (Appended Tables 1 and 2) does not include publishing (4141), the news industry, academic societies, or research institutes; nor does the list of "core business sectors" include publishing. Therefore, acquiring a Japanese publisher or publishing services company alone generally does not trigger a prior notification obligation, requiring only an ex post report.

But there are two key exceptions:

1. If the target also operates broadcasting, telecommunications, or cybersecurity software businesses, or holds a natural science research institute with export-controlled technology, prior notification is triggered. A publishing services company taking on business from research institutes involving sensitive technology requires particular attention.
2. Chinese state-owned entities are the largest institutional variable. Foreign governments and enterprises controlled by them may not use the prior-notification exemption system and must file prior notification for all designated business sectors. The 2025 cabinet order already tightened the rules; the amendment to the Foreign Exchange and Foreign Trade Act submitted to the Diet in March 2026 adds five elements (50% foreign control, risk-mitigation notification, deemed proxy holding, ex post disposition orders, etc.), to take effect within one year of promulgation.

Operational recommendations: (1) obtain a formal legal opinion from a Japanese lawyer that "publishing is not a designated business sector"; (2) prioritize a market-oriented entity without a state-owned background or a joint-venture SPV as the transaction vehicle to avoid the strict rules for "foreign governments, etc."; (3) continue tracking the implementing details of the 2026 amendment.

Three substantive risks that must be addressed up front

Risk one: academic freedom and public backlash (high probability, high impact). This is the factor most likely to cause the deal to fail, and no financial clause can defuse it. Japanese academia is highly sensitive to "Chinese capital controlling academic publishing." Response: strictly limit the public narrative to "enhancing the international dissemination and open access of Japanese research"; do not replace editors-in-chief, interfere in topic selection, or alter editorial board composition; preserve full Japanese autonomy over editorial matters in the governance structure and lock it in through the charter.

Risk two: an incomplete copyright chain (medium probability, high impact). AI training uses require a clear rights basis. Author copyright arrangements in Japanese journals are not uniform, and co-author consents and third-party figure permissions may both be missing. Response: prioritize journals where authors retain rights or that already carry CC licenses; for long-tail corpora, use "title-by-title confirmation plus indemnity warranty clauses"; require the licensor in the license agreement to make representations and warranties regarding the completeness of rights.

Risk three: exclusivity in existing contracts (high probability, medium impact). The contract publishing agreements of most high-impact English-language journals are locked up long term by OUP, Springer Nature, Wiley, and IOP (public samples run about 3 to 3.5 years; 3 to 5 years is inferred), and exit clauses are not disclosed. Response: secure renewal windows in advance; compete on "service upgrade commitments" rather than "price"; for targets still within contract terms, pursue data licensing first and negotiate publishing rights once the contracts expire.

Other risks

16Action Roadmap

0–6 months: intelligence, relationships, and the corpus pilot

1. Assemble a local Japanese advisory team (lawyers, tax advisors, industry consultants) to issue formal opinions on FDI screening and ownership structures
2. Complete ownership and contract due diligence on the 25 targets, focusing on the term, exclusivity, and exit clauses of contract publishing agreements
3. Establish non-transactional cooperation (translation, XML conversion, international promotion) with 2–3 learned societies to build trust first
4. Sign the first long-tail Japanese journal data licensing pilot, covering about 100,000 articles, to validate the processing pipeline and model returns
5. Key decision point: if the model gains from the pilot corpus are not significant, scale back to pure service cooperation and do not proceed to an equity transaction

6–18 months: pipeline M&A and productizing capability

1. Acquire a controlling stake in a publishing services company (preferably a Kokusai Bunken-sha-scale target), retaining the existing management and brand
2. Productize Japanese JATS XML conversion capability to directly address the conversion imperative created by the 2029 mandate
3. Take over contract publishing for 1–2 English-language society journals at contract renewal windows
4. In parallel, apply for DOAJ certification for the journals taken over, delivering on commitments to societies through "enhancing open access"
5. Key decision point: if the society renewal rate falls below the agreed threshold, trigger an adjustment under the earn-out clauses

18–36 months: platform coupling and barrier formation

1. Embed deeply in the J-STAGE ecosystem through a joint venture, taking on XML production and internationalization services
2. Build a million-article-scale exclusive Japanese-language corpus and establish a terminology database, citation network, and retrieval layer
3. Replicate the entire "internationalization service capability of a Chinese publisher" into other language markets
4. Set up phased earn-out and buyback clauses, continuously assess geopolitics and exchange rates, and preserve an exit path

17Appendix A: Data Methodology

Data sources and bases

Three methodological limitations that must be stated

1. OpenAlex's coverage of Japanese-language journals is insufficient. OpenAlex indexes only 39,489 Japanese journal papers from 2023, while J-STAGE adds about 188,000 items a year—a coverage rate of roughly 20%. This report therefore uses OpenAlex for structural and impact analysis and J-STAGE and Crossref for corpus size estimation, without mixing the two.
2. OpenAlex's language tagging is unreliable. About 93% of its language tags for Japanese journal papers are English, a clear distortion. The Japanese-to-English ratio therefore uses J-STAGE's official language structure data.
3. Differences in disciplinary taxonomy. Chinese official data (WoS/SCIE 178 disciplines) uses a different taxonomy from this report's OpenAlex 26 disciplines, so absolute values are not directly comparable; however, the two are highly consistent on the conclusion that "China's domestic capacity rate is about 5%" (the official figure of 19.27 times versus this report's 19.5 times), indicating that the method is reliable.

