Gradium raises $100M seed, Nvidia backs Paris AI voice startup
Gradium reopened its seed, reached $100 million with Nvidia among new investors.
TL;DR
- 01Gradium reopened its seed, reached $100 million with Nvidia among new investors.
- 02Gradium reopened its seed round and has now raised $100 million in total, with Nvidia among the new investors, the Paris-based voice AI startup said on Thursday.
- 03The company plans to use the cash to open an office in the Bay Area and compete for talent there, while continuing development of its low-latency audio models.
Gradium reopened its seed round and has now raised $100 million in total, with Nvidia among the new investors, the Paris-based voice AI startup said on Thursday. The company plans to use the cash to open an office in the Bay Area and compete for talent there, while continuing development of its low-latency audio models.
What did Gradium announce?
Gradium said it reopened its seed and has now raised $100 million, adding Nvidia to a roster of investors. The company originally launched out of stealth in December with $70 million from backers including FirstMark Capital, Eurazeo, DST Global Partners, Eric Schmidt, and French telecom billionaire Xavier Niel.
Gradium was spun out of French AI lab Kyutai, a lab backed by Xavier Niel, and was co-founded by Neil Zeghidour, a researcher who previously worked at Google Brain, DeepMind, and Facebook. The startup builds audio models designed to deliver voice at scale with ultra-low latency, aiming to remove the pause that can interrupt AI agent conversations. Since its December launch, Gradium says it has landed some customers, including French auto manufacturer Renault.
How does Gradium compare to rivals?
Gradium competes directly with specialist voice startups and larger model makers where speech is a feature. Competitors named in the announcement include ElevenLabs, which the article notes was valued at $11 billion in February, and major model makers such as Google with its Gemini models. Gradium positions its offering around "ultra-low latency" audio that responds almost instantly, targeting use cases where conversational lag degrades the user experience.
The startup’s investor list and its founder’s pedigree — Zeghidour’s background at Google Brain, DeepMind, and Facebook — are presented as part of why it has attracted customers quickly. Gradium also signals a talent strategy shift by opening a Bay Area office to place itself close to firms like Anthropic, Google, Meta, and OpenAI, which the company described as part of the world’s leading AI ecosystem.
Why it matters
Gradium’s $100 million seed shows investor appetite for specialized multimodal and voice-first AI companies, and the round underlines that large seed checks remain available for teams with strong research roots and early commercial traction. Moving to open a Bay Area office matters because it acknowledges the perceived recruitment and partnership advantages of proximity to major U.S. AI labs and platforms. The combination of capital, customer wins such as Renault, and ties to an AI research lab give Gradium a clearer runway to scale voice models and compete with both startups and incumbent model makers.
What to watch
Watch whether Gradium’s Bay Area office becomes a hub for U.S. customer wins and hires, and whether the company announces additional commercial partnerships beyond Renault. Also watch for product benchmarks or latency claims that quantify the "ultra-low latency" performance Gradium emphasizes, and for any future funding rounds or strategic ties between Gradium and its new investor Nvidia.
Written by The Brieftide · Source: TechCrunch
The Brieftide Daily · 06:00
Briefs like this one, in your inbox every morning.
Continue reading
More in Open Source AIOpenAI GPT-5.6, Anthropic models: U.S. government control
Two weeks after the government pulled Anthropic’s Fable and Mythos, GPT 5.6 will ship in a limited preview approved "customer by customer".
OpenAI joins Appia Foundation to build shared AI standards
OpenAI supports evaluation frameworks, safety practices and global cooperation through the Appia Foundation.
Zhipu AI GLM-5.2: 1M-token context, closes gap with Opus 4.8
GLM-5.2 ships under the MIT license with a stable one-million-token context and scores 74.4% on FrontierSWE, one point behind Opus 4.8.
OpenAI: PRC-linked influence operations target US AI debates
OpenAI says PRC-linked campaigns are using AI to push narratives on U.S. tech debates, data centers, tariffs and false ChatGPT claims.