How LanceDB launched Series B on X
Multimodal lakehouse for AI data in video, driving, and robotics
What LanceDB launched

Our customers are building AI that generates video, drives cars, and helps robots understand the world. We just raised a $57.5M Series B to help them do that faster, led by @BessemerVP and @iendeavors . Years ago, I helped write pandas for a world where data just meant numbers https://t.co/q5se1CSz4Z
Transcript
How LaunchVideo.com rated the launch
LaunchVideo.com rating
About the launch
LanceDB offers a multimodal lakehouse for research labs and platform teams that train models on video, images, sensor streams, and embeddings. Teams working on video generation, autonomous driving, and robotics use it to curate datasets, search and retrieve examples, engineer features, and load training data from one table, rather than moving the same files across separate storage, search, and pipeline systems. The company just raised a $57.5 million Series B led by Bessemer Venture Partners and Innovation Endeavors, with Theory Ventures, Dimension, Swift Ventures, CRV, and Y Combinator also participating, bringing total funding to $98.5 million. The platform is built on Lance, an open-source columnar format that keeps raw media, metadata, annotations, and vectors together in a versioned table on object storage. A team can add an embeddings column or new labels without rewriting the underlying video or image files, then query that table with SQL and vector search and stream batches into training. CEO Chang She, an early co-author of pandas who previously ran machine learning engineering at Tubi, co-founded LanceDB with CTO Lei Xu, whose earlier work includes HDFS at Cloudera and machine learning infrastructure for Cruise's self-driving fleet. The company names Netflix, Microsoft AI, Runway, Luma AI, and Midjourney among its customers. It reports that open-source packages have been downloaded more than 120 million times, and that revenue has grown more than 10 times year over year since its Series A in June 2025. The new capital is aimed at extending the product across more of the model development cycle, including large-scale curation, feature engineering, and data loading that can keep training clusters supplied.
Tags
Questions
Frequently asked questions
The LanceDB launch post reached 12K views on X, with 107 likes, 15 reposts, 29 comments, 9 bookmarks.
LanceDB launched on X on November 17, 2026 with the message "Multimodal lakehouse for AI data in video, driving, and robotics".
LanceDB offers a multimodal lakehouse for research labs and platform teams that train models on video, images, sensor streams, and embeddings.
LaunchVideo.com rated the LanceDB launch 3.40 out of 5: hook 3/5, editing / creativity 3/5, copy 5/5, sentiment of launch 5/5, distribution strategy 1/5.







