HeyGen

HeyGen is an American artificial intelligence video company known for photorealistic digital-twin avatars and one-click video translation. Founded in 2020 by Joshua Xu and Wayne Liang, it grew into one of the largest AI video platforms for business communication, reporting more than US$200 million in annual recurring revenue, 30 million users and customers in 196 countries by mid-2026.^[1][2][3]

History

HeyGen was founded in 2020 by Joshua Xu, a former Snapchat engineer, and Wayne Liang. The service launched under earlier names before the company rebranded as HeyGen and established its headquarters in Los Angeles in 2022, initially backed by investors including Sequoia China (later HongShan) and ZhenFund; a US$5.6 million round from Sarah Guo's firm Conviction followed in November 2023. A widely shared demonstration that translated a speaker's video into another language while matching their lip movements helped popularise the product internationally.^[1]

In June 2024, HeyGen raised a US$60 million Series A led by Benchmark at a US$500 million valuation, with participation from Thrive Capital, BOND and Conviction, disclosing more than 40,000 customers and about US$35 million in annualised revenue at the time. Growth continued with unusual capital efficiency: annual recurring revenue climbed to roughly US$57.5 million by the end of 2024 and US$123 million by the end of 2025, passing US$200 million by June 2026 — which the company and its investors described as the rise of identity-first AI video, in which content is fronted by a real person's licensed likeness rather than a generic synthetic presenter. HeyGen says it reached that scale having raised only around US$70 million in total and while remaining cash-flow break-even.^[2][4][5]

Technology

HeyGen builds its own full stack, from avatar models to the editing application. Its products include Avatar IV, which turns a single photograph into a talking, gesturing video; digital twins trained from short recordings of a user's appearance and voice; video translation across more than 175 languages and dialects with lip-sync and voice preservation; and an enterprise API. The company says it can train effective avatar models using a fraction of the video data available to general-purpose systems because it focuses narrowly on one person delivering a message, and it markets consent and identity-verification controls as core features.^[2][5][6]

Applications and impact

The platform is used for corporate training, sales outreach, e-learning and social media content, with examples including creators who continued publishing through illness by using avatars, and firms that produce training and marketing video without studios. The rapid spread of avatar video has also placed HeyGen in a wider debate over deepfakes, likeness rights and consent: analysts and regulators increasingly expect synthetic video to be labelled and for cloned identities to be authorised, and platforms offering digital twins face scrutiny over misuse.^[3][5]

>See Also

• ElevenLabs
• Voice Cloning
• Text-to-Speech
• AI Video Generation
• Digital Twin