(Bloomberg) -- Captions, a startup that uses artificial intelligence to let people create and edit videos, has raised a new funding round from a star-studded list of investors, vaulting its valuation to $500 million.
The startup raised $60 million in a deal led by Index Ventures, the company plans to announce Tuesday. That brings its total funding raised to $100 million. Other investors in the round include previous backers Kleiner Perkins, Andreessen Horowitz and Sequoia Capital, as well as new ones, like actor Jared Leto. (Leto also backed another AI video company, Pika, in June).
Captions is one of a growing number of companies aiming to make it cheaper and easier to create videos with generative AI. Founded in 2021 by Gaurav Misra and Dwight Churchill, the company’s software helps people make, edit and distribute videos whether or not they have filmmaking experience, said Misra, Captions’ chief executive officer.
As the quality of AI videos has rapidly improved, the technology has heightened fears about deepfakes. Particularly in a big year for global elections, some have raised alarms about video tools that are increasingly capable, cheap and accessible. But there is also significant demand: People publish nearly 3.5 million videos per month with Captions’ app, Misra said, and more than 10 million people have downloaded it so far.
Captions’ users range from realtors to salespeople to educators using it in the classroom, the CEO said. “It’s very wide ranging in terms of the use case — it’s wherever video is relevant, especially short-form video,” he said.
Captions’ specialty is videos that feature a person speaking. Users can type text into the software that can then be spoken by a photorealistic AI avatar. Or they can simply write in a topic for the software to build a script around, like “instructions for making a balloon dog.” The dialogue can be translated into a slew of languages. Users can also upload their own videos and translate the captions and dialogue into another language, using AI to make the translated voice sound like their own.
To do all this, Captions combines its own AI models for making videos with those built by others. It employs large language models from companies such as OpenAI and Anthropic to generate text, and software from voice-cloning startup Eleven Labs Inc. to generate voices.
To try to curb potential misinformation, Misra said the company made choices with the design of its software and included limitations. The app won’t generate videos about some topics. For example, in a recent test by Bloomberg News, it refused to make a video including disinformation about vaccines.
Misra said the rapid progress and fast-rising popularity of generative AI over the past few years have helped the company speed up the development of its own tools. The company will spend some of its new funding on research and development, along with compute power, seeking to improve its AI models.
“What we thought would take 10 years takes one year now,” Misra said.
©2024 Bloomberg L.P.
