CedarDB wants one database to handle both transactions and analytics

CedarDB, the Munich engine for analytical questions and day-to-day operational records, announced limited availability of fully managed CedarDB Cloud on Wednesday at TechCrunch Disrupt. TechCrunch's report places that cloud tier beside on-premise CedarDB Enterprise and the free CedarDB Community Edition. Co-founder and CEO Moritz Sichert told TechCrunch that agentic AI is why one database for both jobs matters now.

This piece covers the Cloud launch, the Umbra research under the engine, PostgreSQL compatibility, Databricks' June LTAP launch, and the 14-person plan for the first hosted users. It is for engineers who store orders and payments in an operational database, architects weighing a second system for natural-language questions, and readers following Startup Battlefield.

CedarDB wants one database to handle both: what actually changed

Analytical databases help businesses spot trends and answer questions. Operational databases store day-to-day records such as orders and payments. CedarDB, the Munich database engine, supports both and has been in the making for the last decade. Sichert said demand comes from enterprises whose stacks store years of transactions yet answer natural-language analytical questions slowly. In May of last year the company launched on-premise CedarDB Enterprise beside free CedarDB Community Edition and announced a €5.3 million seed round, approximately $6 million.

The team strongly debated launching a cloud offering right away. The cloud offering was only delayed, though, not abandoned. After beta testing with waitlist users, CedarDB on Wednesday announced limited availability of CedarDB Cloud onstage at TechCrunch Disrupt, competing with 20 other startups to win TechCrunch's prestigious Startup Battlefield competition. The fully managed service is meant to reach users beyond large enterprises. The founders, listed from left to right, are Moritz Sichert, Lukas Vogel, Philipp Fent, Dominik Durner, and Christian Winter, the five Umbra researchers.

CedarDB wants one database to handle both: how it works

Questioning the split between analytical and operational queries is a by-product of CedarDB's focus on efficiency. The company is a Technical University of Munich spinout based on Umbra, a research project built from scratch to beat legacy database engines at processing speed by making the most of modern hardware. Five Umbra researchers co-founded the company and kept that approach. CedarDB promises faster performance without requiring more CPUs or processing power and memory, and it requires few changes from its customers.

In particular, CedarDB is compatible with PostgreSQL, an open source database system. Sichert said that match keeps switching costs low, because users do not rebuild their apps, and that the engine follows logic large language models are most familiar with from their training. Co-founder Christian Winter's blog post listed priorities after the team left academia, "such as reliability, fault tolerance and compatibility with popular tools." Sichert called the PostgreSQL choice one of many that proved lucky in retrospect. It was not all luck.

CedarDB wants one database to handle both: why it matters now

Sichert told TechCrunch the CedarDB approach gained relevance with agentic AI. Enterprise stacks were well suited to storing years of transactions and less suited to quickly answering analytical questions asked in natural language about that data. Questioning the divide between analytical and operational queries is a by-product of its focus on efficiency. Umbra, the research project under CedarDB, was built to beat legacy engines on processing speed by using modern hardware, and CedarDB kept that approach for a single engine that does both.

At its Data + AI Summit in June, Databricks cited "the agentic era" as the reason for launching LTAP, a data architecture aiming to "unify transactions, analytics, streaming, and operational data on a single copy of storage in the lake." It is too early to tell how disruptive this might be. Sichert interprets that move as another sign of demand for what CedarDB does. CedarDB Cloud leaves the German startup better prepared to compete with Databricks, the major American data company that reached a similar conclusion.

CedarDB wants one database to handle both: who is affected

Enterprise stacks that already hold years of transactions, and that are less suited to quick natural-language analytical questions, are what Sichert describes. PostgreSQL users are the migration path: his compatibility claim means they do not rebuild apps to try on-premise CedarDB Enterprise or free CedarDB Community Edition. On-premise CedarDB Enterprise is the offer for organizations that choose to run the database themselves. CedarDB Cloud is the fully managed, cloud-hosted version meant to help expand the user base beyond large enterprises.

Waitlist users from the beta are the first CedarDB Cloud accounts, and the company plans to serve them with its current team of 14 people. Sichert expects a larger team once the hosted product gains traction. Twenty other startups are competing with CedarDB for TechCrunch's Startup Battlefield win at Disrupt. Databricks' LTAP, which puts transactions, analytics, streaming, and operational data on one copy of storage, is the design set beside CedarDB. Vogel, Fent, Durner, Winter, and Sichert are the five Umbra co-founders.

CedarDB wants one database to handle both: what to watch

CedarDB plans to serve its first cloud users, under limited CedarDB Cloud availability, with its current team of 14 people. Sichert expects to need a larger team once it gets more traction. "And this is where we would raise our next round," he said. The round already announced is the €5.3 million seed, approximately $6 million, from May of last year, paired with CedarDB Enterprise and CedarDB Community Edition. The on-premise product and the free edition remain available while the hosted tier stays in limited availability.

Databricks' LTAP aims to unify transactions, analytics, streaming, and operational data on a single copy of storage in the lake, and Sichert takes that goal as demand for CedarDB's combined engine. It is too early to tell how disruptive the shift might be. After leaving academia, Winter detailed priorities "such as reliability, fault tolerance and compatibility with popular tools." Sichert's PostgreSQL terms sit with that list: low switching costs, no application rebuild, and a dialect large language models already know from training.