Loading organizations...
Gradium, based in Paris, France, develops audio language AI models for ultra-low latency voice interactions, enabling real-time, natural conversations between people and machines. The company launched out of stealth on December 2, 2025, with a $70 million seed round, co-led by FirstMark Capital and Eurazeo, and included DST Global Partners, Eric Schmidt, and Xavier Niel. Gradium provides APIs and foundation models supporting multiple languages, including English, French, German, Spanish, and Portuguese, building on technology from non-profit AI lab Kyutai, such as its Moshi conversational model and Hibiki speech-to-speech translation. These models power voice products for developers in applications like conversational assistants and real-time speech translation. Gradium was founded in September 2025 by Neil Zeghidour, Olivier Teboul, Alexandre Défossez, and Laurent Mazaré, a team of ex-Meta and Google DeepMind experts.
Gradium has raised $70.0M across 1 funding round.
Gradium's technology enables natural, expressive, and ultra-low latency voice interactions, supporting capabilities such as speech-to-text, text-to-speech, smart turn-taking, and instant voice processing.
Gradium provides its audio language models to customers through real-time, scalable voice APIs, allowing for integration into various modern applications and AI agents.
Gradium's technology enables natural, expressive, and ultra-low latency voice interactions, supporting capabilities such as speech-to-text, text-to-speech, smart turn-taking, and instant voice processing.
Gradium provides its audio language models to customers through real-time, scalable voice APIs, allowing for integration into various modern applications and AI agents.
Gradium is at the forefront of developing audio language models, positioning itself as a key player in the AI voice technology sector. The company's core mission revolves around enabling highly natural, expressive, and ultra-low latency voice interactions, crucial for the next generation of AI-powered applications and agents. By focusing on these attributes, Gradium addresses the growing demand for more human-like and responsive digital communication.
The company's technological foundation is built upon sophisticated speech-to-text and text-to-speech artificial intelligence models. These models are made available to clients through a robust API framework, allowing developers and businesses to integrate Gradium's advanced voice capabilities into their own products and services. This API-driven approach ensures scalability and ease of adoption across various industries.
Gradium's offerings extend beyond basic voice conversion, incorporating features such as smart turn-taking and instant voice processing. These functionalities are vital for creating seamless and intuitive conversational AI experiences, where the system can intelligently manage dialogue flow and respond without perceptible delay. The emphasis on high-fidelity models ensures clarity and naturalness in spoken output.
Operating from Paris, France, Gradium contributes to the European AI ecosystem by innovating in the field of foundation models for voice. The company's work supports a wide array of applications, from customer service bots to interactive virtual assistants, by providing the essential voice interface.
Gradium's commitment to developing scalable and high-performance voice solutions underscores its strategic importance in the evolving landscape of artificial intelligence. Its technology is designed to empower modern applications with a sophisticated and reliable voice layer, facilitating more engaging and efficient user interactions.
Gradium has raised $70.0M across 1 funding round. Most recently, it raised $70.0M Seed in December 2025.
| Date | Round | Lead Investors | Other Investors | Status |
|---|---|---|---|---|
| Dec 2, 2025 | $70M Seed | Eurazeo, FirstMark Capital | Guillermo Rauch, Eric Schmidt, Rodolphe Saad, Xavier Niel, Yann Lecun, DST Global | Announced |