Japan Company Durability: What 530 Assessments Show
530 Japanese companies assessed July 2026: 248 rated Hard to displace, 100 losing its protection. Who holds scarce assets and who is exposed to automation.
Japan's corporate landscape splits into two very different stories in our data. Of the 530 Japan-based organizations assessed in July 2026 within WillItSurvive's assessment of 10,525 companies, 248 earned a Hard to displace standing while 100 landed losing its protection. The dividing line is not size or age. It is whether the company controls something scarce, such as land, licenses, physical networks, or institutional trust, versus something software can copy.
Most of the 530 Japanese companies assessed in July 2026 sit toward the calmer end of the spectrum, with 34 companies flagged as high pressure. That range hides extremes: rail operators like JR Central, Keio, Tokyu, and Hankyu Hanshin show almost no exposure, while DMM Bitcoin, already folded into SBI VC Trade after a 2024 hack wiped out its customer base, sits at the opposite end of the Japan set.
The durable core: rails, banks, and regulated franchises
Japan's rail operators are the calmest corner of the entire country dataset. JR Central, which runs the only Tokyo-Nagoya-Osaka bullet train line, shows almost no exposure among the 530 Japanese companies assessed in July 2026, and JR East, JR West, Odakyu, Tobu, Nankai, and Keisei all carry Hard to displace standings alongside it. A physical rail network with station real estate is not something an AI product replaces.
The same logic holds for finance. Mitsubishi UFJ, Sumitomo Mitsui Financial Group, Mizuho, Nomura, Tokio Marine, Dai-ichi Life, and Nippon Life all carry Hard to displace standings in the July 2026 assessment. Nippon Life, a mutual insurer with roughly ¥96 trillion in assets and no public shareholders, runs on long-term trust rather than a product a chatbot can substitute. Institutions round out the picture: the University of Tokyo and Kyoto University both rate Hard to displace, since degree-granting authority and research reputation are not copyable.
Manufacturing with deep technical moats also holds up. Fujifilm spans medical imaging, drug manufacturing, and chip materials, businesses with regulatory barriers, and rates Hard to displace in the July 2026 assessment. Keyence, Nintendo, FANUC, Hitachi, Sony Group, Tokyo Electron, and Shin-Etsu Chemical all sit in the same durable band.
Where the pressure is landing
Anywhere a Japanese business sells a service that can be automated, undercut, or matched by a cheaper digital rival, the July 2026 data shows movement, and often the losses are already in revenue, not projections.
Advertising is the clearest case. Hakuhodo DY Holdings, rated losing its protection in the July 2026 assessment, saw revenue fall 9.7% year over year as platform-native tools like Google Buyer Direct absorb the media-buying commissions agencies used to bill for. Dentsu carries the same losing its protection standing, and Digital Holdings (Opt) was absorbed by Hakuhodo DY itself.
Retail middlemen are next. Askul's sales dropped roughly 95% after an October 2025 ransomware attack froze ordering and shipping for weeks, and Amazon Business Japan and Kokuyo's Kaunet picked up the stranded customers; Askul rates losing its protection in the July 2026 assessment. Shiseido, rated losing its protection in the July 2026 assessment, watched its Drunk Elephant brand fall 65% year over year, forcing job cuts in the Americas while L'Oreal grows faster.
Autos show a distinct national-level strain. Nissan rates losing its protection in the July 2026 assessment, its US market share down from about 9% to about 6% over a decade as BYD, Geely, Hyundai, and Kia sell cheaper, faster EVs. Mazda, Mitsubishi Motors, Hino, and Subaru all rate losing its protection alongside it. Toyota and Honda both rate defensible for now. By contrast, suppliers with hard assets, Aisin, Bridgestone, Isuzu, and Denso, keep Hard to displace standings.
Education subscriptions are eroding too. Benesse, rated losing its protection in the July 2026 assessment, is losing Shinkenzemi students to Recruit's Studysapuri and free AI homework tools. Z-kai, DMM Eikaiwa, and cram schools Sundai and Yoyogi Seminar all show elevated pressure.
Already gone: the extreme end of the distribution
Some entries in the Japan set are cautionary records rather than going concerns. DMM Bitcoin, which our July 2026 assessment placed as the most exposed company in the Japan set, was folded into SBI VC Trade after a 2024 hack wiped out its customer base. Mercari Hallo, the gig-work app, closed in December 2025 with Timee taking over its users and employer relationships; it rates Little protection left in the July 2026 assessment. Crypto exchanges broadly sit under strain: Coincheck and bitFlyer both rate losing its protection, and GMO Coin and BITPoint both rate protection under test.
Enterprise software middlemen are squeezed from a different direction. Treasure Data, rated losing its protection in the July 2026 assessment, faces Salesforce bundling customer data management directly into its CRM suite while warehouse-native tools let companies skip a dedicated data platform entirely.
How to read this without predictions
WillItSurvive never declares an outcome for any company. Standings describe evidence as of the July 2026 assessment window, and forward-looking material is published only as dated watch-items. For Japan, the watch-items on record include: whether Hakuhodo DY's revenue keeps declining year over year (which would confirm platform ad buying is a permanent shift, not a dip); whether Askul's next quarterly report, checkable by December 2026, still shows revenue below pre-attack levels; whether Nissan's US share stays at or below 7% through March 2027; whether Shiseido reports further Drunk Elephant declines or Americas restructuring by January 2027; whether Sakura Internet's GPU cloud growth offsets falling legacy hosting; and whether Treasure Data loses named enterprise contracts to Salesforce Data Cloud. Each is something a reader can verify directly.
That confidence rests on a backtest of 116 companies that later failed, using only facts that were knowable 18 months before each one's collapse: the assessment flagged nearly all of them as high-risk in advance, with none rated safe. The supported claim is separation, not prophecy.
Questions this data answers
Which Japanese companies show the least AI-era pressure? Rail operators dominate the low end: JR Central, Keio, Tokyu, and Hankyu Hanshin show almost no exposure in the July 2026 assessment. Megabanks, top universities, and mutual insurers like Nippon Life sit close behind, all with Hard to displace standings.
Which Japanese companies face the heaviest pressure right now? DMM Bitcoin, already absorbed into SBI VC Trade, Mercari Hallo, closed in December 2025, and Nissan top the list of the most exposed companies in the July 2026 assessment. Askul, Shiseido, Dentsu, Mazda, Hino, and Coincheck all rate losing its protection, with revenue damage already visible in most cases.
Is Japan overall holding up better or worse than the pattern suggests? Of the 530 Japanese companies assessed in July 2026, 248 (about 44%) rated Hard to displace against 100 losing its protection and only 6 Little protection left. The mix reflects a country whose durable core of infrastructure, finance, and precision manufacturing outweighs its exposed middleman and consumer-service layer.
Are Japan's ad agencies losing to automation? The July 2026 data says yes, measurably: Hakuhodo DY's revenue fell 9.7% year over year as Google Buyer Direct and similar platform tools replace agency media-buying commissions, and its subsidiary Opt was absorbed after a buyout. Dentsu carries an losing its protection standing as well.
What separates the durable Japanese companies from the exposed ones? Control of something scarce: land, licenses, physical networks, regulatory barriers, or institutional trust. Companies with those assets, like JR Central or Fujifilm, hold steady even without an AI strategy, while middlemen, subscription tutors, and standalone data platforms show losses already appearing in revenue.
Assessments reflect evidence available July 2026. Companies change; treat watch-items as things to verify, not conclusions.
Figures are as of each company's assessment date unless dated otherwise.