Gradium secures $100M seed round with Nvidia backing to advance ultra-low latency voice AI

Gradium Secures $100M Seed Round with Nvidia Investment

Gradium, a Paris-based AI voice startup, announced Thursday that it has raised $100 million in total seed funding after re-opening its round to new investors, including Nvidia 1

1

. The company originally launched out of stealth in December with $70 million from FirstMark Capital, Eurazeo, DST Global Partners, Eric Schmidt, and French telecom billionaire Xavier Niel 1

1

. The funding extension marks a significant milestone for the startup as voice AI reaches what co-founder and CEO Neil Zeghidour calls an "inflection point" 2

2

.

Founded in September and spun out of French AI lab Kyutai, Gradium was co-founded by Zeghidour, a researcher who previously worked at Google Brain, DeepMind, and Facebook 1

1

. The startup is developing ultra-low latency voice AI models that deliver voice at scale with near-instant response times, eliminating the awkward pauses that often plague AI agent conversations 1

1

.

Strategic Bay Area Expansion to Compete for AI Talent

The new capital will accelerate Gradium's AI research, product development, and international expansion, with a particular focus on establishing a new office in the San Francisco Bay Area 2

2

. This move represents a strategic shift for the company, positioning itself "at the heart of the world's leading AI ecosystem" to compete for talent alongside Anthropic, Google, Meta, and OpenAI 1

1

. While Paris remains a major European hub for AI, Gradium's decision to establish a presence in the Bay Area acknowledges the competitive advantages of proximity to the world's leading AI companies.

Rapid Product Development and Enterprise Adoption

Since its December launch, Gradium has expanded its capabilities significantly across speech generation, speech recognition, translation, and developer tooling 2

2

. The company has introduced a new generation of its flagship real-time text-to-speech model and launched several specialized products. These include Gradium Translate, an ultra-low-latency speech-to-speech translation model; Phonon, an on-device text-to-speech solution for edge devices; and Gradbot, an open-source framework for building production-ready voice agents 2

2

.

The startup has already secured enterprise customers across multiple sectors, including customer experience, healthcare, media, AI agents, and consumer applications 2

2

. Notable clients include French auto manufacturer Renault 1

1

.

Competing in a Crowded Voice AI Market

Gradium faces stiff competition from established players like ElevenLabs, which reached an $11 billion valuation in February, as well as major model makers such as Google Gemini 1

1

. However, the company's technical foundation may give it an edge. Zeghidour emphasized that while many businesses are building around voice AI, "developing very strong models for transcription, synthesis, the technological layer of AI, is very difficult. Only a few people in the world know how to do it properly. In our case, we have invented most of the technological steps and algorithms that are powering current technology" 2

2

.

As voice technology becomes foundational infrastructure for agentic commerce and the digital economy, Gradium's ability to deliver real-time speech-to-speech translation and ultra-low latency responses positions it to capture growing enterprise demand for seamless AI-powered voice interactions.