How to reproduce the scoring model

Composite score = 100 × (0.22×D1 + 0.20×D2 + 0.18×D3 + 0.27×D4 + 0.13×D5) ÷ 5 − 18 × D6 ÷ 5

Here D1–D5 are better the higher they are, and D6 is the risk score (worse the higher it is, and subtracted). All sub-scores are based on the evidence in the body of this report and can be reproduced item by item in japan_ma/scripts/ma_data.py.

18Appendix B: Key Uncertainties

This report contains three irreducible uncertainties that must be clearly communicated to decision-makers:

1. Valuation multiples cannot be reliably provided. The overwhelming majority of Japanese academic publishing targets are unlisted, with no comparable transactions and scarce public financial data. What this report provides are order-of-magnitude estimates, which cannot serve as a basis for quotation.
2. The shareholder structures and succession intentions of small and mid-sized publishers are generally not public. Igaku-Shoin, Shokabo, Asakura Shoten, Kyoritsu Shuppan, Tokyo Kagaku Dojin, Nanzando, and Chugai Igakusha have not disclosed shareholder or succession arrangements. The ranking of "most likely to sell" is an analytical inference and does not represent actual willingness to sell.
3. The AI value of Japanese-language corpora lacks transaction evidence. There is no publicly disclosed case anywhere in the world of a Chinese AI company purchasing Japanese-language academic corpora. Corpus value can currently only be argued as synergy gains, not as the primary consideration.

19Appendix C: Target Scoring Details

The table below lists the six-dimension scores and composite scores of all 25 candidate targets. D1 strategic leverage, D2 discipline fit, D3 transferability, D4 AI corpus value, and D5 valuation and financial feasibility are all scored 0 to 5 (higher is better); D6 is integration and compliance risk (higher is worse, and is subtracted). The composite score formula is in Appendix A.

Category codes: A society journal portfolios; B publishing services and production; C secondary indexes and data; D commercial publishers; E university presses; F long-tail corpora.

20Appendix D: Main Data Sources

All quantitative conclusions in this report can be traced to the public sources listed below. Any judgment marked as unverified is explained in the body.

