# Delivery Costs Cut by Up to 60%! MSK Express Now Streams Kafka Data Straight into Apache Iceberg Tables

Hey everyone, it's Shiichan! I've got some big news today for anyone running Kafka, so let's dive in!

## What was announced?

AWS's What's New page announced that Amazon MSK Express brokers now support continuously materializing Apache Kafka topics as Apache Iceberg tables in Amazon S3 Tables. It's a fully managed capability for delivering data straight into streaming tables.

## The story so far

A lot of teams ingest real-time data with Kafka for things like fraud detection and personalization, and more and more of them want to combine that data with Apache Iceberg tables for near real-time analytics.

Until now, though, connecting the two meant operating complex custom pipelines and handling format conversions yourself. On top of that, high-volume ingestion often triggered the small-file problem, where huge numbers of tiny Parquet files piled up, slowing downstream queries and driving up costs.

## What changes

With this new capability, intelligent inline compaction automatically eliminates the performance hit from small files, keeping query performance predictable without sacrificing data freshness. Built-in coordination also resolves conflicts even when multiple high-throughput consumers write concurrently.

The cost impact is significant too: compared to self-managed deployments, you can cut the cost of ingesting and delivering Kafka data into Amazon S3 Tables by up to 60%, and cut downstream query costs by up to 30% compared to self-managed Apache Kafka deployments. That's great news for data engineering teams who've been building their own pipelines and wrestling with format conversion and small-file headaches.

## Dive Deep

Here are the concrete details worth knowing:

- Delivery throughput to Apache Iceberg supports up to 10 GB/s
- This capability adds no broker egress throughput, so you avoid the extra infrastructure cost that typically comes with scaling connector-based pipelines, and can match capacity to actual demand
- Getting started is simple: open the Amazon MSK console, select your Express cluster, and enable the capability in a few clicks. You can also configure it through the MSK APIs or the MCP server
- It's available today in every AWS Region where MSK Express brokers are offered
- Check the Amazon MSK Developer Guide for setup details and the MSK pricing page for cost details

## Wrap-up

- MSK Express brokers can now continuously materialize Kafka topics as Apache Iceberg tables in Amazon S3 Tables
- Inline compaction eliminates the small-file problem, and built-in coordination automatically resolves concurrent writer conflicts
- Ingestion and delivery costs drop by up to 60%, and downstream query costs by up to 30%
- Throughput reaches up to 10 GB/s, and you can enable it from the console in just a few clicks

If your team already uses Kafka data for Iceberg or lakehouse analytics and you're tired of the cost and hassle of running your own pipeline, this update is well worth trying out.