Cloudflare on Wednesday gave a name to a data service it has been testing for a year. Basin collects events, stores them, and lets a team ask questions. An event is a record of something that happened: a page view, an error, an order, a login. Companies already collect these records. The argument is about where they sit and who sends the invoice.
The usual path lands the records in cheap file storage, then pays a second product to search them. Amazon Athena is one of those search products. Snowflake and Databricks are the larger ones, used for reports the finance and operations teams already trust. Many companies pay to keep that second system ready even when nobody is asking a question, and pay again when the files have to be copied so another tool can see them.
Basin stores the records as Apache Iceberg tables on R2, Cloudflare's file storage. Iceberg is an open layout for those tables, so more than one product can read them. R2 does not charge a fee to read the files from another cloud. Basin's own question tool runs when someone asks, and Cloudflare bills for ingest, processing, and queries. There is no hourly charge for a cluster sitting idle. Dane Knecht, chief technology officer, said developers should not have to become data infrastructure operators to answer questions from their own data.
What a company actually gets
- A place to take events in, from an application or from Cloudflare's own traffic logs, and strip fields before anything is stored. Cloudflare showed a filter that keeps failed web requests and stores a scrambled customer address instead of the real one. The stored table is smaller, and a raw address never lands in the file.
- Tables kept in order, with old copies cleared out, without a separate cleanup job.
- A question tool in the Cloudflare dashboard. Joins and rankings were added during the year of testing.
- A bill that arrives when data comes in or a question runs, not an hourly charge for a cluster left on overnight.
- The same files readable by Snowflake or Databricks, without a transfer fee.
- Old setups kept. Customers who configured last year's product names do not have to rebuild them.
Rules for which country the data must sit in are still on the roadmap. A leadership team should treat Basin as a home for high-volume event feeds, not as a replacement for the warehouse that closes the books.
Cloudflare keeps the files
Cloudflare already runs customer applications on its network, already exports traffic logs, and already sells R2 storage. Basin connects those. If the event files stay there, Cloudflare collects storage fees, fees to take the data in, and fees on the questions that stay in Basin. Snowflake or Databricks can still run the monthly report. Spark, the calculation program Databricks was built on, is one of the outside tools Cloudflare says can read the tables. The warehouse vendor keeps the heavy jobs. Cloudflare keeps the files.
Amazon can hold the same kind of files, and Athena can search them. Moving those files to another cloud often carries a transfer charge. The missing charge on R2 is the opening against that bill. Cloudflare's own billing and infrastructure teams use Basin. That shows the path runs inside the company. It does not say how many outside customers have moved a production feed.
Anomaly left an Amazon setup
Anomaly makes OpenCode, an open-source coding assistant that runs in a terminal, a desktop app, or an editor and can call more than one model provider. Dax Raad, a co-founder, said the company moved its event pipeline to Basin and replaced an Amazon storage and Athena setup. A coding-assistant company produces a steady stream of event records and usually does not want a separate team maintaining a warehouse for them.
Julien Grobbelaar, head of platform at Bobsled, said Basin lets the company build data products that partners can open in any region of a major data platform, because there is no transfer fee. Bobsled's business is handing data to other companies. Neither company published row counts or the bill before and after the move. A buyer should ask Cloudflare for the price per gigabyte stored and per question before treating either quote as a savings figure.
Sources
- Selwan, Marc, and Micah Wylde. "Introducing Cloudflare Basin: An Open, Serverless Data Platform, Now Generally Available." Cloudflare Blog, 1 Oct. 2026, blog.cloudflare.com/cloudflare-basin/.
- Cloudflare. "Cloudflare Launches Basin: A More Open and Accessible Data Platform for Developers." Press release, 1 Oct. 2026, cloudflare.net/news/news-details/2026/Cloudflare-Launches-Basin-A-More-Open-and-Accessible-Data-Platform-for-Developers/default.aspx.
- Cloudflare. "Basin." Cloudflare Docs, updated 1 Oct. 2026, developers.cloudflare.com/basin/.
