Loading
Gradium is a Paris-based artificial intelligence startup that specializes in audio language models for natural and expressive voice interactions. Founded in September 2025, the company emerged from stealth mode in December 2025 after securing a $70 million seed round. Gradium aims to create the "voice layer" of AI, focusing on ultra-low-latency voice applications with a benchmark of sub-300ms time-to-first-audio for streaming text-to-speech. The company offers a suite of B2B API services, including ultra-low-latency text-to-speech, real-time speech recognition, voice cloning, live speech-to-speech translation, and dialogue generation. Gradium's technology is built on audio-native language models, allowing for efficient and scalable voice processing. With multilingual support from the start, Gradium serves various industries, including gaming, healthcare, and customer service, providing innovative solutions for real-time voice interactions.