Cloudflare Basin: serverless Iceberg analytics platform hits general availability

General availability under a new name

Cloudflare's analytics platform is now generally available, and it arrives with a new identity: Cloudflare Basin. The suite was first announced during Birthday Week 2025 as the Cloudflare Data Platform, and according to the Cloudflare blog the rename is intended to pull the individual products under a single umbrella.

Basin is a serverless platform for analytical data built on Apache Iceberg — which Cloudflare describes as the standard open table format for data lakes — and R2 Object Storage. It spans three products, each renamed from its beta-era title:

• Basin Pipelines, formerly Cloudflare Pipelines, receives events from Workers, HTTP endpoints or Cloudflare Logpush, transforms them with SQL, and writes them as Apache Iceberg tables or as files in R2.
• Basin Catalog, formerly R2 Data Catalog, manages Iceberg metadata and automatically maintains tables to keep them fast and cost-efficient.
• Basin SQL, formerly R2 SQL, is a serverless, distributed SQL engine for querying Iceberg tables directly on Cloudflare.

Cloudflare says it began building the platform after noticing two shifts: Apache Iceberg consolidating the market around an open table format portable across nearly every major query engine, and developers moving analytics data into R2, where the absence of egress charges made access from different tools, teams, regions and cloud providers practical.

What changed since the beta

During the beta, adopters — including Cloudflare's own billing and infrastructure teams — used the services for real-time e-commerce optimization, long-term storage and reporting of billing metrics, and infrastructure telemetry, and Cloudflare reports that users created tens of thousands of Pipelines. Anomaly co-founder Dax Raad said the company moved its entire event data pipeline onto Basin, replacing an AWS S3 and Athena setup.

The GA release raises ingestion capacity to up to 3GB/s per stream, alongside several integration improvements. Cloudflare Logpush can feed logs through SQL transforms into compressed Parquet files or Iceberg tables; Worker bindings are schema-aware, with wrangler types generating TypeScript types from a stream's schema so missing fields and type mismatches surface before deployment; data-quality errors such as missing fields, type mismatches, parse failures and null values are visible in the dashboard and via the GraphQL API; and Terraform resources cover the catalog, stream, sink and the SQL that connects them.

Cloudflare says planned additions include custom partitioning when writing to Basin Catalog and schema migrations.

Speed, openness and pricing

Cloudflare frames the past year of engineering work around speed, openness and cost efficiency. A catalog, an ingestion pipeline and a first query can be set up in seconds — a property the company ties to the growth of data applications built around prompts and coding agents, which would otherwise wait and poll for resources. As datasets grow, Basin Catalog compacts metadata and data files to reduce I/O and generates statistics for query planning, and Basin SQL uses those statistics to split queries into smaller tasks distributed across Workers.

On openness, data remains in Iceberg, so it can be read and written with any compatible engine, including PyIceberg, DuckDB, Snowflake and Apache Spark. Julien Grobbelaar, head of platform at Bobsled, said this portability combined with zero egress fees lets his company make data products accessible in any region of every major data and AI platform at a fraction of the cost.

Pricing is usage-based: customers are billed when data is ingested, processed or queried, with no hourly charges or separate infrastructure costs. Cloudflare positions this as workable for hobby projects at little or no cost while scaling for larger enterprise use cases.

Why it matters

Basin bundles ingestion, table management and querying into one serverless offering, removing much of the infrastructure work normally required to run an Iceberg-based lakehouse. The Iceberg foundation means the data is not locked to Cloudflare — any compatible engine can query it, and R2's free egress lowers the cost of exercising that portability. Together, those properties put Cloudflare in more direct competition with cloud data warehouses and managed lakehouse services, and the launch arrives as AI agents and other data-heavy applications increase demand for analytics that can be stood up and queried immediately rather than provisioned in advance.