CedarDB's unified database approach gains traction as AI agents demand faster analytics on transactional data

Database systems have long been divided into two distinct categories: analytical engines that help organizations identify patterns and derive insights, and operational systems that manage everyday transactional data such as orders and payments. CedarDB, a database engine based in Munich, breaks this traditional divide by supporting both workloads simultaneously—a capability that has become increasingly valuable as agentic AI systems proliferate, according to cofounder and CEO Moritz Sichert.

The company has spent the past decade in development, but its recent momentum reflects a growing need among enterprise customers. Many organizations maintain data infrastructure optimized for storing years of transaction records, but struggle to quickly answer analytical questions posed in natural language against that same data.

CedarDB's rejection of the analytical-operational divide stems from its core design philosophy centered on efficiency. The startup emerged from the Technical University of Munich as a spinout based on Umbra, a research initiative created specifically to outperform conventional database engines by leveraging modern hardware capabilities more effectively.

Five researchers from the Umbra project serve as CedarDB's cofounders and have maintained this performance-first approach. The team engineered the system to deliver superior speed without demanding additional CPUs, processing power, or memory—and without forcing customers to overhaul their existing applications. A critical design decision was ensuring compatibility with PostgreSQL, the widely-used open-source database platform.

According to Sichert, PostgreSQL compatibility reduces the friction of switching to CedarDB since users can retain their existing applications without rebuilding them. The choice carries another benefit: because large language models were trained extensively on PostgreSQL syntax and logic, LLMs naturally align with CedarDB's interface.

While some of these decisions proved fortuitous, Sichert emphasized that strategy guided many choices. In a blog post, cofounder Christian Winter outlined the priorities the team established upon transitioning from academia, including "reliability, fault tolerance and compatibility with popular tools."

https://www.youtube.com/embed/oLAH4RfCZXQ?si=Z7LUr39_aXW4uwTd&autoplay=1&mute=1

The path forward involved significant internal debate. The team deliberated extensively about whether to launch a cloud product immediately. Ultimately, CedarDB released CedarDB Enterprise as an on-premise solution alongside CedarDB Community Edition, a free tier, both unveiled in May of the previous year alongside a €5.3 million seed funding round, equivalent to approximately $6 million.

The cloud product was postponed rather than abandoned. Following beta testing with users on the company's waitlist, CedarDB announced limited availability of CedarDB Cloud on Wednesday during a presentation at TechCrunch Disrupt, where it competes among 20 other startups in the TC Startup Battlefield competition.

A fully managed cloud-hosted platform should enable CedarDB to reach beyond its current enterprise customer base and strengthen its competitive position against established players like Databricks, which appears to have arrived at similar conclusions about market direction.

At its Data + AI Summit in June, Databricks announced LTAP, a new data architecture designed to "[unify] transactions, analytics, streaming, and operational data on a single copy of storage in the lake," citing "the agentic era" as the driving rationale.

While the ultimate impact remains uncertain, Sichert views Databricks' announcement as validation that demand exists for CedarDB's approach. The startup intends to serve its initial cloud customers with its current team of 14 employees, though Sichert anticipates needing additional headcount as the business scales. "And this is where we would raise our next round," he stated.