基盤モデル・インフラ. Models, platforms and the infrastructure that makes AI usable at scale. 26 companies in central Tokyo, most active first.
Claude foundation-model lab with Anthropic Japan GK; Tokyo office and AISI cooperation for safe AI evaluation.
Leadership: 東條 英俊 (代表取締役社長 / Head of Japan)
AI-native GPU cloud and inference platform expanding US–APAC infra, with Japan sovereign-AI plans.
Leadership: Alex Yeh (Founder & CEO)
Physical-AI startup developing robotics foundation models (RLDX-1) for human-level five-finger manipulation in industrial settings; Japan entity in Marunouchi.
Leadership: リュ・ジュンヒ (Ryu Jung-hee) (代表 (RLWRLD, HQ US)) · 李フン (Lee Hoon) (代表 (日本法人))
SoftBank's wholly owned AI arm, researching and developing domestically built large language models specialised for Japanese and the generative AI services on top of them.
Leadership: Yoshihisa Ijiri (井尻 善久) (代表取締役社長 兼 CEO (Representative Director, President & CEO))
Italian physical-AI cybersecurity company (Photon runtime security + foundation models) expanding into Japan with a local entity and Tokyo office.
Leadership: Gianni Cuozzo (Founder & CEO)
Vertically integrated Japanese AI company spanning its own AI chips, computing infrastructure, generative AI foundation models, and industry solutions.
Leadership: Toru Nishikawa (Co-Founder, Representative Director & Chairman) · Daisuke Okanohara (Co-Founder, Representative Director & President (CEO))
Tokyo-based AI research company building frontier foundation models, including a Japanese-specialised LLM API and a model-orchestration layer over frontier models.
Leadership: David Ha (Co-Founder & CEO) · Llion Jones (Co-Founder & CTO)
Operates Japan satellite-data PaaS Tellus after MBO from Sakura Internet; expanding satellite-data×AI for defense and overseas markets.
Leadership: Hideto Yamazaki (山﨑 秀人) (Representative Director & CEO)
Japan's national industrial research institute operating ABCI open AI compute and co-developing physical-AI foundation models.
AWS Japan delivering Bedrock and cloud AI infrastructure for Japanese generative AI workloads.
Shibuya AI infra startup; CC Gov-LLM on 源内 trial and CC AI Factory for enterprise-owned AI stacks.
Leadership: 木下 寛士 (代表取締役社長 CEO)
Internet infrastructure operator whose GMO GPU Cloud provides NVIDIA HGX B300/H200 clusters with Spectrum-X networking in Japanese data centres for large-scale AI training.
Leadership: Tadashi Ito (伊藤 正) (代表取締役 社長執行役員 (Representative Director, President & Executive Officer))
Google Cloud's Japan arm delivering Vertex AI, Gemini, and hyperscale AI infrastructure.
Major Japanese telecom co-providing ELYZA Japanese LLM for government 源内 trial and enterprise generative AI.
Frontier voice AI company building end-to-end real-time speech models for East Asian languages, licensed to enterprises via API and SDK.
Leadership: Noriyuki Kojima (Co-Founder & CEO) · Jungo Kasai (Co-Founder & CTO)
Microsoft's Japan subsidiary for Azure OpenAI, Copilot, and enterprise AI cloud.
Developer of Tsurugi, a government-backed, domestically-developed next-generation relational database combining high-speed batch and online processing, built to provide a low-latency execution environment for local LLMs and MCP-based AI wor
Leadership: Hiroshi Kanbayashi (神林飛志) (Chairman & Representative Director (代表取締役会長))
Japanese IT major developing cotomi LLM / generative AI stack for enterprise and government (源内 trial).
The Japan subsidiary (GK) of global intelligent data infrastructure company NetApp, providing the ONTAP-based data platform plus AI Data Engine and automation to keep enterprise data across on-premises and cloud continuously AI-ready.
Leadership: Chiharu Saito (斉藤千春) (Representative Executive Officer, President (代表執行役員社長), NetApp Japan)
NII-led LLM-jp consortium building and releasing open, Japanese-strong large language models.
Consortium-backed startup building Japan multimodal foundation models for physical AI / robotics (Noetra + AIST NEDO program).
Leadership: 丹波 廣寅 (代表取締役社長)
Listed telecom and R&D group that develops tsuzumi, its own from-scratch Japanese large language model designed to run on a single GPU with 40GB or less of memory.
Global foundation-model lab with Japan entity; SB OAI Japan JV brings Crystal intelligence / enterprise OpenAI to Japanese companies.
Former PFN-group entity that built Japan-made PLaMo LLM from scratch; absorbed into Preferred Networks in July 2025.
Leadership: 岡野原 大輔 (代表取締役社長 (at merger announcement))Japanese generative-AI company (originated as Microsoft's Rinna team) building Japanese LLMs plus avatar, speech, and custom-model products such as Tamashiru Custom and Koemotion.
Leadership: Juhon Son (宋 珠憲) (Representative Director)
Diversified tech group and Noetra core backer advancing multimodal / physical AI for sensing, robotics, and industry.