Gradium, a rising player in voice artificial intelligence, has boosted its seed funding to $100 million by bringing Nvidia on board as a new investor. This capital injection will support the company’s expansion efforts, boost research and product development, and establish a new office in the San Francisco Bay Area.
Founded recently in September, Gradium has quickly broadened its technology portfolio, enhancing capabilities in speech generation, recognition, and translation. The company introduced an innovative real-time text-to-speech model, alongside an ultra-low-latency speech-to-speech translation tool named Gradium Translate. It also launched Phonon, an on-device text-to-speech model designed for edge devices, and Gradbot, an open-source framework aimed at building production-ready voice agents.
Gradium’s voice AI solutions have attracted enterprise clients spanning customer experience, healthcare, media, AI agents, and consumer applications. The co-founder described this funding milestone as pivotal, enabling the company to fast-track its roadmap and extend its reach to developers and businesses worldwide.
With founders formerly affiliated with DeepMind and Meta, Gradium leverages proprietary algorithms developed over several years, which underpin much of today’s voice technology. The company emphasizes the technical challenges involved in developing top-tier models for transcription and synthesis, areas in which only a handful of experts excel.
Gradium’s progress reflects a broader trend toward voice technology becoming a critical infrastructure in agentic commerce and the digital economy. Investments by major tech companies and the rise of voice-focused startups highlight the growing importance of voice AI across multiple sectors.