Gradium, a Paris-based startup building ultra-low-latency voice AI models, has raised $100 million in total seed funding after reopening its round to new investors including Nvidia, TechCrunch reported on July 9, 2026. The funding positions the company as a serious challenger in the increasingly competitive voice AI market.
The startup originally launched out of stealth in December 2025 with $70 million from a roster of prominent investors. The reopened round added Nvidia as a strategic backer, bringing the total to $100 million. Gradium said it will use the capital to open an office in the San Francisco Bay Area, strengthening its position at the heart of the world's leading AI ecosystem. For the latest AI developments, Gradium's transatlantic expansion reflects a broader pattern of European AI startups establishing U.S. presences to compete for talent.
Spun Out of Kyutai
Gradium was spun out of Kyutai, a French AI research lab backed by telecom billionaire Xavier Niel. Both Kyutai and Gradium were co-founded by Neil Zeghidour, a researcher who previously worked at Google Brain, DeepMind, and Facebook before establishing the lab.
The connection to Kyutai gives Gradium a deep research foundation. Kyutai has built a reputation for open-source AI research in Europe, and the spinout model — where a research lab commercializes its work through a dedicated startup — mirrors successful patterns seen at institutions like DeepMind, which spawned multiple ventures after its acquisition by Google.
Zeghidour's background across three of the world's leading AI labs informs Gradium's technical approach. The company is focused on audio models that deliver voice at scale with ultra-low latency — meaning AI voices that respond almost instantly, without the awkward pause that often characterizes AI agent conversations.
A Competitive Landscape
Gradium enters a crowded market. ElevenLabs, the dominant player in AI voice generation, was valued at $11 billion in February 2026. Google's Gemini platform also offers voice capabilities integrated into its broader AI assistant ecosystem. Other startups, including well-funded competitors in the United States and Asia, are racing to build voice models that can power everything from customer service agents to entertainment applications.
Despite the competition, Gradium has already landed notable customers. Since its December launch, the company says it has signed on French auto manufacturer Renault, among other enterprise clients. The Renault deal suggests that Gradium's technology is mature enough for production deployment in automotive and industrial settings, where voice interfaces are increasingly important.
Gradium's emphasis on ultra-low latency addresses a specific pain point in the voice AI market. Current AI voice systems often suffer from noticeable delays between a user's speech and the AI's response, which disrupts natural conversation flow. Gradium's models are designed to minimize this latency, potentially enabling more fluid human-AI voice interactions for applications like real-time translation, conversational agents, and in-vehicle assistants.
The Nvidia Factor
Nvidia's participation in the round is strategically significant. As the dominant supplier of GPUs for AI training and inference, Nvidia has positioned itself as a kingmaker in the AI startup ecosystem through its Nvidia Inception program and direct venture investments.
For Gradium, Nvidia's backing provides more than capital. Access to Nvidia's latest GPU architectures, developer tools, and optimization frameworks could give the startup a technical edge in training and deploying voice models at scale. Nvidia's involvement also signals confidence in Gradium's approach to ultra-low-latency voice AI — a technology that could become increasingly important as AI agents move from text-based interfaces to voice-first interactions.
Paris to Silicon Valley
Gradium's decision to open a Bay Area office is notable for a company rooted in the Paris AI ecosystem. Paris has emerged as a major European AI hub, home to Mistral AI, Kyutai, and other prominent labs and startups. But the company acknowledged the benefits of proximity to Anthropic, Google, Meta, and OpenAI — all headquartered in the San Francisco Bay Area.
The move reflects a reality of the global AI talent market. While European cities have produced world-class AI research, many startups find that competing for top engineering talent requires a presence in Silicon Valley, where the concentration of AI companies, researchers, and venture capital is unmatched.
Gradium's existing investor roster reads like a who's who of tech and venture capital. FirstMark Capital, Eurazeo, DST Global Partners, Eric Schmidt, and Xavier Niel all backed the initial round. The addition of Nvidia in the reopened round adds a strategic compute partner to a financial base that already spans U.S. and European capital.
Stay Ahead of AI
For more coverage of AI startup funding, voice technology, and the companies shaping the future of artificial intelligence, follow AI Buzz Wire for the stories that matter.
Read more AI startup news →



