Japan's Demand for Already-Landed Embodied-AI Tech × Where Our Software Fits — Otsuka Shokai channel needs + macro direction (labor shortage · Labor-Saving Investment Subsidy · RFA standards) + concrete pain points (warehouse utilization stuck at 60–65% · zero cross-brand fleet management in food service · siloed inspection data · care not wired to nurse call) + one match table

⚠️ Research snapshot (2026-07-02).

Building on the previous piece Landed Domains × Borrowable Software — that one laid out "six already-landed domains + a five-layer software stack (open vs closed) + the four entry points where we can build software."

This piece investigates specifically [Japan]'s demand for those technologies, landing it in three blocks + one match table, to see clearly "which of our entry points squarely hit Japan's (especially Otsuka Shokai's) needs."

Method: three-track parallel investigation (Otsuka Shokai / macro direction / concrete pain points) → each key figure/policy/case carries an external link, aggregator-site dates re-verified against official sources, and external links self-checked for reachability via curl before being committed (this round's key links were batch-curled, all HTTP 200). Flagged as CONFIRMED (verified against a primary source) / uncertain.

Sister pieces: Coordination-software landscape · SEER architecture teardown.

🧭 One-line overview

Japan's demand structure is extremely favorable to us — serving / cleaning / transport / inspection hardware has been amply supplied and deployed at scale by Chinese and Japanese vendors, and the real gap lies entirely in the upper software layer:

🎯 Multi-brand unified orchestration　+　integration with existing IT · elevators · POS · nurse call · MES　+　O&M SaaS

This is precisely the main direction ① (heterogeneous fleet orchestration middleware / marsupial coordination) we selected last round. And Otsuka Shokai happens to be the ideal channel target customer for this path.

One-line conclusion:

✅ Making "a neutral heterogeneous fleet orchestration middleware + a systems-integration gateway + O&M SaaS, white-labeled to SI/channels" the main thrust simultaneously hits three things — Japan's macro policy direction, its concrete industry pain points, and the procurement logic of SI channels like Otsuka Shokai.

Broken into three sentences:

1. Macro — the labor shortage is an irreversible hard demand, and the government is underwriting "heterogeneous orchestration + systems integration" along two lines: subsidies + standards.
2. Micro — every landed industry shows the same pattern: single machines are mature, but the moment you mix brands or connect systems, there's no neutral unifying layer.
3. Channel — Otsuka Shokai doesn't build hardware; it treats robots as "a commodity it can bundle into a monthly subscription," and naturally needs middleware like ours.

---

1. What kind of robot software does 【Otsuka Shokai】 want as a channel/SI (🌟 focus)

Positioning (CONFIRMED, primary source otsuka-shokai.co.jp)

• Founded 1961; FY ending 2025/12 consolidated revenue ¥1.3228 trillion (a record high for three straight years), consolidated headcount 10,079¹.
• About 295,000 client companies (≈8% of all Japanese firms), centered on SMBs².
• Its two core businesses = SI (system deployment) + S&S (post-deployment O&M / consumables mail-order "たのめーる" · subscription IT support "たよれーる" / Tayoreru), with a headline promise of "one-stop resolution of your troubles (お困りごと)."

⚠️ Contrast note: It is a pure channel reseller + deployment-support partner that does not build robot hardware; robots are just one emerging category in a vast catalog, with a tiny revenue share and no standalone IR disclosure. "Largest independent-system SI" is a phrasing the media commonly uses; there is no such original sentence on the official site (uncertain).

Robots it already resells

(models CONFIRMED; vendor names are mostly industry common knowledge, not directly labeled on the site = uncertain)³

• Reception / serving: temi · temi GO (Robotemi)
• Cat-shaped serving: Mars (Pudu)
• Transport: kachaka Pro (Preferred Robotics)⁴
• Commercial cleaning: Phantas = Gausium (cross-confirmed)⁷
• Home cleaning: RC300DZ (labeled on Makita's official site)
• RPA: たよれーる WinActor (NTT-AT)

Not found: a clear product page for it reselling standalone AGV/AMR, collaborative arms, or drones (uncertain / to be verified).

What "software/integration capability it wants" (CONFIRMED signals)

1. Multi-brand/multi-unit unified management + cloud single-pane management

   temi has a web platform "temi Center" for managing multiple units; it simultaneously resells Pudu / Gausium / Preferred / temi / Makita, so cross-brand unified management is a natural hard requirement (today each vendor has its own App/cloud, with no unifying layer).

2. Integration with existing IT systems

   Its food-service DX solution page explicitly says "self-order + robot serving + seamless system integration to lower the SG&A ratio," landing on its own core "SMILE V" ERP, cloud CRM, and document management⁵. → Robots must be able to connect to accounting / ERP / office / CRM, not be islands.

