Your Data Lives Where You Say It Does. Real-Time Analytics Should Too.
With the acquisition of Altinity, EDB Postgres® AI brings Postgres and ClickHouse® software together on one road map, uniting transactions, real-time analytics, and agentic AI wherever your data has to live.
Postgres is the most-used database among professional developers. It’s the system of record, where a business’s core transactions live, and the teams I speak with love it for its reliability, open source roots, and huge ecosystem of extensions.
But when someone needs a live answer from millions of records, Postgres starts to sweat. So teams add a second system, their data ends up in two places, the analytics side runs a step behind, and someone has to maintain the pipelines in between.
EDB Postgres AI takes Postgres further for the enterprise, with up to 99.999% availability, built-in tools for AI agents, governance controls, and the freedom to run wherever the data has to live. It also helps with analytics, because it shares live Postgres data with analytics tools without a separate pipeline. EDB has offered supported ClickHouse software, a fast open source analytics database, for real-time queries.
The data has to be in one architecture, and in this case, an open source one. So it was inevitable that the teams behind the two of the most popular open source databases, Postgres and ClickHouse software, would come together. That’s why I’m excited to share that yesterday EDB announced that it acquired Altinity, the company behind many of the contributions to ClickHouse software’s open source ecosystem. Now, we ship both engines in one multi-model data layer, all working together in one sovereign AI data system.
Two engines, two jobs
The magic is two leading open source databases working together in EDB Postgres AI, uniting transactions, real-time analytics, and agentic AI.
Postgres is your system of record. It’s what’s true right now, where transactions commit, and consistency matters. ClickHouse software is your system of insight. It’s a columnar engine built to scan billions of events (clicks, carts, sensor readings) and answer questions about them in under a second. It tells you what’s happening right now. That’s the speed a dashboard, an analyst, or an AI agent needs before it acts.
Put them together, and each engine does the job it’s good at. Analytical queries run on ClickHouse software, so they add no load to the system of record, and both engines work together in one multi-model data layer. That's why Forrester named EDB a Leader in The Forrester Wave: Multimodel Data Platforms, Q2 2026.
We first invested in ClickHouse software capabilities in June to support real-time analytics, and with Altinity now part of EDB, that engine sits next to Postgres on one roadmap. The real power of open source is the collective of contributors who keep building new innovations and upstreaming them into those communities. EDB is already the number-one commercial contributor to Postgres, and Altinity is the second-largest independent contributor to ClickHouse software (behind only ClickHouse, Inc.).
Holding the number-one and number-two positions across the two leading open source databases means we can innovate more and faster together than we ever could separately. Both communities benefit, and so do our customers.
What it looks like: Fresh answers from live data
Let’s see what this looks like in practice. Picture a retailer on a new product launch day. Orders, payments, and inventory commit in Postgres. Clickstream and cart events pour into ClickHouse software.
A dashboard, or an agent, asks which SKUs are selling through fastest in which regions, and the answer comes from data that’s only seconds old. Not yesterday’s data. The follow-up action—say, reallocating inventory—goes back to Postgres as an ordinary transaction.
Getting live data from Postgres to ClickHouse software is the part people expect to be hard. Today there are three good paths.
- Through Iceberg: EDB’s Postgres Analytics Accelerator (PGAA) publishes your Postgres tables to Apache Iceberg in your own object storage, and ClickHouse software reads that same copy. There’s no ETL job to write, and other lakehouse tools can read the same open tables.
- Through change data capture: If you want a direct stream, you can carry changes from Postgres into ClickHouse software with Debezium on Kafka. Kafka stays your event backbone, and ClickHouse software can consume its topics natively.
- Through the PostgreSQL table engine: ClickHouse software can also query a live Postgres table in place, reading with SELECT and writing with INSERT. Simple filters run on the Postgres side, while joins and aggregations run in ClickHouse software, so it works well for pulling in reference data or making an initial copy.
Run it where your data has to live, the same way everywhere
The same multi-model system runs on-premises, in any cloud, as bring-your-own-cloud, in hybrid setups, and fully air-gapped. You decide where the data lives and who holds the off switch, and in a customer-controlled deployment EDB has no access to your data and no switches. That’s truly differentiated among real-time analytics database offerings today.
Both engines are open source and Iceberg is an open format, so you can read the code, move your workloads, and take your data with you. The choice remains yours.
Running anywhere only works if you can operate the same way everywhere, and that’s where Kubernetes comes in. More than 56% of enterprises deploy mission-critical workloads in a hybrid model, and nearly half of organizations run at least half of their data workloads on Kubernetes in production (Data on Kubernetes, 2025). Your database needs to run wherever your applications do and be managed the same way everywhere, especially now that AI agents need to act on all of that data in real time, with security and control.
That consistency comes from a Kubernetes operator, a controller that keeps your database matching the state you declare. Both Postgres and ClickHouse software have a leading one. For Postgres, that’s CloudNativePG, which EDB created and donated to the community, and it’s the #1 Postgres operator for Kubernetes. With Altinity, we now bring the ClickHouse software side too: the first Kubernetes operator built for real-time analytics, pulled more than 100 million times.
Having both gives your team one operating model for both engines, the same declarative, version-controlled way to run your transactional database and your analytics engine in any environment. Your platform team learns one pattern, and moving between on-premises, cloud, and hybrid doesn’t mean relearning how to run your data. One company now stands behind a leading operator for each database. That lets a team start with the Postgres it already runs, add real-time analytics beside it, and give agents live data without rebuilding the estate.
Come say hi at KubeCon
Keep running the Postgres your business already depends on, add ClickHouse software beside it for real-time answers, and run both wherever your data needs to live.
I’ll be at KubeCon in Salt Lake City next week with Robert Hodges, CEO of Altinity, talking about how both of our Kubernetes operators work together in practice. Stop by Booth #280 with your questions.
ClickHouse is a registered trademark of ClickHouse, Inc. https://clickhouse.com.