TileDB builds a "modern database" that consolidates multiple data modalities, reproducible runnable code, and serverless distributed compute in a single product. Its platform treats multi-dimensional arrays as a first-class data structure to handle tables, genomics, images, graphs, key-values, point clouds and more. TileDB offers an open-source ecosystem alongside a commercial product, TileDB Cloud, for secure multi-modal data management, discovery, interactive analysis and collaboration. The company emphasizes cost savings by reducing software licenses, lowering cloud consumption, and cutting data engineering time while improving governance and time-to-insight. TileDB has seen traction with Fortune 500 customers, top pharmaceuticals, hospitals, defense and government organizations, and proven use cases in life sciences, geospatial and traditional analytics. The product now supports vector search and Generative AI workloads, positioning the platform to handle embeddings, ML code and secure private LLM compute. TileDB provides a universal database that models and stores complex data as multi-dimensional arrays and enables analysis with any programming language and tool. Its product stack includes TileDB Embedded, an open-source array storage engine, and TileDB Cloud, a serverless universal database offering secure governance, distributed compute, and data, code and model sharing. The company targets challenging data domains such as point cloud, imaging, video, location data, weather, time-series and genomics and works with IT leaders to consolidate enterprise data infrastructure. TileDB emphasizes tailored performance, elastic scalability, extreme interoperability and reduced total cost of ownership by eliminating cluster management. The company says recent funding will be used to accelerate development of advanced features and to scale go-to-market functions. TileDB was spun out of MIT and Intel Labs in May 2017 and is based in Cambridge, Mass. TileDB offers a universal database that models and stores any type of data using multi-dimensional arrays and can run on any backend (cloud or on-prem) and be analyzed with any programming language or tool. The platform unifies images, tables, videos, transcript data, functional genomics, genomic variants and flat files on a single storage layer to enable secure, performant, and elastic analytics workflows. TileDB ships TileDB Embedded, an open-source array storage engine, and TileDB Cloud, which adds secure governance, access management, scalable compute and data sharing for reproducible research. The company positions its technology to reduce data wrangling, infrastructure overhead and governance costs in genomics and healthcare and to enable global-scale secure collaboration. The funding announced will be used to advance the vision and development of the TileDB universal database. TileDB has developed a universal database that models and stores diverse data types using multi-dimensional arrays, enabling analysis on any backend and with any programming language or tool. It targets complex, non-tabular data such as point cloud, genomics, imaging, video, and location data. The product suite includes TileDB Embedded, an open-source storage engine for arrays, and TileDB Cloud, which provides secure governance, sharing, access management and scalable serverless compute. The company says standardizing on arrays delivers superior performance, elastic scalability, extreme interoperability, and global-scale collaboration. The announced funding will be used to continue key projects and advance the vision and development of the universal database. Customers and partners cited include NTT DOCOMO, which is working with TileDB on Mobile Spatial Statistics and time-series modeled as versioned, sparse arrays. TileDB builds a "universal data engine": a multi-dimensional array database and serverless infrastructure (TileDB Cloud) that manages complex data beyond tables and SQL. It targets workloads such as whole-genome population data, single-cell gene data, spatio-temporal satellite imagery, and asset-trading data, and provides APIs and integrations with Spark, Dask, MariaDB and PrestoDB. Product capabilities called out include optimized storage across numerous backends, data versioning, metadata, access control, logging, and a pluggable compute model. Customers and partners cited in the articles include Capella Space, the Chan Zuckerberg Initiative (cellxgene), and Helix, which use TileDB for geospatial, single-cell transcriptomics, and population-scale genomic storage and analysis. TileDB was spun out of MIT and Intel Labs in May 2017 and is backed by Two Bear Capital, Uncorrelated Ventures, Nexus Venture Partners, Intel Capital, and Big Pi. The company says the new funding will support expanded go-to-market and continued product development for its universal data engine and TileDB Cloud.

Total raised

$54M

Funding rounds

7

Latest round

Series B

Latest activity

Oct 2023