What was claimed
CedarDB, a Munich startup spun out of the Technical University of Munich, says a single database can handle both operational work and analytics. CEO Moritz Sichert ties this to AI agents raising the stakes, and sees proof of demand in Databricks launching a similar product, LTAP, in June.
What holds up
Not a lot, but some. The company exists, has 14 employees, raised €5.3 million in 2025 and ships a free Community edition and an enterprise one. PostgreSQL compatibility does lower the cost of switching. A limited cloud launch was announced on October 14.
What remains words
The central pitch, that CedarDB is faster without needing more CPUs or memory, is not backed by anything: no benchmarks, no comparisons, no figures. No customers are named, and there is nothing on users, revenue or pricing.
The tie to AI agents also looks stretched. Merging transactions and analytics is an old idea that predates agents, and the point that LLMs handle PostgreSQL well speaks to compatibility, not to a new need. Databricks launching a product shows a big player finds the topic interesting, not that the market will pay for CedarDB. The reporter herself notes it is "too early to tell" how disruptive the Databricks move will be.
Bottom line
The idea is sensible and the research-backed team is serious. For now, though, it is a promise without measurements, and worth testing on your own data.