Gradium Announces its Launch with $70 Million in Seed Round Funding

Gradium, building the technological backbone, models, and infrastructure to support all voice applications, led by Olivier Teboul, Alexandre Défossez, Neil Zeghidour, and Laurent Mazaré, has announced its launch with a $70 million seed funding round led by FirstMark Capital and Eurazeo with participation from DST Global Partners, Eric Schmidt, Xavier Niel, Rodolphe Saadé, Korelya Capital, Amplify Partners, Liquid2, Drysdale Ventures, and angels including Yann LeCun, Olivier Pomel, Ilkka Paananen, Thomas Wolf, Guillermo Rauch, and Mehdi Ghissassi.

Voice is our most natural interface, yet today’s voice AI remains brittle, slow, and costly. Achieving effortless, human-like interaction is a true technological frontier, demanding deep expertise to produce speech that sounds organic, natural, and convincingly human.

Gradium builds foundational audio language models: an audio-native architecture that unifies generation, transcription, transformation, and dialogue. The platform delivers ultra-realistic, emotionally expressive speech with low latency, while remaining efficient and scalable, so high-quality voice becomes broadly accessible and affordable.

The ambition is to establish the technical backbone of global voice technology by eliminating the long-standing trade-offs between naturalness, speed, and cost, making realistic and interactive voice the default experience.

Gradium launches with real-time, multilingual transcription and synthesis (English, French, German, Spanish, Portuguese), available as standalone capabilities or combined through flexible plans for developers and enterprises. The platform already supports voice agents in healthcare, customer support, and market research, as well as NPCs in gaming and avatars in digital advertising.

Gradium is built by a team of pioneers in generative audio: Neil Zeghidour (Meta, Google DeepMind), Olivier Teboul (Google Brain), Laurent Mazaré (Google DeepMind, Jane Street), and Alexandre Défossez (Meta). This group collectively invented and open-sourced neural audio codecs and audio language models, enabling the first wave of voice cloning, text-to-music generation, and speech-to-speech translation. They later launched Kyutai, the non-profit research lab behind the first real-time conversational model released in 2024. The founding team is further strengthened by Constance Grisoni (BCG X) as Chief Growth Officer and Eugene Kharitonov (Meta, Google DeepMind) as Founding Scientist.

Gradium extends more than a decade of frontier research while maintaining close collaboration with Kyutai’s researchers and engineers. This tight link to foundational work in generative audio enables rapid translation of breakthrough research into reliable, production-grade products.

Within just three months of its inception, Gradium’s streaming transcription and synthesis APIs are already serving customers in production. Studios and game developers are deploying the technology to create immersive characters; language-learning platforms are integrating instant translation and natural-sounding voices; and healthcare innovators are testing conversational assistants designed to meet strict latency and privacy requirements. The platform accommodates both developers and enterprises, offering lightweight API access for rapid prototyping as well as robust, scalable deployments for production environments.