---
title: "AI needs the best-of-breed data stack: Postgres and ClickHouse"
description: "AI-native applications generate large volumes of operational data, while agents and users increasingly require real-time answers from that data. This pressure raises both query concurrency and reliability requirements, with a cited sample of companies using Postgres and ClickHouse showing more than 1,000% data growth in six months and over 85 TB added. The proposed stack assigns Postgres low-latency transactional workloads and ClickHouse real-time analytics, connecting them through PeerDB and ClickPipes CDC with single-digit-second replication, pg_clickhouse query pushdown, and pg_stat_ch telemetry streaming. The conclusion is that AI makes the OLTP/OLAP distinction more important, favoring an open best-of-breed stack over a single platform while preserving each database's specialized role."
---

# AI needs the best-of-breed data stack: Postgres and ClickHouse

[Clickhouse](https://yomu.fyi/company/clickhouse) · Jul 7, 2026

**Type:** Explainer

## Summary

AI-native applications generate large volumes of operational data, while agents and users increasingly require real-time answers from that data. This pressure raises both query concurrency and reliability requirements, with a cited sample of companies using Postgres and ClickHouse showing more than 1,000% data growth in six months and over 85 TB added. The proposed stack assigns Postgres low-latency transactional workloads and ClickHouse real-time analytics, connecting them through PeerDB and ClickPipes CDC with single-digit-second replication, pg\_clickhouse query pushdown, and pg\_stat\_ch telemetry streaming. The conclusion is that AI makes the OLTP/OLAP distinction more important, favoring an open best-of-breed stack over a single platform while preserving each database's specialized role.

## Context

AI applications generate rapidly growing volumes of data and require real-time analytical answers alongside transactional operations. The post states that hourly or minutely ETL is insufficient when only seconds should separate a write from its analytical availability, and that production AI systems also raise reliability and security requirements.

## Approach / What changed

Use Postgres for transactional workloads and ClickHouse for analytics, integrating them with PeerDB and ClickPipes for CDC replication, pg\_clickhouse for querying ClickHouse from Postgres, and pg\_stat\_ch for streaming query telemetry into ClickHouse. The post also presents managed Postgres in ClickHouse with built-in high availability, PITR, read replicas, forks, encryption, and native integration.

## Takeaways

- A cited sample of AI-native companies using Postgres and ClickHouse added more than 85 TB of data and grew its data volume by over 1,000% in six months.
- PeerDB and ClickPipes provide CDC replication with single-digit-second latency, along with parallel snapshotting, schema evolution, monitoring, and resilient replication.
- pg\_stat\_ch stores Postgres query telemetry in ClickHouse, supporting months of retention and high-cardinality analysis without impacting OLTP performance.

**Tags:** [Data Pipelines](https://yomu.fyi/topic/data-pipelines), [Observability](https://yomu.fyi/topic/observability), [Open Source](https://yomu.fyi/topic/open-source), [Performance](https://yomu.fyi/topic/performance), [Postgres](https://yomu.fyi/topic/postgres)

- Source: [Clickhouse](https://clickhouse.com/blog/ai-best-of-breed-data-stack)
- Source URL: https://clickhouse.com/blog/ai-best-of-breed-data-stack
- Ingested by Yomu: 2026-08-29T01:38:02.399Z

[Read original post](https://clickhouse.com/blog/ai-best-of-breed-data-stack)