Moments Lab is part of a growing number of AI startups raising money to reshape Hollywood and related fields.
Moments Lab provides tools that index video libraries and create new videos from raw footage.
The company recently raised $24 million in Series B funding from Oxx, with participation from Orange Ventures, Kadmos, Supernova Invest, and Elaia Partners. It plans to use the funding to build on its existing tech and roll out an agentic AI tool. To date, it’s raised $37.4 million.
Paris-based Moments Lab was founded by twin brothers Fred and Philippe Petitpont in 2016. Philippe was a product manager for French media company TF1 Group, while Fred was a tech lead at French ad holding company Havas’ BETC Digital. They say they saw a need for media and entertainment companies to produce more video at a lower cost to feed the growing streaming ecosystem. Their clients include Warner Bros. Discovery, Banijay Entertainment, Fullwell Entertainment, Hearst, Thomson Reuters, Sinclair, and Amazon Ads.
Moments Lab’s core tool, MXT-2, breaks down and indexes video footage into components, like people who are on screen and what they’re doing, and generates descriptions of them.
Its key pitch is that it saves users time and money. Philippe Petitpont said the tech can identify soundbites for clips for social media and things like trailers and highlight reels about seven times as fast as an employee could, citing internal research. He also said some clients have reported making twice as much revenue from social media using the product.
“For them, it’s a way to create new revenue streams,” Petitpont told Business Insider, referring to revenue from social media. “Before now, it was very complicated for production companies to create a revenue stream because there was a huge need for humans — it’s a very tedious task.”
Moments Lab’s newer agentic AI tool, which it says Hearst is among those testing, takes raw video material and turns it into rough cuts using written prompts.
Petitpont said the AI agent can do this at a fraction of the time it takes people to do the work. He said its most promising application so far is in reality TV. Scripted TV is still a work in progress.
Petitpont said Moments Lab has gained momentum with Hollywood companies since the beginning of the year, as they face pressure to make more shows at lower costs.
“The demand from Netflix, Amazon’s Prime Video, Peacock, is so strong that production companies need to produce more content than ever before, and they don’t have more money to produce a show,” he said. “Being able to produce at lower cost is more important than ever. That’s where they’re interested in using new approaches.”
AI tools are widely used in Hollywood to make production processes more efficient and, for now, generate short videos from text.
But Hollywood is highly protective of its intellectual property, which is a barrier to adoption. Another is its labor unions, which worry that AI will replace their members.
Petitpont said one of the most time-consuming parts of the sales process is assuring companies’ legal teams that it won’t use their IP to train its model.
Moments Lab says its model is trained on a dataset of 1.5 billion assets that it describes as a mix of open-source content and content from partners that are part of its research program (a consortium of research labs, media rights owners, and tech companies).
The company doesn’t shy away from the idea that automating work done by assistant editors and others can reduce the need for human workers. Petitpont said one US financial media client told him it expects to use fewer editors as a result of using its tech.
“The big question is: Will the assistant start to be the senior editor, or will the job disappear?” he said. “We don’t know yet.”
One thing there seems to be broad agreement on in the industry is that AI usage will grow — not only to save time on pre- and post-production functions, but make high-quality original video fast.
Other startups that tackle film editing functions include Runway, Filmustage, and Imaginario.
Petitpont said the ability for media companies to use AI to help make full-length documentaries based on their video libraries is only several months off, imagining a company building a film on the history of America using decades of news footage.
In a year, he expects Moments Lab to be able to provide predictive modeling tools that will make suggestions about changes to a video that could boost its audience.
“That’s what we believe will be the next thing, and we’re not very far from that because audience data is very easily available on YouTube,” he said. “The tech is not the issue, it’s more the rights.”
Check out key slides from the pitch deck Moments Lab used to raise its Series B, with some confidential information removed.
Read the full article here