The AI avatar-video leader whose revenue jumped 82% in 2024 on filed UK accounts, on the way to $100M+ ARR and a $4B valuation — here is what's driving the growth and how a creator uses its output.
Last verified · 2026-10-07 · by Moe Ameen
Synthesia is the AI avatar-video platform that turns a typed script into a narrated talking-head video — no camera, no studio. What makes it worth understanding in 2026 is not just the product but the trajectory: its filed UK accounts show turnover of about £58 million for 2024, up 82% from £32 million the year before, and it has since crossed $100 million in annual recurring revenue, reaching roughly $146 million ARR by September 2025. In late January 2026 it raised a $200 million Series E at a $4 billion valuation, led by GV with Nvidia's NVentures, Accel, Kleiner Perkins, and NEA participating — about double its $2.1 billion Series D a year earlier.
The single clearest driver of that growth is localization at enterprise scale. Synthesia's signature is taking one script and regenerating it across 140+ languages with 1-click translation and AI dubbing that re-syncs the avatar's lips to each new-language audio track — so one explainer ships as dozens of localized versions without a reshoot. Pair that with hundreds of stock avatars, cloned "personal" avatars built from a short recording, PowerPoint-to-video, and enterprise governance (SOC 2, GDPR, SSO), and you have exactly what a global L&D or comms team wants to buy. Net revenue retention above 140% and penetration of more than 90% of the Fortune 100 confirm where the demand sits.
Founded in London in 2017 by Victor Riparbelli, Steffen Tjerrild, and professors Lourdes Agapito and Matthias Niessner, Synthesia has layered on interactive avatars that can listen and respond. But the honest framing for a creator is unchanged by the valuation: Synthesia makes one thing — avatar video — meters it by the minute, and stops at the exported file. The growth story is a market signal, not a distribution tool. It renders a localized MP4; it does not clip, caption for feeds, reframe per platform, diversify into other formats, or publish anything.
Read the growth story for what it tells you to copy: Synthesia's revenue is compounding because it turns one script into a wall of localized, lip-synced versions, and markets pay to be reached in their own language. You can borrow that exact play without the enterprise contract — but only if you solve the thing a Synthesia export leaves open. Each localized render lands as a standalone MP4, in one aspect ratio, metered against a monthly minute cap, with nothing built to turn it into a presence in that market. [Kompozy](/) is where those exports become a published, per-market feed. Drop a Synthesia video in and Kompozy clips it into vertical [Clipped Shorts](/), reframes it to 9:16, 1:1, and 16:9 for each platform, and burns in feed-native captions in the matching language — the finishing and distribution layer Synthesia stops short of.
The deeper leverage is that Kompozy also generates the avatar video itself, so you are not tied to a minute meter at all. A face-locked HeyGen-based [Persona Short](/glossary/persona-shorts) narrated in a target language, held to your voice by the [Persona Brief](/glossary/persona-brief), becomes the seed for a full local content set — a brand-exact Carousel, a Quote Graphic, a Photo Post, a Blog Article, and an Email Newsletter from the same idea. Then it ships: schedule and publish each market's set across eight social platforms plus blog and email from one queue, behind a per-post review. Synthesia proved localization is the growth engine; Kompozy is how a creator runs that engine as a published feed in every market instead of a folder of exported files.
Its filed UK accounts show 2024 revenue up 82% year over year, and it has since passed $100 million ARR, reaching about $146 million by September 2025. The driver is enterprise localization: one script becomes dozens of lip-synced language versions, which global L&D and comms teams buy at scale, and existing customers expand spend year over year (net revenue retention above 140%).
Synthesia raised a $200 million Series E in late January 2026 at a $4 billion valuation, led by GV with Nvidia's NVentures, Accel, Kleiner Perkins, and NEA — roughly double its $2.1 billion Series D a year earlier.
No. Synthesia renders and exports a video file, metered by video minutes. It has no scheduler and no social publishing — you distribute the file with other tools. Kompozy takes a Synthesia export and clips, captions, reframes, schedules, and publishes it across nine platforms, and can generate the avatar video itself.
Yes. Export Synthesia's localized versions and Kompozy will clip, reframe, and caption each for the feed in the matching language — or generate a face-locked Persona Short in the target language directly. Either way it then publishes across nine platforms, so localization becomes a per-market published presence rather than a stack of MP4s.