• JST, "J-STAGE no Genkyo to Zenbun XML Tosai Hissuka ni Tsuite" (On the Current Status of J-STAGE and the Mandatory Full-Text XML Submission), 2026-07 (FY2025 fiscal year-end figures) — J-STAGE holdings, number of organizations, full-text XML readiness rate, download geography, and the FY2029 full-text XML mandate. https://www.jstage.jst.go.jp/
• Crossref REST API full member registry (local tabulation) — 723 Japanese members, 2,768,775 DOIs, with Japan accounting for 1.59% of global DOIs. https://api.crossref.org/members
• OpenAlex API (local tabulation, 2026-09) — 1,797 Japanese sources, more than 40,000 journal-level metrics, and the disciplinary and language structure of Japanese journal papers. https://api.openalex.org/
• DOAJ public metadata snapshot (local tabulation) — 23,555 titles worldwide in DOAJ, 95 in Japan (0.40%); license and APC structures. https://doaj.org/api/v4/docs
• MEXT/NISTEP, Kagaku Gijutsu Shihyo 2025 (Science and Technology Indicators 2025) and Kagaku Kenkyu no Benchmarking 2025 (Benchmarking Scientific Research 2025), 2025-08 — Japan ranks 5th globally in paper volume, 13th in Top 10%, and 12th in Top 1%. https://www.nistep.go.jp/archives/61221/
• JUSTICE, 2024 Nendo Keiyaku Jokyo Chosa Kekka Hokoku (Report on the Results of the FY2024 Contract Status Survey), 2025-03 — member libraries' foreign journal subscriptions of 43.03 billion yen and domestic journals of 3.45 billion yen. https://contents.nii.ac.jp/
• JUSTICE, Ronbun Kohyo Jittai Chosa Hokoku (Survey Report on the Actual State of Paper Publication), 2025-03 (FY2023 actuals) — Japanese authors spent 12.94 billion yen on APCs in 2023. https://contents.nii.ac.jp/
• JSPS Kakenhi, Reiwa 7 Nendo Bumpai Jokyo Ichiran (FY2025 Allocation Status List), 2025-12 — the Grant-in-Aid for Publication of Scientific Research Results: 424 million yen in new allocations for FY2025, or 711.7 million yen including continuing allocations. https://www.jsps.go.jp/
• Association of Japanese University Presses (AJUP) membership directory and new title bulletins, 2025-2026 — 24 member presses and 535 new titles in 2025. https://www.ajup-net.com/
• Chuo University FY2024/2025 Financial and Business Report — the only university press to disclose profit and loss: contract publishing accounts for 87.4% of revenue, with deficits for two consecutive years. https://www.chuo-u.ac.jp/
• Japan Publishing Association / Research Institute for Publications, Shuppan Shihyo (Publishing Indicators), 2026-01 — the 2025 publishing market was 1.5462 trillion yen, with print media at 964.7 billion yen. https://shuppankagaku.com/
• Cabinet Office, "Gakujutsu Ronbun To no Sokuji Open Access no Jitsugen ni Muketa Kihon Hoshin" (Basic Policy for the Realization of Immediate Open Access to Academic Papers), 2024-02 — the 2024 immediate OA policy: immediate green OA for outputs of designated competitive funding, applicable from FY2025. https://www8.cao.go.jp/cstp/oa_houshin_faq.pdf
• Yoshiko Tanahashi et al., "Nihon no Gakkyokai Journal Shuppan no Genjo" (The Current State of Journal Publishing by Japanese Learned Societies), Joho no Kagaku to Gijutsu 69(11), 2019 — all journals that held a Q1 ranking for three consecutive years are published under contract by international publishers; the ESCI review rejection rate is about 54%. https://www.jstage.jst.go.jp/
• JaLC materials for the 12th Dialogue Meeting, 2026-02 (end-2025 figures) — 86 full members, 3,448 associate members, and 14.14 million cumulative DOIs. https://japanlinkcenter.org/
• Chemical Society of Japan announcement of the transfer of BCSJ/Chemistry Letters to OUP, 2024-01 — publishing taken over by Oxford University Press from January 2024. https://www.chemistry.or.jp/
• Physical Society of Japan public statements and business plans concerning PTEP — after JSPS funding ended, PTEP developed a structural deficit and APC waivers are reviewed case by case. https://www.jps.or.jp/
• Wiley press release, 2013-09 / Cancer Science journal page — Cancer Science converted to full OA with Wiley in 2014. https://www.wiley.com/
• Science Council of Japan recommendation, 2020-09-28 — four reform proposals: a unified rate offset, a journal publishing services company, and a Japanese-language citation index. https://www.scj.go.jp/
• Ministry of Finance, Gaikoku Kawase oyobi Gaikoku Boeki Ho (Foreign Exchange and Foreign Trade Act), inward direct investment, designated business sectors, Appended Tables 1 and 2 — the positive list of designated business sectors does not include publishing (4141), the news industry, academic societies, or research institutes; the core business sectors also do not include publishing. However, foreign governments and enterprises controlled by them may not use the prior-notification exemption. The March 2026 amendment bill to the Foreign Exchange and Foreign Trade Act adds 5 items. https://www.mof.go.jp/policy/international_policy/gaitame_kawase/fdi/
• Kokusai Bunken-sha Co., Ltd. company profile (2026-06) — founded in 1951; revenue of 1.877 billion yen and about 180 employees as of June 2026; contracted by about 200 learned societies and handling 244 J-STAGE journals. https://bunken.co.jp/
• Nakanishi Printing Co., Ltd. company profile — founded in 1865; eighth-generation president Hidehiko Nakanishi (2016); capital of 20 million yen; long-standing partnership with OUP. https://the.nacos.com/nakanishi/
• Teikoku Databank / Research Institute for Publications, publisher operating statistics — in FY2023, 36.6% of 675 publishers ran deficits and 66.1% saw deteriorating performance (the worst in nearly 20 years); in FY2024 there were 31 publisher bankruptcies, 1.8 times the 17 of the previous year. https://www.tdb.co.jp/
• The journal Primates moving from Springer back to a Japanese society in 2025 — a precedent for the "return" of society journal publishing rights: proof that contract publishing relationships can be reversed. https://www.springer.com/journal/10329
• Oh-Ebashi LPC & Partners, interpretation of inward direct investment regulations, 2025-10 — the 2025 cabinet order tightened the rules; foreign governments and the enterprises they control must file prior notification for all designated business sectors. https://www.oslaw.jp/

Data calculated by this report

• Structure and impact of Japanese academic publishing: OpenAlex API, retrieved September 2026; 1,797 Japanese journal sources and 478 Japanese publishers; sources with anomalous metadata (source merges) excluded.
• DOIs and member registration: full Crossref REST API member scan; 723 Japanese members and 2.769 million DOIs.
• Domestic journal capacity rates and disciplinary gaps for China and Japan: OpenAlex API, filtered by authorships.countries and primary_location.source.id, type:article, 2023. Chinese authors 812,907 / Chinese journals 41,765; Japanese authors 139,037 / Japanese journals 39,489.
• Open access structure: OpenAlex API, tabulating the OA status composition of 2023 journal articles by author country.
• Corpus size estimation: based on 6.002 million J-STAGE documents and 2.769 million Crossref DOIs, estimated at 8,000 to 12,000 tokens per article.