Samsung backs a video AI startup that can analyze thousands of hours of footage

Many AI tools can watch a video today and summarize what is going on, but things become a bit tricky if you ask models about multiple videos and images that span many hours.

This is a major restriction for security companies that want to use AI to scrub thousand hours of images from different cameras, as well as marketing companies that want to study different video campaigns and product shoots.

Memories.ai Want to tackle that problem with its AI platform that can process up to 10 million hours of video. For companies with many videos to analyze, the startup wants to offer a contextual layer, complete with searchable indexing, tagging, segments and aggregation.

The co-founder, Dr. Shawn Shen, was a research scientist at Meta’s Reality Labs while he was pursuing his promotion, and his counterpart Enmin (Ben) Zhou worked at Meta as an engineer of Machine Learning.

“All top AI companies, such as Google, OpenAi and Meta, are aimed at producing end-to-end models. Those possibilities are good, but these models often have restrictions on understanding video context after one or two hours,” Shen told WAN.

“But when people use visual memory, we search a large context of data. We are inspired by this and wanted to build a solution to better understand video over many hours,” he said.

Opposite that goal, the company has now collected $ 8 million in a seed financing round led by SUSA Ventures, and with the participation of Samsung Next, Fusion Fund, Crane Ventures, Seedcamp and Creator Ventures. Shen said the company initially aimed to collect $ 4 million, but ended with an excess round due to investor’s interests.

“Shen is a very technical founder and he is obsessed with pushing boundaries of video comprehension and intelligence,” said Misha Gordon-Rowe, a partner at Susa Ventures. “Memories.ai can unlock a lot of visual intelligence data in the first party with his solution. We thought there was a gap in the market for a long context of visual intelligence, which attracted us to invest in the company,” he added.

Samsung then had a slightly different statement: the samsung investment sees sees memories. AI’s solution is useful for consumers.

“One thing that we liked about memories. Ai is that it could do a lot of computer use on the devices. That means that you do not necessarily have to store video data in the cloud. This can unlock better security applications for people who are afraid to place security cameras in their home due to privacy problems,” said Sam Campbell, said Sam Campbell, said Sambell, Next, said Sambell, Next.

Memories.ai says it uses its own technical stack and models to perform analyzes. Firstly, the noise removes videos and passes the output through a compression layer to store only important data. Then there is an indexation layer, which makes the video data searchable (using natural language queries) with segmentation and tags. There is also an aggregation layer that summarizes data from the index, making reports.

The startup is currently suitable for two types of companies: marketing and security. Marketing companies can use the tools of the startup to look up trends related to their brands on social media and identify what kind of video they want to make. Memories.AI also offers marketers tools to make those videos.

The company also works with security companies to help them analyze security images to determine potentially dangerous actions of people in the videos by reasoning through patterns.

Currently, companies that work with memories have to upload their video library to the platform to have these clips analyzed. But Shen said that in the future his customers will be able to create a shared drive and to synchronize content more easily. The plan is to enable customers to ask questions such as: “Tell me everything about people I interviewed last week.”

Shen presents an AI assistant to get that context in the life of a user through their photos or when they activate smart glasses. He also sees the technology playing a role in training humanoid robots to perform complex tasks or to help self -driving cars remember different routes.

The company currently employs 15 people and it is planning to use the fund to increase its team and improve its search.

Memories.ai goes against similar startups, such as Mem0 And Letta, who work on offering a memory layer for AI models, although they currently offer limited video support. It must also compete with companies such as Twelvelabs and Google, who have worked to help AI models to understand videos.

However, Shen believes that the solution of his company is more horizontal, so it could also work with different video models.