Samsung invests in AI startup tackling long-term video understanding with advanced indexing and search capabilities

Samsung has invested in Memories.ai, a video-focused AI startup aiming to revolutionize how organizations process and analyze vast volumes of video data. The company’s platform is designed to handle complex tasks like cross-video context understanding, a challenge for existing AI tools that struggle with multi-hour footage. This capability is critical for security firms managing thousands of hours of camera data and for marketing teams evaluating video campaigns and product shoots across extended periods. Memories.ai’s solution addresses this gap by offering a comprehensive system that processes up to 10 million hours of video. Its technology includes a noise-removal layer, a compression step to retain only essential data, and an indexing system enabling natural-language queries to search, tag, and segment video content. An aggregation layer then synthesizes insights into actionable reports. The startup’s co-founders, Dr. Shawn Shen and Enmin (Ben) Zhou, previously worked at Meta’s Reality Labs and as machine learning engineers, respectively. The company recently secured $8 million in seed funding, led by Susa Ventures, with additional support from Samsung Next, Fusion Fund, Crane Ventures, Seedcamp, and Creator Ventures. Initially targeting $4 million, the round exceeded expectations due to strong investor interest. Misha Gordon-Rowe of Susa Ventures highlighted Shen’s technical expertise and the startup’s potential to unlock “first-party visual intelligence data,” noting a market gap for long-context video analysis. Samsung Next, the investment arm of Samsung, emphasized the appeal of Memories.ai’s on-device computing capabilities. This approach allows video data to be processed locally without requiring cloud storage, addressing privacy concerns for consumers hesitant to use security cameras. Sam Campbell, a partner at Samsung Next, noted that the technology could enhance consumer-facing applications by reducing reliance on centralized data repositories. Currently, Memories.ai serves two primary industries: marketing and security. For marketers, its tools help identify brand trends on social media and guide video content creation. Security clients use the platform to detect suspicious behaviors by analyzing patterns in surveillance footage. While the company currently requires users to upload video libraries to its platform, it plans to introduce shared drives for easier content synchronization, enabling queries like, “Tell me all about people I interviewed in the last week.” Shen envisions broader applications, including AI assistants that contextualize a user’s life through photos or smart glasses, and systems that aid humanoid robots in learning complex tasks or help self-driving cars retain route memory. The startup, which employs 15 people, will use the funding to expand its team and refine its search capabilities. Despite competition from startups like mem0 and Letta, which focus on AI memory layers with limited video support, and companies such as TwelveLabs and Google, which are advancing video understanding, Shen claims Memories.ai’s approach is more versatile. Its horizontal architecture allows compatibility with diverse video models, positioning it as a flexible solution for evolving AI needs. The investment underscores the growing demand for tools that bridge the gap between raw video data and actionable insights, as companies race to harness AI’s potential in increasingly complex applications.