The Ever Expanding World of New Database Systems
The Ever Expanding World of New Database Systems
Carnegie Mellon University’s tracker shows dozens of new DBMSs are in development
Welcome to the Cloud Database Report. I’m John Foley, a long-time tech journalist who also worked in strategic comms at Oracle, IBM, and MongoDB. A big welcome to new subscribers and thank you to paid subscribers — your support is appreciated. Connect with me on LinkedIn.
Net new DBMS development
In the age of AI, there will be more data, more database instances deployed in compute infrastructure, and — here’s the surprising part many people may not realize — more new types of database management systems.
That’s my prediction, and I’m sharing it because it’s easy to fall into the trap of thinking of databases as a mature, legacy market. That’s partly true, but database innovation is alive and well.
There’s no escaping the fact that the first DBMSs were introduced 60 years ago in the age of Cobol. Or that the top five database providers — AWS, Microsoft, Oracle, Google Cloud, and IBM — have been entrenched at the top of the market for seven consecutive years. So yes, there may be cobwebs on some old database servers out there.
However, barely a day passes that I don’t see or hear about a new database innovation, platform, or startup. Recent examples: CedarDB, DuckLabs, Keewano, pgEdge, and Supabase.
And that’s just a sampling. Dozens more are in the pipeline. In fact, 55 new database systems have gone into development thus far in 2026, according to Carnegie Mellon University’s comprehensive Database of Databases.
Because tech teams are dealing with so many new data types and use cases, these emerging systems are designed around various data models: vector, graph, relational, key value, document, array, and more.
You can view Carnegie Mellon’s Database of Databases here.
Hobby vs. commercial systems
Some of the 55 new databases are categorized as hobby projects. Those include AionDB, described as “a multimodal database engine written in Rust that integrates relational, graph, and vector data models into a single system.” And GitDB, a “GPU-accelerated, version-controlled database system designed for vectors, documents, and tables,” per Carnegie Mellon.
Others are intended for commercial use. Among them:
ChronoVec, described by its project team as “vector memory that thinks in time.” Its Github site includes an agent memory demo.
Jazz, promoted by its developer, Garden Computing, as “the database that syncs.”
L, a new runtime for k and q. In other words, it’s a fast database and array language interpreter used on Wall Street, according to its developer.
Nexir, a system for establishing a durable, replicated state for AI agents. It’s in public beta now.
Not all of the work underway are indie projects. There are also new database systems from market leaders AWS (ExtendDB) and Microsoft (Azure HorizonDB).
More of everything: data, instances, DBMSs
What I’m talking about in this post are newly developed database management systems.
That said, the sheer number of database instances (i.e. Amazon Aurora, Google Spanner, Oracle, SQL Server, etc.) being deployed is also on the rise. That’s due in part to the fact that more databases are being created autonomously by AI agents, according to Databricks, which has described a trend toward “small, ephemeral databases” that may have a compute lifetime of less than 10 seconds.
Big picture: As I’ve been reporting, there are new data types and more data than ever, with petabytes growing into exabytes. To keep up, these new — often special-purpose — databases are being developed to handle AI, agents, and other modern workloads.
Going back to 1964, Carnegie Mellon’s Database Group has tallied a total of 1,228 database systems. Of those, 382 (nearly a third) are identified as “abandoned,” so it’s always good to have contingency and migration plans if needed.
I will note that some new data architectures don’t required databases as all. Vector indexes and event streaming are two examples, and there are others. That’s a blog post for another day.
Meantime, my expectation is that Carnegie Mellon’s Database of Databases will keep growing.
More from the Cloud Database Report
© 2026 John Foley
Huntington NY 11743