Samsung invests in AI startup for fast video analysis at scale
▼ Summary
– Current AI models struggle to analyze and answer questions about multiple videos or footage spanning many hours, limiting their use for security and marketing companies.
– Memories.ai aims to solve this with a platform capable of processing up to 10 million hours of video, offering searchable indexing, tagging, and aggregation.
– The startup raised $8 million in seed funding, led by Susa Ventures, to expand its team and improve its video analysis technology.
– Memories.ai’s tech stack removes noise, compresses data, and indexes videos for natural-language queries, serving marketing and security clients with trend analysis and threat detection.
– The company plans to expand its capabilities, enabling shared drives for easier content syncing and envisioning AI assistants for personal and robotic applications.
AI-powered video analysis is reaching new heights as startups develop solutions capable of processing massive amounts of footage with unprecedented efficiency. While current AI models excel at summarizing short clips, they often struggle with longer videos or cross-referencing multiple sources, a critical gap for industries like security and marketing that rely on analyzing extensive visual data.
Memories.ai, a promising new player in this space, has developed a platform that can process up to 10 million hours of video, offering advanced indexing, tagging, and contextual search capabilities. Founded by former Meta researchers Dr. Shawn Shen and Enmin (Ben) Zhou, the company aims to mimic human visual memory by enabling AI to understand and analyze video content across extended timeframes.
The startup recently secured $8 million in seed funding, doubling its initial target due to strong investor interest. The round was led by Susa Ventures, with participation from Samsung Next, Fusion Fund, Crane Ventures, and others. Investors see potential not just for enterprise applications but also consumer use cases, particularly in on-device processing, a key advantage for privacy-conscious users.
Memories.ai’s proprietary tech stack processes video in three stages: noise reduction, compression to retain only essential data, and indexing for natural-language search queries. The system also includes an aggregation layer that generates summaries and reports, making it easier for businesses to extract actionable insights.
Currently, the platform serves marketing and security firms, helping marketers identify trends in social media videos and assisting security teams in detecting suspicious behavior through pattern recognition. Future plans include seamless cloud integration, allowing users to query their video archives with commands like, “Show me all interviews from last week.”
Looking ahead, Shen envisions broader applications, from AI-powered life assistants that organize personal media to training robots and autonomous vehicles by analyzing vast video datasets. The company is also expanding its team to refine its search capabilities and compete with rivals like mem0, Letta, and established players like TwelveLabs and Google.
With its unique approach to long-context video understanding, Memories.ai is positioning itself as a versatile solution in an increasingly data-driven world, one where efficient video analysis could redefine industries from surveillance to content creation.
(Source: TechCrunch)
