Memories.ai is building the visual memory layer for wearables and robotics

Memories.ai, founded by Shawn Shen and Ben Zhou after their work on Meta’s RayBan AI glasses, is developing a visual memory layer for AI systems in wearables and robotics. The company emphasizes the importance of AI being able to remember visual data to function effectively in the physical world, a capability that lags behind text-based memory tools currently dominating AI development. To address this, Memories.ai has partnered with Nvidia to leverage Nvidia’s Cosmos Reason 2 vision language model and Metropolis video search technology, enhancing its ability to embed, index, and recall video data for AI applications. Launched in 2024, Memories.ai has raised $16 million through seed funding led by Susa Ventures and others. The company has built proprietary hardware called LUCI to collect video data for training its models, focusing on efficient video formats suitable for AI memory rather than high-definition recording. Their visual memory model, comparable to recent multimodal indexing models, is set to run on Qualcomm processors later this year. While