3. Monthly subscription + one-stop O&M bundling

   temi goes via "initial ¥64.2k + operation coaching ¥42.8k + monthly ¥33k lease (incl. faulty-unit replacement / contact center)"; RPA goes via "たよれーる WinActor monthly ¥128k + 3 on-site visits/year"⁶. → The robot software you sell to it had best fit into this "subscription + support" model.

4. End-to-end support from pre-deployment assessment → build → O&M

   What it wants is software that comes with matching SI/O&M services, not bare machines.

✅ What this means for us: The correct posture toward Otsuka Shokai = a neutral middleware for "multi-brand robot unified management + integration with SMILE V/office/elevators/POS + white-labelable into Tayoreru subscriptions" — aligned with the direction of entry point ①, and a low-CAC path of "sell to the channel, don't face the end customer directly."

---

2. Japan's society-wide "directional (macro)" demand

Population / labor shortage = irreversible hard demand (CONFIRMED)

• The working-age population (15–64) peaked at 87.26M in 1995 → 45.35M by 2070 (about a 40% drop)⁸.
• A labor supply shortfall of about 11M people by 2040, with mechanization / automation named the top countermeasure⁹.
• Labor-shortage bankruptcies 350 cases in FY2024, a record high for two straight years (construction 111 · logistics 42)¹⁰.
• The effective job-openings-to-applicants ratio for FY2025 = 1.20, characterized by the MHLW as a "long-term and sticky" structural problem¹¹.

Policy and subsidies = directly paying to subsidize robots (CONFIRMED)

• ★ SMB Labor-Saving Investment Subsidy (new in 2024, ~¥500B scale)

  Catalog-order type with a 1/2 subsidy rate, cap of ¥2M/¥5M/¥10M by headcount, rolling acceptance from 2024-06-25, operated by SMRJ. The catalog scope explicitly includes: cleaning robots, serving robots, automated guided vehicles AGV·AMR (transport), automated warehouses · picking systems¹². → This is currently the strongest lever getting SMBs to buy robots.

• ★ RFA = the Robot Friendly Facility Promotion Organization (一般社団法人, founded 2022-08)

  Led by METI/NEDO, four standardization areas = elevator linkage, security linkage, physical environment, multi-robot (fleet) management¹³. It has already produced a robot–elevator sharing interface standard RFA B 0001:2025, but the industry notes that "the technical spec is set, yet the concrete methods for risk assessment / safety measures are unclear, which stalls rollout" — the standard exists, but landing / safe implementation is still a gap¹⁴·¹⁵.

  🎯 Policy is treating "elevator linkage + multi-robot fleet management" as the landing obstacles to overcome, on the same wavelength as our middleware direction.

• Others: the Monozukuri Subsidy, the Business Improvement Subsidy; the Robot New Strategy 2015 → NEDO 2023 "Grand-Scale Action Plan" sequel.

Major directions (CONFIRMED figures)

3. Japan's society-wide "concrete (micro)" demand: pain-point tasks × existing supply × gaps

🔑 The same pattern runs through every industry: single-machine hardware is mature and already deployed at scale; the gap lies entirely in "cross-brand unified orchestration + integration with existing systems + O&M SaaS."

🚚 Logistics warehouses / transport

• Existing supply: Rapyuta (PA-AMR picking), MUJIN (depalletizing / automated warehouses), Doog (Thouzer follow-me carts), Geek+/Locus.
• Gap (CONFIRMED): warehouses run 3–5 brands of robots simultaneously, each with its own software stack unaware of the others → hardware utilization stuck at 60–65%; and mainstream orchestration covers only AMRs / forklifts, with humanoids not yet included in the mixed fleet²⁶. Japan's humanoid pilots are all deployed unit-by-unit with no fleet layer (INSOL-HIGH × Yamazen using AgiBot G1)²⁷.

🍜 Food service / dining

• Existing supply: Pudu / Bear / Keenon / OrionStar, roughly 15,000 serving robots across Japan (2025-08).
• Gap (CONFIRMED): fleet management is limited to same-vendor same-model only, cross-brand unified orchestration does not exist; bussing / POS / KDS rely on one-off custom builds (Sojitz Robotics: doable if there's a public API, but secondary development is time-consuming and costly) with no standard middleware²⁸. METI's "Robo-Friendly × Food" is pushing standardization of the plating-step interface²⁹.

