Synthesia co-founder Victor Riparbelli just raised $180 million ($2.1 billion) to make his startup’s AI-generated video avatars more lifelike.
synthesis
While OpenAI, Google and Adobe compete with startups like Runway to develop AI video tools for Hollywood use, a British startup is taking a totally different approach. The goal is to assist firms create training videos that individuals actually watch.
That’s a boring but lucrative area of interest for Synthesia, which just raised $180 million in a round led by NEA that valued the London-based startup at over $2.1 billion. Chemical giant DuPont, printer maker Xerox and airline Spirit now use Synthesia’s avatars to deliver safety briefings and other training videos in over 20 languages with only one click.
“People are now taking text and slide content and turning it into video,” says Victor Riparbelli, CEO and co-founder of Synthesia. “You can think of it like Powerpoint 2.0 with the same user type and use case.”
Synthesia was reported generated sales of over $70 million last 12 months, greater than doubling its revenue $31.3 million in 2023, company filings show. With its $180 million in funding and backing from latest investors Atlassian Ventures and billionaire Penny Pritzker fund PSP Growth, Synthesia plans to speculate in improving its avatars and in tools that make creating and hosting videos easier .
Riparbelli says a few of its customers are already using Synthesia avatars to market videos and a few Tiktok creators are using them to provide videos, however the technology is not able to move into promoting or content production yet. “We have some major models training now that I think we’ll probably get over that threshold within the next three to six months,” he says.
That could put Synthesia in competition with AI giants like OpenAI and its Sora video generator and startup Runway. “We don’t care about AI video in the sense of, ‘Here’s something that you can do absolutely anything you can imagine with,’” says Riparbelli, a 30 Under 30 alum. “We only care about people and videos and moderators for business content.”
Riparbelli believes Synthesia’s narrower concentrate on pure human proxies quite than modeling spaceships, dragons and other fantastical creations will keep it in competition with rivals with more firepower. And his customers usually tend to be overworked HR teams than Hollywood directors or computer graphics experts. “I don’t see the competition at the model level … and none of these companies are venturing into our space yet,” he says.
Synthesia says it has also taken a cautious approach to content moderation and reviewing scripts and videos produced on its platform. Only enterprise customers may produce news-oriented or political content behind the New York Times Synthesia avatars were reportedly used to post videos for a pro-China disinformation campaign. The risks could increase if Synthesia begins introducing generic avatars that should not trained by a selected human moderator. “As AI models become more powerful, content moderation becomes more important,” says Riparbelli.
