Gradium

Artificial Intelligence · Speech & Audio AI
Lately:16 news stories in 90 daysFollow changes →

Facts

as of Nov 20, 2026

Gradium builds ultra low latency voice AI models for developers. Generate expressive text to speech, accurate speech to text, and high fidelity voice clones with a single API.

Employees
51-200estimated range
Website
gradium.ai

These are the key facts about Gradium on Alion as of Nov 20, 2026. It has raised $170M in 2 funding rounds. The latest round is Seed of $100M, announced in 2026.

Sources: company registers, GLEIF, Wikidata, the company's site, Alion job postingsReport a mistake

What the registers and the postings say about this employer.

Sanctions lists
OFAC SDN, EU consolidated · sanctionssearch.ofac.treas.gov ↗
No match
Phishing reports
OpenPhish, Phishing.Database · openphish.com ↗
None
Layoff notices
US WARN notices on file · dol.gov ↗
None
Full employer checkScam signs in adverts, mail records, visa registers, domain history
Sources: company registers, the Alion truth index, OFAC and EU sanctions lists, US WARN notices

Financials

Funding rounds, investors and revenue on record.

$170MTotal raised
SeedLatest round
2Rounds and grants
Money by yearUSD, rounds and grants
2026$100M1 record
2025$70M1 record
Largest rounds and grants
2026
$100M
2025
$70M
Sources: funding announcements, grant and company registers

Money

Gradium has $170M paid in by investors, lenders and donors.

$170MPaid in
StagePaidSignedAnnouncedSourceFiledAnnouncedPress
Money in
Equity raised · NVIDIA
Jul 2026 · Seed
$100MPaid
Equity raised
Dec 2025 · Seed
$70MPaid
Taxes
No public tax figures for this company or its entities.
Every row links its own source. A flow two sources state counts once; a figure a source does not state reads "not stated".

Owners & entities

Sources name NVIDIA among the owners or investors of Gradium.

Named investors, no share stated1
Holders and shares as each register or filing states them.

Milestones

Each milestone is dated by its source: registers, filings, the press

Newsroom

top 0.38%Web presenceof 133.2M domains by links, Common Crawl
722 ★GitHub stars12 public repositories · +146%
Sep 2026Launched on Hacker News1 points
#655,199Web rankTranco top 1M
Oct 262026
Blog
Gradium's voice models now run natively inside Yampa, the agentic platform enterprises use to build, deploy and supervise customer care agents, in SaaS on EU infrastructure or on…
Oct 212026
Blog
Trained on NVIDIA GPUs, Phonon generates natural speech locally in as little as 64.5 MB, on phones, laptops, embedded hardware and in the browser.
Oct 202026
News · Gradium
Phonon is Gradium's family of open-weight text-to-speech models that run fully on-device: 31m and 40m parameters, five languages, offline, faster than real time on a Raspberry Pi…
Oct 202026
Blog
Today, we're open-sourcing Phonon, now a family of 31m and 40m parameter text-to-speech (TTS) models that set a new standard for naturalness and accuracy on-device.
Oct 192026
Blog
Eden AI is an AI gateway: one API for 500+ models across LLMs, speech, vision and translation, with routing, fallbacks and a dedicated EU endpoint for European data residency. Gr…
Sources: the outlets named on each headline, each story once