AI Data Startup LanceDB Raises Series B; 100x Parquet Speed Claim Rests on Its Own Preliminary Test
Reported by Tech Times1 source ↓
LanceDB announced a $57.5 million Series B co-led by Bessemer Venture Partners and Innovation Endeavors, bringing its total funding to $98.5 million. Its 100x random-access speed claim comes from a preliminary test in its own repository.
Key points
- Bessemer Venture Partners and Innovation Endeavors co-led the round
- LanceDB says the funding will expand its platform into data preparation and training-run data loading
- New Enterprise features include real-time ingestion, distributed indexing and background compaction
- The repository’s 100x benchmark does not specify hardware or testing method
Details
Tech Times reports that the round also included Theory Ventures, Dimension, Swift Ventures, CRV and Y Combinator. LanceDB last raised $30 million in a Series A led by Theory Ventures in June 2025; the coverage does not state a valuation.
LanceDB says it plans to use the funding for workflows before and during model training, including data preparation, loading and caching. The company also announced Enterprise features for real-time ingestion, indexing across large datasets and background compaction, which it says supports up to 100 billion multimodal rows.
The 100x random-access comparison with Parquet appears in LanceDB’s repository and is based on a test the repository labels preliminary, without details on hardware or method. Tech Times notes that other published comparisons vary: analyst Ben Lorica cites 3x to 35x faster random reads, while a Vortex benchmark lists Lance at about 32x faster than Parquet.
Written from coverage as of
Images
- Tech Times