APOLLO11 (EdgeLLM)
Category: AI Infrastructure
Japanese AI infrastructure company (Nagoya) offering EdgeLLM, a fixed-price, zero-external-transmission on-premise generative AI platform for law firms, hospitals, and government agencies, starting at ¥9.8M (~$66K). APOLLO11 (EdgeLLM) was founded in 2011. The company is led by Sho Yoshimaru (吉丸彰). Based in Nagoya, Aichi, Japan. Team size: 1-10. Latest round: Bootstrapped.
- Founded
- 2011
- Headquarters
- Nagoya, Aichi, Japan
- Team size
- 1-10
Value proposition
Enables organizations that cannot send confidential data to external servers (law firms, medical institutions, government agencies) to safely use frontier-level generative AI entirely within their own LAN — with zero external transmission of inference data, zero usage-based fees, and no vendor lock-in.
Products and solutions
EdgeLLM (on-premise generative AI platform, launched Aug 2026, packages: Vision Pack S ¥9.8M, Agent Pack M ¥19.8M, High-Context Pack ¥19.8M), APOLLO11-FUSION (domestic LLM API, launched Aug 2026), NuraGrid (general-purpose business AI app with chat UI, agent builder, auth management), marketing SaaS, managed server business.
Unique value
Fully on-premise, air-gap-capable generative AI with published transparent pricing (initial fee + fixed monthly, no usage-based fees), OpenAI-compatible API for drop-in integration, ~240W low-power single-node design installable in an office (no server room), and legal-risk resolution (Lawyer Act Art.23, APPI, Unfair Competition Prevention Act).
Target customer
Law firms, hospitals/medical institutions, government agencies/public sector, financial institutions, and manufacturers holding trade secrets that are restricted by law or security policy from using cloud/SaaS AI.
Industries served
Legal, Healthcare, Government/Public Sector, Manufacturing, Financial Services
Technology advantage
Deploys open-weight models (Qwen, DeepSeek, Apache 2.0/MIT licensed) on NVIDIA GB10 nodes with vLLM/LiteLLM inference stack; OpenAI-compatible API endpoint inside LAN; air-gap and closed-network operation; full internal audit logging; quarterly model updates; 5-6 week standard deployment; 3-year hardware warranty.
How they differentiate
Unlike cloud API usage (which leaks data externally even via VPN) and self-built LLM infrastructure (costing hundreds of millions of yen), EdgeLLM offers a turnkey, transparently-priced on-premise package at ¥9.8M entry, with zero external transmission, zero usage fees, and no vendor lock-in — targeting mid-size organizations that can't afford custom builds.
Main competitors
Sakana AI, Preferred Networks, cloud-based frontier AI providers (OpenAI, Anthropic/Claude) via VPN/API, other on-premise/private LLM deployment vendors in Japan.
Key partnerships
JETRO Global Startup Acceleration Program (GSAP) selected (2025), Mouseflow (web analytics) Japan distributor, Lenovo PGX hardware (standard), partnership with Fumore (business alliance).
Major milestones
Founded April 28, 2011, launched APOLLO11-FUSION domestic LLM API (Aug 2026), launched EdgeLLM on-premise generative AI platform (Aug 31, 2026), selected for JETRO GSAP (2025).
Market positioning
Positioned as a "third option" in the Japanese generative AI market: between cloud API usage (data-leak risk) and self-built LLM infrastructure (too expensive), targeting regulated/confidential-data organizations. Aims for 10 deployments in 2026 across legal, medical, and government sectors.
Geographic focus
Japan (domestic focus), with HQ in Nagoya and Tokyo office; GSAP program for overseas expansion.
About Sho Yoshimaru (吉丸彰)
Ex-Island Brain (intern then employee, 2005); Ex-Ateam (2008). Nagoya University Faculty of Economics (enrolled 2003, withdrew 2005). Founded APOLLO11 in 2011. Studied Experimental Interaction Design at Tallinn University of Technology (Estonia, 2013); Stanford d.school BootCamp (2025).
Latest news about APOLLO11 (EdgeLLM)
More AI Infrastructure companies
Official website: https://apollo11.co.jp/
