Startup profile for Gradium: latest funding, latest known valuation, total disclosed funding, investors, status, sources, and why the company is interesting. Part of the Venture Capital Tracker startup directory.

Startup profile · funding coverage

Gradium funding, valuation and investors

Paris-based low-latency speech AI developer spun out of Kyutai research for real-time voice agents.

Verified company research · funding, investors and latest developments

Keep track of Gradium

Save this profile to your VCT watchlist for a quick return.

View watchlist

Funding, valuation & investors

Answer-first snapshot

Latest funding

Seed extension; cumulative total $100M

Incremental amount not separately disclosed; $100M total seed including original $70M · July 8, 2026

Latest known valuation

Not publicly disclosed

Total disclosed equity funding

$170M

Excludes debt, grants, acquisitions, secondaries, and IPO proceeds.

Current status

private

seed · Paris, France / San Francisco Bay Area, United States

Investors in latest funding

Other: NVIDIA

Sources: latest funding · gradium.ai Last verified: 2026-11-13

Overview

Gradium was founded in 2025 by researchers associated with the Kyutai speech AI laboratory. It builds real-time speech generation and speech-to-text models and serves agentic audio APIs. The company announced a $70M seed on December 2, 2025, co-led by FirstMark Capital and Eurazeo, with participation from DST Global Partners and other investors.

On July 8, 2026 it announced an extension of the seed to $100M cumulative, with NVIDIA participating; the exact incremental amount was not separately stated in the company release.

Why Gradium is interesting

A clear commercial use of efficient model inference in the voice category, where latency and interruption quality are measurable. Its 2026 $100M figure is cumulative seed capital, not a new $100M round.

Product & use cases

Low-latency multilingual speech foundation models, voice agents, transcription and audio streaming APIs.

  • Interactive low-delay audio assistants
  • Speech generation and transcription at scale
  • Voice interfaces that need fast turn taking

Key facts

  • $70M seed announced December 2, 2025
  • July 2026 extension brings cumulative seed capital to $100M, including initial $70M
  • NVIDIA joined the extension; no separate $100M new round announced

Funding history (newest first)

Seed extension; cumulative total $100M

2026-07-08 Incremental amount not separately disclosed; $100M total seed including original $70M

Source: Read source · gradium.ai

Seed

2025-12-02 $70M

Source: Read source · gradium.ai

Competitive landscape

Edge: Kyutai research lineage and specialist audio efficiency focus; voice latency is more commercially testable than claims of general intelligence.

Competes with Cartesia and other speech API providers. Better streaming latency does not imply cheaper training of general-purpose text models. July 2026 company statement should not be double counted as a separate $100M raise.

Market / IPO context

Editorial / static context — not a live quote.

FAQs about Gradium

Practical answers founders, operators, and investors typically search for.

No. Gradium announced a $70M seed in December 2025 and later said its seed financing had reached $100M cumulatively, including the original amount.
No. It creates low-latency voice models and APIs and is adjacent to, not a developer of, Stanford NCM.

By Venture Capital Tracker

Last updated:

Editorial note: AI tools assisted with research, structure, or drafting. Venture Capital Tracker retains human editorial responsibility for factual accuracy, relevance, and source quality before publication.