# High-Level Overview
Kotoba Technologies is a cross-border AI startup developing speech and language AI systems specifically optimized for Japanese and Asian languages.[1] The company builds real-time speech translation, voice generation, and automatic speech recognition (ASR) technologies that address a critical gap in generative AI adoption for non-English languages. Founded in October 2023, Kotoba serves enterprises, developers, and consumers through both mobile applications and enterprise APIs, with its DOTSU iOS app achieving over 500,000 sessions within three months of launch in March 2025.[3]
The company's mission centers on democratizing advanced AI capabilities across language barriers. Rather than adapting English-first models to Asian languages, Kotoba develops language-centric foundation models that deliver superior performance for Japanese, Korean, Chinese, and Spanish.[3] This approach positions the startup at the intersection of two powerful trends: the global expansion of generative AI beyond English-speaking markets and the rising demand for real-time multilingual communication in business and entertainment.
Origin Story
Kotoba Technologies was co-founded in October 2023 by Dr. Hiroyuki "Nori" Kojima (CEO) and Dr. Jungo Kasai (CTO), both former Meta employees with PhDs in Computer Science from Cornell University.[4] Kojima earned the Best Paper Award at EMNLP 2022, one of AI's leading conferences, and co-initiated the "Fugaku-LLM" project, which leveraged Japan's Fugaku supercomputer—then the world's second-fastest—to develop Japanese-language large language models.[1][4]
The founding insight emerged from observing a critical market gap: while generative AI was rapidly advancing in English, non-English languages lagged significantly in adoption and capability. Rather than starting from scratch, the founders built on the momentum of the Fugaku-LLM project, assembling top researchers and engineers to bridge this divide.[1] The company's early traction was substantial—it was selected twice for Japan's government-backed GENIAC generative AI project, which provides access to large-scale computational resources.[3] By 2024, Kotoba released the beta version of Kotoba SpeechGen, its flagship voice generation technology using preset voices and voice cloning.[3]
Core Differentiators
- Language-First Architecture: Unlike English-centric models adapted to other languages, Kotoba builds foundation models optimized natively for Japanese and Asian languages, delivering superior fluency and naturalness.[1][2]
- Real-Time Speech Translation: The company's simultaneous translation technology goes beyond transcription to enable true real-time interpretation with voice cloning and emotion rendering capabilities.[1]
- Rapid Commercialization: The DOTSU app demonstrated product-market fit, accumulating 500,000 sessions in just three months, signaling strong consumer demand.[3] Enterprise API rollout began in late July 2025 with ASR and voice agent components.[3]
- Supercomputer-Grade Infrastructure: Access to Fugaku and participation in Japan's GENIAC program provide computational advantages that competitors without similar resources cannot easily replicate.[1][4]
- Dual-Market Positioning: The company operates across both consumer (mobile apps) and enterprise (APIs for business meetings, entertainment, events) segments, diversifying revenue streams.[3]
Role in the Broader Tech Landscape
Kotoba Technologies operates at a critical inflection point in AI's global expansion. The generative AI boom has been overwhelmingly English-centric, leaving massive markets—particularly in Asia—underserved by high-quality AI tools. Japan alone represents a $5+ trillion economy where language barriers have historically limited AI adoption. Kotoba's timing is optimal: enterprise demand for multilingual communication is accelerating as companies globalize, while consumer appetite for seamless translation is evident from the DOTSU app's rapid adoption.
The startup also benefits from Japan's strategic push to develop indigenous AI capabilities. Government backing through GENIAC and partnerships with Japanese universities and companies position Kotoba as a national champion in speech AI infrastructure.[3][4] This creates a virtuous cycle: government support attracts talent and capital, which accelerates product development, which in turn strengthens Japan's AI ecosystem.
Beyond Japan, Kotoba's expansion roadmap—adding Chinese, Korean, and Spanish support—targets the world's largest non-English-speaking populations.[3] This positions the company to influence how generative AI develops globally, potentially establishing new standards for language-specific model optimization that other startups and incumbents will need to match.
Quick Take & Future Outlook
Kotoba Technologies is well-positioned to become the dominant speech AI platform for Asia and a blueprint for how non-English AI infrastructure should be built. The company's $11.83 million Series Seed 2 funding round, led by Globis Capital Partners and Boost Capital, provides runway to accelerate the transition from R&D to commercialization.[3][5] Immediate priorities include hiring machine-learning and application engineers to advance model development and scaling the enterprise API business.
The next 18-24 months will be critical. Success hinges on converting early DOTSU momentum into sustained enterprise adoption—particularly in remote business meetings, where real-time translation delivers immediate ROI. Expansion into entertainment and events represents a significant upside opportunity. If Kotoba can establish itself as the default speech AI platform for Asian languages while maintaining technical superiority over adapted English models, it could capture outsized value in a market that has historically been overlooked by Silicon Valley incumbents. The company's ability to leverage Japan's computational infrastructure and government support while maintaining startup agility will ultimately determine whether it becomes a category leader or a strong regional player acquired by a larger AI platform.