Building a modern data lakehouse requires moving data between specialised query engines and BI tools across different environments.
Cloud data warehouses charge massive exit fees for egress, ie moving data across platforms, and these bills stack up when querying or federating data.
Cloudflare recently launched a serverless data platform called Basin, built on Apache Iceberg and the company’s R2 object storage, which gives teams an open way to ingest, store, catalogue, and query analytical data across the tools they already use.
The release marks the company’s foray into the full ingest-to-query analytics stack
The platform became generally available at the beginning of this month and renames its three existing pieces: Basin Pipelines (formerly Cloudflare Pipelines), Basin Catalogue (formerly R2 Data Catalogue), and Basin SQL (formerly R2 SQL).
With Basin, Cloudflare pitches for a standard setup in the age of demanding analytics, provisioning servers ahead of demand, wiring together separate tools for collection, storage, and querying, and paying to move data between clouds.
Basin attempts to flatten that stack by storing data as open Apache Iceberg tables on R2, so any compatible engine can read it directly without copying or reformatting.
Besides, ingestion, catalogue management, and querying are all serverless on Cloudflare’s network, meaning no sizing, no separate systems to stitch, and no egress fees to access the data elsewhere.
“Basin’s architecture is what allows it to be economical,” Meagan Gamache, Director of Product Management of Developer Platforms at Cloudflare, tells AIM. “Cloudflare has its own globally distributed network and infrastructure that Basin utilises to be completely serverless.
She explains that this means developers and agents can go from getting started to getting answers from their data in seconds, without provisioning or managing resources.
Since it is also built around Apache Iceberg—an open format that keeps the data portable and accessible—it lets developers bring the query engines they want alongside their data.
“The combination of our serverless architecture, global infrastructure, and zero egress fees allows us to provide developers with a low-cost, purely usage-based billing model without any hourly or idle compute costs,” she says.
Its usage-based billing also ensures developers pay only when Basin ingests, processes, or queries data beyond the included usage.
“But we’re not just seeing smaller teams adopt Basin,” says Gamache. “Basin’s zero egress fees have made it a very appealing solution for enterprise use cases where developers want to store their data once and make it accessible to a wide variety of tools—and even external users—without paying a toll every time the data is accessed.”
Basin is usage-based with no hourly warehouse fees. Pipelines has free ingress and 50GB free for transforms and sinks (then $0.04/GB for SQL, $0.03-0.06/GB for sink egress), and Basin SQL is $2.50/TB of compressed data scanned after a 10GB free tier—with zero egress fees from R2 underlying it.
Another advantage of Basin is that it is tied to Cloudflare’s developer platform, “which is known for having generous included free usage,” adds Gamache. “This allows hobbyists and smaller teams to get started and operate at little-to-no cost.”
Basin lands in an already crowded market. Snowflake, BigQuery, and Athena own most analytics workloads today, and the Iceberg ecosystem itself is now contested territory—with Snowflake, AWS, and Databricks all pushing their own catalogue and query layers.
Basin’s pitch, however, focuses on what Cloudflare builds around Iceberg, which is serverless ingestion and querying, R2 as the underlying storage layer, and no egress fees.
The question is whether Cloudflare can persuade enterprises to move analytical data onto R2 in the first place. If customers continue to do their core analytics through Snowflake, Spark or other engines, how much value does Cloudflare ultimately capture from Basin beyond storage and ingestion?
Cloudflare’s position is that it doesn’t need to replace those engines to be useful.
On whether Basin is meant to be a full replacement for traditional data platforms or an infrastructure layer that sits underneath them, Gamache says the intent is to be both. “We see Basin as both an end-to-end platform for many analytics use cases and an open infrastructure layer for customers who already have tools they like.”
She explains that for teams without an existing warehouse, it means using Basin as a complete stack: ingest with Pipelines, manage Iceberg tables with Catalog, query with SQL. For teams that already run Snowflake or Spark, it’s positioned as a storage and ingestion layer that writes open Iceberg tables to R2, which those engines can then read directly without copy or egress fees.
“Our view is that open wins over time because it gives customers more choice, more portability, and more control over their data, rather than forcing everything into one closed system,” she says.
She also notes Cloudflare, with Basin, doesn’t see value as something that needs to be ‘captured’. “Rather, we see value when the Internet is more open, interoperable, and resilient.”
What is worth watching is whether Cloudflare’s model is compelling enough to change how enterprises build their data infrastructure, or whether Basin will primarily become a low-cost storage and ingestion layer feeding analytics platforms that remain elsewhere.
