Joshua Xu

Joshua Xu is co-founder and CEO of HeyGen, the AI video platform that generates talking-head and avatar videos for marketing, training, and localization. He came out of Snap's advertising and AI camera teams convinced that the camera itself was becoming optional — and that businesses still drown in production cost when they try to live in a video-first world.

Snap ads and the AI-augmented camera

At Snap he integrated AI into advertising, then moved to the AI-augmented camera around 2018, when generative video still felt fringe. In a Unite.AI interview he recalled realizing "the computer can create high-quality and realistic videos" and getting excited about changing how people make content. Mobile cameras had unlocked Instagram and TikTok-style creation, but on-camera skill, recording time, and production cost still blocked first-class video.

Founding HeyGen to replace the camera

He co-founded HeyGen in 2020 on a blunt product thesis: "At HeyGen, we believe that the camera is replaceable." If you can remove the camera, you remove a barrier to visual storytelling. The company builds studio-quality videos with AI avatars and voices so teams skip crews, sets, and reshoots. Xu's commercial case is speed and overhead: nix the film crew and "HeyGen users create videos 10x faster."

Create, localize, personalize

Public customers span marketing teams that need volume, plus high-profile localization experiments — McDonald's Sweet Connections messages in a grandmother's language, Weather Kids forecasts with UNDP and The Weather Channel. Text-to-speech and lip-sync across many languages sit at the center of that localization story. Trust and safety work, including live consent for avatars, became part of the product narrative as deepfake anxiety rose.

On video without a crew

Looking ahead he talks about full-body avatars and personalized video at scale, and about generative editing assembling polished cuts without traditional production. Film disruption, in his telling, includes localization without reshoots. The work is business video treated as infrastructure: create, localize, and personalize without hiring a studio every time.