Seven months is hardly enough time to settle into new office furniture, much less double your seed funding. But Gradium, the Paris voice AI startup, doesn't seem particularly interested in conventional timelines.
On July 8, the company confirmed it had extended its December seed round by adding Nvidia as a strategic investor—pushing the total to $100 million. Sifted reported the extension brought in roughly $30 million, though Gradium declined to confirm the exact figure or disclose its current valuation. For a company that only launched last December, that's the kind of momentum that makes venture capitalists sit up and pay attention.
The original $70 million seed—already among the largest in European AI history—was led by FirstMark Capital and Eurazeo, with DST Global Partners, Eric Schmidt, Xavier Niel, and Rodolphe Saadé joining the round. A who's who of AI luminaries signed on as angels: Yann LeCun, Datadog CEO Olivier Pomel, Supercell's Ilkka Paananen, Hugging Face's Thomas Wolf, Vercel founder Guillermo Rauch, and Mehdi Ghissassi.
Most startups would be planning their Series A pitch by now. Gradium just kept the seed round open.
Pedigree and Product Velocity
The founding team reads like a DeepMind and Google Brain alumni directory. CEO Neil Zeghidour spent years at both; CTO Olivier Teboul came from Google Brain; Laurent Mazaré, the Chief Coding Officer, worked at DeepMind and Jane Street; Chief Scientist Alexandre Défossez hails from Meta. They founded Gradium in September 2025 as the first commercial spin-out from Kyutai, the Paris-based nonprofit AI lab backed by approximately €300 million as of November 2023 from Niel, Saadé, and Schmidt's foundation.
And they've been moving fast. Between December and the Nvidia news, Gradium shipped an upgraded text-to-speech system, launched Gradium Translate—a real-time speech-to-speech translation service—and released Phonon, an on-device TTS model the company claims runs six times faster than real-time on a MacBook CPU. There's also Semantic VAD, a voice-agent turn-detection feature, and Gradbot, an open-source agent framework.
The company says it started generating revenue "within weeks of launch" and now serves enterprise customers across customer experience, healthcare, media, AI agents, and consumer applications. No names yet—though that reticence is hardly unusual for a young infrastructure play still sorting out partnerships.
Heading West
Gradium is using the new capital to open an office in the San Francisco Bay Area, planting a flag in the heart of the U.S. AI ecosystem. It's a logical move, if not exactly a groundbreaking one. Voice infrastructure has become a crowded battlefield: ElevenLabs is reportedly targeting a $22 billion valuation in a July secondary, and frontier model providers like OpenAI and Google are layering voice capabilities directly into their platforms. Then there's Cartesia, Inworld, and a dozen other contenders jockeying for position.
Nvidia's involvement may matter more than the cash itself. Strategic investors don't just write checks—they offer compute access, integration pathways, and a certain Silicon Valley credibility. For a European startup trying to crack the U.S. market, those advantages can be worth more than another line on the cap table.
Voice AI funding across Europe surged in the first half of 2026, with AI-native voice startups pulling in €536 million—up nearly 50% year-over-year, per Sifted. Gradium is riding that wave, but the company faces the central dilemma of all infrastructure providers: staying ahead of larger, better-funded competitors who can afford to play the long game.
The Road Ahead
Perhaps the biggest question isn't whether Gradium can sustain this velocity—it's whether $100 million in seed capital will be enough. Infrastructure businesses burn through capital quickly, and the company hasn't disclosed whether it's aiming for profitability or gearing up for a much larger Series A.
For now, the focus seems to be product expansion and customer acquisition. The technical chops are there. The funding is in place. What remains to be seen is whether Gradium can turn rapid-fire product releases into durable market position before the next wave of competition arrives.
And in voice AI, that wave is always just around the corner.
