TileDB, Inc., the modern database provider, is announcing its recent financial milestone, having raised $34 million in its Series B funding round, led by AlleyCorp, with participation from Two Bear Capital, Nexus Venture Partners, Big Pi Ventures, Intel Capital, Uncorrelated, Lockheed Martin Ventures, Amgen Ventures, NTT Docomo Ventures, Verizon Ventures, S Ventures, LDV Partners, and Scale Asia Ventures. The funding will serve to advance TileDB’s “modern database” mission, including support for product development, growth endeavors, and accelerating go-to-market initiatives.
TileDB is an innately scalable, multi-modal database that manages a variety of data modalities—including tables, files, images, video, genomics, ML features, key-values, point clouds, metadata, and more—as multi-dimensional arrays. The database also offers reproducibly runnable code and versatile compute, all from a single pane of glass.
TileDB’s assertion as a “modern database” aims to accommodate any type of data modality, from those that exist now, to those that will exist in the future. This sort of accommodation is a logical steppingstone from the humble beginnings of the original database, according to the company.
“Any other database, from the beginning of relational databases in the 70s up until now, is handling a single data type. We started with tables in relational databases, but we expanded this to graphs, time series, images, and so on. But for every single data type, you would have to build a different database. That was the status quo—up until now,” explained Dr. Stavros Papadopoulos, founder and CEO of TileDB.
The key to the versatility and value of TileDB is multi-dimensional arrays, which actively morphs and captures any diverse data while allowing users to construct, maintain, and run any sophisticated ETL process, pipeline, workload, or query algorithm.
All of this occurs within TileDB’s serverless distributed computing environment, ultimately centralizing data, code, and compute in a single location, eliminating silos and increasing productivity.
While other databases claim to be multi-modal, they lack this functionality at their core; instead, these databases attempt to force fit different data types into tables, ultimately sacrificing performance, according to Dr. Papadopoulos.
TileDB’s innovative design offers faster time-to-insight through strong performance and streamlined governance, as well as overall cost reduction by eliminating extraneous software licenses, lowering cloud consumption, and saving on data engineering time.
With the growing popularity in generative AI (GenAI) use cases, TileDB is fundamentally designed to support these endeavors. Vector embeddings, the data modality that serves GenAI, is just that—another type of data modality, specifically, 1D arrays natively handled by TileDB. This exists as a testament to TileDB’s invaluable, future-proof architecture, according to the vendor.
The future of TileDB, now supported by its recent funding, will continue to delve into both the education of developers on arrays and agnostic functionality.
“We want to…understand the customer even better and provide solutions always built upon this generic platform,” said Dr. Papadopoulos. “[While] we never customize for customers, we go to understand them deeper and give them tools on the database that make their lives extremely easy.”
“We are the only database that brings all data modalities, code, and flexible compute, in the same product, and that's spectacular,” he concluded.
To learn more about TileDB, please visit https://tiledb.com/.