The news
Cloudflare (NET) made Cloudflare Basin, its serverless analytics platform built on Apache Iceberg and R2 storage, generally available on October 1, 2026, and opened K2 event streams in public beta. The company pitches Basin on zero egress fees and no clusters to manage.
Egress fees are what cloud providers charge when data leaves their storage. Iceberg is an open table format that lets different query engines work on the same data files. Basin is the new name for what Cloudflare announced in 2025 as the Cloudflare Data Platform. Chief Technology Officer Dane Knecht said the aim was "no clusters to manage, no unnecessary data movement."
The platform combines three renamed products. Basin Pipelines, formerly Cloudflare Pipelines, takes in events from Workers, HTTP requests or Cloudflare Logpush, transforms them with SQL and writes them to R2 as Iceberg tables or files. Basin Catalog, formerly R2 Data Catalog, manages Iceberg metadata and automatically handles upkeep such as compaction and snapshot expiration. Basin SQL, formerly R2 SQL, is a serverless, distributed engine that queries those tables directly on Cloudflare.
For general availability, Cloudflare said Basin Pipelines can ingest up to 3GB/s per stream and supports Terraform, while Basin SQL adds inner, outer, semi and anti joins, window functions and more than 190 scalar and aggregate functions. Basin Catalog gains per-table compaction policies and cleanup of unreferenced data files. The company said users have created tens of thousands of pipelines since the beta. Dax Raad, co-founder of Anomaly, said his company replaced an AWS S3 and Athena setup with Basin.
Cloudflare describes Basin's pricing as usage-based, with no hourly charges or separate infrastructure costs, but its announcement does not list rates. SiliconANGLE reported that customers are charged based on the amount of data they analyze. The Register pointed out a caveat: charges labeled "Sinks (egress)" apply when Basin Pipelines loads streaming data into Iceberg or Parquet files in R2 beyond included allowances.
K2 stores events as a partitioned, durable log on R2, which Cloudflare said removes the operational overhead of traditional broker clusters. The company said it first built K2 as the ingestion layer for Basin Pipelines. The beta is open to Workers Paid accounts, capped at 10GB of storage per account and 30 MB/s of writes per stream, with seven-day default retention. It is free during the beta; Cloudflare's anticipated rates are $0.04 per GB produced, $0.04 per GB consumed and $0.02 per GB per month retained.
The numbers
- Basin Pipelines ingest limit
- Up to 3GB/s per stream
- Basin SQL functions
- 190+ scalar and aggregate
- K2 beta storage cap
- 10GB per account
- K2 beta write limit
- 30 MB/s per stream
- K2 anticipated price, produced or consumed
- $0.04 per GB
- K2 anticipated retention price
- $0.02 per GB per month
Why CEOs should care
For CFOs and data leaders, the first payoff is negotiating leverage. The Register noted that Amazon S3, Google Cloud Storage and Azure Blob Storage charge egress fees, which add up when the same data is read by several engines or from several regions. Cloudflare says Basin's Iceberg tables can be read by Snowflake, Apache Spark, DuckDB and PyIceberg. That gives buyers a credible alternative to cite at renewal time with Snowflake, Databricks or a hyperscaler, even if they never move a workload.
Before moving anything, get the full price in writing. Cloudflare published no Basin rates, and the "Sinks (egress)" line The Register flagged means loading data into R2 is not free beyond the included allowances. Ask Cloudflare for a modeled bill using your real event volumes, query patterns and retention needs, then compare it with your current warehouse bill including egress and idle compute. Ask, too, how long any committed rates will hold.
For CIOs and platform teams, scope this as a pilot, not a replatforming. Constellation Research analyst Michael Ni told SiliconANGLE he does not expect larger companies to replace Databricks or Snowflake soon, but sees room for Cloudflare at the edge of the warehouse market. Log and telemetry data, which Basin can take straight from Logpush, are natural first candidates. K2's beta limits and the lack of Kafka client support mean it cannot yet stand in for an existing Kafka deployment, so treat it as a test bed for now.
The bigger picture
Basin extends Cloudflare from network and security services deeper into data infrastructure, built on the no-egress pricing of R2. Because the data sits in Iceberg tables that other engines can read, a customer could keep its existing query tools while holding the data on Cloudflare. Knecht cited open standards and customer control over data as selling points, the main contrast Cloudflare draws with closed warehouses.
SiliconANGLE described the likely buyers as small businesses and developer teams that want analytics without dedicated infrastructure. Cloudflare said its own billing and infrastructure teams used the beta for billing metrics and telemetry, the kind of high-volume event data that already flows through its network.
What’s next
Watch for a published Basin price list and for the date K2 leaves beta and billing starts; Cloudflare has announced neither. The K2 roadmap includes multi-GB/s write parallelism, message keys with key-based ordering, push-based Worker consumers, a lower-latency Express tier and Apache Kafka client compatibility. The Kafka piece will decide whether existing Kafka applications can connect without new client code.
What “Fact-checked” means
Fact-checking means testing a story’s facts against the evidence before it is published. This story went through at least two separate checks before this version was published.
- What we checked
- Its names, figures, dates, job titles, quotes and who said what were checked against the story’s sources, including its main source where it could be opened. The headline was checked for accuracy and overstatement.
- How
- A first check reviewed the whole story. If it passed, a second, skeptical check went back to the sources to look for mistakes in the most important facts. If a check flagged the story, it was edited to fix the problems found, and a separate re-check then reviewed the whole story again.
- Who
- The checks are made by our newsroom, as steps kept separate from the writing, under rules set by our editor, Hussein Mukhtar. A story the checks still flag is held for the editor, who decides whether it is fixed, published or dropped.
- If something is wrong
- “Fact-checked” does not mean error-free. If a material error is found after publication, we correct the story and add a note saying what changed. Report an error
Companies in this story
