Business

Nvidia Backs French Voice AI Startup Gradium at $100M Seed

PARIS: Gradium extends its seed round past $100 million, adding roughly $30 million with Nvidia as the headline new investor. The company is seven months old.

The July 9 extension tops up a $70 million first tranche from December, led by FirstMark Capital and Eurazeo. DST Global, Eric Schmidt, and French billionaires Xavier Niel and Rodolphe Saadé backed that opening round. Valuation remains undisclosed.

Also read: LiveKit Achieves $1B Valuation with $100M Series C for Real-Time AI Voice Platform

What justifies seed money at this scale is the founding team’s paper trail. Neil Zeghidour, Laurent Mazaré, Olivier Teboul, and Alexandre Défossez spun Gradium out of Kyutai, the Niel-backed Paris research lab. Between them, they built EnCodec, SoundStream, and Moshi: the foundational research that most modern voice AI runs on, including systems at OpenAI. The people who invented the plumbing now want the market.

The product is speed. Gradium ships speech-to-text, text-to-speech, and real-time translation models with latency around 200 milliseconds in practice. That kills the awkward pause that makes most AI phone conversations feel like talking to a satellite. Moshi, the underlying research, was the first real-time full-duplex spoken language model.

Traction came fast. The company launched from stealth in December with production-ready models and says it generated revenue within weeks. Renault is already a customer. The new money funds a San Francisco Bay Area office, planting the French team inside the talent pool it competes for.

Competition is not gentle: ElevenLabs at $11 billion, Deepgram, Mistral, and OpenAI’s own voice stack. Yet capital keeps flooding the category. Rime raised $24 million this week for enterprise calls, and OpenAI’s GPT-Live now reaches 150 million weekly voice users.

Voice, long the neglected AI interface, is suddenly the crowded one.

Anurag Shukla

Anurag Shukla is a Senior Journalist with over two decades of experience across television, digital, and print media. He has worked with leading national news organisations and has also served as a Research Officer in the Prime Minister’s Office (PMO), contributing to media research and policy-level content. A former journalism academic, Anurag brings strong editorial depth and a keen understanding of how technology, governance, and society intersect at Tea4Tech.

Recent Posts

Amid Backlash, Meta Goes Camera-Free With New AI Glasses

California: Meta introduced its first pair of camera-free AI glasses on Wednesday at its Connect…

4 weeks ago

Anthropic Brings Claude Chat and Cowork Together Into a Single Window

Bengaluru: Anthropic has combined Claude's chat and Cowork into one interface, so users no longer…

1 month ago

OpenAI Goes All In on Hardware, Buys Glass Imaging for $300M+

California: OpenAI has reportedly acquired Glass Imaging, a startup specializing in AI-enhanced smartphone camera technology,…

1 month ago

OpenAI Launches GPT-6 Astra With Advanced Computer and AI Capabilities

New Delhi: OpenAI has launched GPT-6 Astra, its latest and most advanced AI model, with…

1 month ago

India’s Rilo Joins Adobe in AI Marketing Automation Deal

New Delhi: Adobe has acquired Rilo, an India based marketing intelligence startup, in a deal…

2 months ago

Anthropic Levels Up Claude Fable 5.1 With More Power, Less Cost

New York: Anthropic has launched Claude Fable 5.1 and Claude Mythos 5.1, its latest AI…

2 months ago