• Mujin

  Participated · Debt Financing · Dec 2025

  Founded in 2011, Tokyo-based Mujin provides MujinOS, a software platform that unifies robot vision, motion planning and real-time digital twin technology to automate complex manufacturing and logistics tasks. The system controls mixed fleets of robotic arms, AGVs and storage robots under a single architecture, eliminating custom system integration and accelerating deployment. Current applications span depalletizing, palletizing, piece picking, truck unloading, warehouse execution and fleet management. With the new funding, the company will pivot from project-based integration to a more scalable product-led model centered on MujinOS, expand its product lineup and accelerate global rollout in Europe and North America. Resources will also go toward enhancing plant- and warehouse-level digital twins and hiring additional world-class engineers to advance next-generation physical-AI capabilities. Including the latest round, Mujin has raised a cumulative ¥596 billion (≈US$4 billion) in equity and debt financing.

• Craif

  Participated · Series C · Apr 2025

  Craif develops a urine analysis platform that combines proprietary biomarker detection and AI to assess disease risk, with cancer as its initial focus. Its flagship test, miSignal, is available at more than 2,500 medical institutions and 4,500 pharmacies in Japan, and the company has performed over 110,000 tests to date. Craif analyzes microRNA signals and has published more than 90 peer-reviewed papers and conference presentations, including two independent publications on early pancreatic cancer detection. The company is advancing a software-based medical device program targeting pancreatic cancer and plans a prospective clinical study in the U.S. from its San Diego laboratory. Craif opened its San Diego lab in April 2026, established a U.S. subsidiary in 2022, and is building clinical, regulatory, and commercial capabilities to support a reimbursement-driven U.S. launch.

• Elsa

  Participated · Series C · Sep 2023

  ELSA, founded in 2016, offers speaking lessons with personalized learning plans, test and interview preparation, and a speech analyzer to help users improve conversational skills. The company recently launched ELSA AI Tutor, a voice-based generative-AI tutor that role-plays conversations and uses speech recognition to give feedback on pronunciation, grammar, cadence and professional tone. The AI Tutor remembers users’ past performance to customize lessons and is aimed at learners who lack access to live tutors. ELSA has offices in San Francisco, Lisbon and Ho Chi Minh City. The company said it will use the new funding to support and grow its platform, expand further in Taiwan, enter the Middle East and Turkey, build on success in Japan, and increase its focus on B2B. To date ELSA has raised $60 million. ELSA is a pronunciation-focused English learning app that uses proprietary speech-recognition AI trained on accented English from millions of users to provide detailed feedback on sounds, intonation, rhythm and fluency. The app offers fine-grained correction of individual sounds and advanced prosodic coaching to help learners speak more naturally. ELSA claims more than 13 million users and was founded in 2015; it is based in San Francisco and Ho Chi Minh City. The company has partnerships with IDP (which now recommends ELSA for IELTS test preparation), language schools in Vietnam (IMAP and Speak Up), online platform YOLA, and corporate clients including Kimberly Clark, Intel and ATAD. Going forward, ELSA plans to expand operations in Latin America, Vietnam, India and Japan and to build a scalable B2B platform to offer its coaching services to companies and educational organizations. Elsa provides an AI and voice-recognition app that grades users' English speaking and pronunciation against standard American English and gives detailed feedback. The service is freemium, offering full access to 1,000 courses for roughly $3–6 per month, with new content added weekly. Elsa reports users in more than 100 countries and two million paying users in its top market, Vietnam. The company supports localized payments and a local sales team in Vietnam and has opened an office in Tokyo as part of expansion into Japan, Indonesia and India. Elsa is preparing to support additional languages (Spanish announced as next) by building voice-recognition models and collecting voice data to train them. It is also pursuing the education market by licensing the app to schools (subsidized per‑student fees) and allowing teachers to upload curriculum and track students' oral homework. ELSA provides an AI- and speech-recognition-powered app that gives personalized pronunciation feedback and coaching to English learners. Users take a five-minute assessment that builds a tailored curriculum from roughly 600 lessons and 3,000+ words; lessons are bite-sized (about two minutes) and include five progressively harder exercises. The app listens to learners, compares speech to American English targets, highlights performance (red/yellow/green) and offers corrective tips on mouth and tongue placement. Since launch the product has been adopted by a few million users in over 100 countries (about half in Southeast Asia) who practice a few million exercises per week. ELSA generates revenue via subscriptions (about $3.99/month or $29.99/year, with a quarterly option and regional price variance). Future plans include building a teacher tool to import classroom curriculum, adding new languages, developing a voice-recognition API, and hiring additional AI talent.

No current team members are available.