🏭 Factories / manufacturing

• Existing supply: FANUC CRX, Yaskawa MOTOMAN NEXT, UR, Techman.
• Gap (CONFIRMED, strong evidence): Panasonic Connect's "Robo Sync" (2025-10, single-pane control of multi-vendor arms / hands / cameras, a 26-company alliance, ¥50k/month·unit) self-proves that "heterogeneous unified management" is a real need — but explicitly does not cover MES / upper-system linkage³⁰. Yaskawa / FANUC each entrench their own ecosystems, which actually amplifies the cross-brand gap.

🔍 Inspection / facility maintenance

• Existing supply: Spot (SoftBank reseller) + its own cloud Orbit, ANYmal, Mitsubishi Heavy Industries EX ROVR.
• Gap (CONFIRMED, strong evidence): each brand comes with its own cloud, data siloed; J-Power's geothermal plant needed a third-party IIU AIMOS-DX just to unify Spot data³². Yokogawa OpreX RMC (2024-03) already does single-pane management of "multiple and heterogeneous robots," integrating ANYmal X in 2026-02 — a leading heterogeneous-orchestration competitor, but bound to Yokogawa's OT ecosystem, leaving room for a neutral cloud-native layer³¹.

🧑‍⚕️ Care (介護)

• Existing supply: transfer (SASUKE / HAL), watch-over/見守り (けあマルシェ), toileting (DFree), communication (PALRO).

• Gap (CONFIRMED): adoption is extremely low (80.9% not deployed, bathing only 8.4%, transfer facility-based 6.2%), with the field complaining "it's slower than human hands." Three gaps =

  • Multi-brand unified management (watch-over has already spawned aggregation platforms proving willingness to pay, but they cover only one category and are vendor-locked)
  • Linkage with nurse call / care records
  • O&M SaaS

  The MHLW's 2025 revision newly adds "linkage with other machines/systems," a policy tailwind³³·²².

🧹 Cleaning

• Existing supply: Gausium / SoftBank Whiz / Amano.
• Gap: cleaning robots and other robots run on independent systems, with no same-layer coordination / time-slot orchestration and no unified O&M panel; cross-floor movement today still relies on people carrying robots into elevators (uncertain: no official figure of comparable rank obtained for Japan's cleaning-robot deployment scale).

4. Match table: Japan's demand × already-landed tech/software × existing supply × the software points we can enter × fit

Fit: ★★★ highest (dead center of the main thrust) / ★★ medium / ★ low (red ocean or high barrier). Entry-point numbering follows last round: ① heterogeneous fleet orchestration middleware / marsupial ② data curation/eval SaaS ③ VLA eval + edge inference ④ vertical integration/O&M.

5. Landing recommendations for the company

1. Main thrust unchanged, and strongly validated by the Japanese market

Entry point ① (a neutral heterogeneous fleet orchestration middleware + marsupial coordination + a systems-integration gateway + O&M SaaS) has a clear gap in every landed industry in Japan, and is endorsed by two policy lines — the Labor-Saving Subsidy + RFA standardization.

2. The best landing form = a white-label middleware aimed at SI/channels (rather than selling to end customers directly)

Otsuka Shokai is the ideal target-customer/channel profile: 295k SMB clients, already resells multi-brand robots, explicitly wants multi-brand management + SMILE V connection + monthly-subscription O&M — our middleware can be white-labeled into its "たよれーる."

3. Priority vertical wedges (★★★ and with controllable barriers)

• Food service: cross-brand serving / cleaning unified backend + POS/KDS middleware
• Logistics/warehousing: marsupial + heterogeneous fleet
• Care: cross-brand + nurse-call / care-record linkage (riding the MHLW "linkage" policy)
• Inspection: Yokogawa RMC has already staked out the OT side, so we differentiate via "cloud-native / multi-CMMS / neutral"

4. Honest caveat (don't oversell the gap)

The middleware space already has players staking claims: PLiBOT (Obayashi affiliate, claiming vehicle/brand-agnostic unified control)³⁴, Panasonic Robo Sync, Yokogawa RMC, the VDA5050 ecosystem; on the construction side, Takenaka + Kajima + Obayashi + Fujita launched a project in 2025-10 for "software standardization · cross-model reusable SI modules," with officials calling out that "each robot is developed separately and the technology can't be reused"³⁵.

🎯 Differentiation should focus on the combination points not yet fully captured by any single player: safe, compliant integration + unified O&M monitoring + white-labelable to SI channels for multi-vendor AMRs + elevators / access control / POS / nurse call — not a generic "unified orchestration."

5. Hidden barriers

The Fleet Adapter must interface with each vendor's private API + elevators / POS / nurse call — the dirtiest, heaviest moat; real-robot closed-loop validation requires real machines or high-fidelity simulation.