Skip to main content

Turing Inc.

Category: Robotics / Embodied AI

A Japanese deep-tech startup developing fully autonomous electric vehicles (EVs) powered by End-to-End (E2E) multimodal generative AI. Turing Inc. was founded in 2021. The company is led by Issei Yamamoto. Based in Tokyo, Japan. Team size: 50-100. Total funding raised: ¥20.0B+ (approx. $130M). Latest round: Series A First Close (¥15.3B, Nov 2025). Key investors include ["JIC Venture Growth Investments","Global Brain","Denso","GMO Internet Group","Yanmar Ventures"].

Founded
2021
Headquarters
Tokyo, Japan
Team size
50-100
Total funding
¥20.0B+ (approx. $130M)

Value proposition

Replacing traditional rule-based/LiDAR-heavy autonomous driving with a camera-first, End-to-End AI approach that uses Large Language Models (LLMs) to understand complex road situations and 'long-tail' scenarios like human drivers.

Products and solutions

["Heron (Multimodal Generative AI Model for autonomous driving)","Terra (Generative World Model for traffic simulation)","Tokyo30 (Autonomous driving project/dataset aiming for 30 mins of intervention-free driving in Tokyo)","Gaggle Cluster (Dedicated GPU computing infrastructure)","THE FIRST (Original concept car/prototype)"]

Unique value

Unlike competitors (e.g., Waymo) that rely on HD maps and modular software stacks (perception -> prediction -> planning), Turing utilizes a pure 'End-to-End' deep learning approach where the AI directly maps camera inputs to steering/braking outputs, similar to Tesla but with a heavier emphasis on LLM-based reasoning.

Target customer

Future: Consumer automotive market (mass production EVs); Current: Automotive OEMs and Tier 1 suppliers (for AI technology/datasets)

Industries served

["Automotive / Electric Vehicles","Artificial Intelligence (Generative AI)","Mobility as a Service (MaaS)","Semiconductors (AI Compute)"]

Technology advantage

Proprietary 'Heron' model allows the vehicle to understand context (e.g., reading road signs, understanding police gestures) via language processing, enabling navigation in unmapped or complex urban environments without expensive LiDAR sensors.

How they differentiate

Turing differentiates through a 'camera-first' End-to-End (E2E) deep learning approach that eliminates LiDAR and HD maps. Unlike competitors using rule-based stacks, Turing utilizes proprietary Multimodal Generative AI ('Heron') to understand complex road contexts (e.g., gestures, signs) and a Generative World Model ('Terra') for realistic simulation.

Main competitors

["Tesla","Tier IV","Preferred Networks"]

Key partnerships

["Denso (Strategic Investor & Tier 1 Automotive Partner)","GMO Internet Group (Cloud/GPU Infrastructure)","NTT DATA / NTT Docomo Ventures (Data Infrastructure)","Renesas Electronics (Semiconductor/Edge AI)","Yanmar Ventures (Strategic Investment)"]

Notable customers

["Denso (Strategic Partner)","GMO Internet Group","NTT DATA"]

Major milestones

["Released 'Heron', a multimodal generative AI model for autonomous driving","Unveiled 'Terra', a generative world model for driving simulation","Secured ¥15.3 Billion Series A funding led by JIC VGI and Global Brain (Nov 2025)","Selected for METI/NEDO's 'GENIAC' AI development project"]

Growth metrics

Valuation reached approx. ¥60 Billion ($388M) post-Series A; conducting 'Tokyo30' project (30 minutes of intervention-free driving in Tokyo).

Market positioning

Deep-tech challenger aiming to be the 'Tesla of Japan' by achieving fully autonomous driving (Level 5) through generative AI rather than conventional sensor fusion.

Geographic focus

Japan (currently), with global expansion plans targeting North America

Patents and IP

Proprietary datasets (Japanese-Heron-Bench) and E2E driving architectures (Specific patent filings undisclosed in public search, but core IP lies in the 'Heron' and 'Terra' foundation models)

About Issei Yamamoto

Creator of 'Ponanza' (first Shogi AI to defeat a Meijin), former Lead Engineer at HEROZ, Inc., and Specially Appointed Associate Professor at Nagoya University.

Official website: