DuckDB

Evidence (10)

• 6Hacker News·10/10/2026framework_update
  DuckDB 2.0 Speeds Up S3 Queries with Async I/O

  The article says DuckDB 2.0 alpha is faster, especially for querying Parquet data on AWS S3. On one benchmark reading a 2.2 GB file, runtime dropped from 18.8 seconds in DuckDB 1.5.5 to 7.7 seconds in the alpha, largely because async I/O overlaps network downloads with CPU decoding.

• 4Hacker News·9/15/2026open_source
  Alternatives to MinIO for Local Single-Node S3

  The article reviews simpler replacements for MinIO in local, single-node S3-compatible setups used for demos and validation pipelines. It compares options against requirements such as Docker support, open licensing, ease of use, and active maintenance, with a focus on running DuckDB and Iceberg against local object storage.

• 5Hacker News·9/12/2026research
  Argument for Replacing Pandas with Faster Single-Machine Tools

  The article argues that Pandas becomes inefficient at larger data sizes and pushes users toward distributed systems too early. It suggests that modern single-machine tools such as Polars and DuckDB can handle many workloads that sit below the threshold where Spark or similar platforms are truly necessary. The post cites Amazon Redshift fleet data to support the claim that most real-world analytics workloads are under roughly 100GB.

• 5Hacker News·9/9/2026open_source
  Planet Labs opens its satellite feed

  Planet Labs has made its open satellite feed available, alongside details from its ephemerides service and satellite catalog. The post highlights the company's current fleet and shows how to query publicly accessible orbital data for Planet satellites.

• 5InfoQ·8/31/2026research
  Running AI Workloads in the Browser at the Edge

  James Hall outlines why AI workloads are moving from cloud services to local edge devices, with a focus on browser-based execution. He describes using WebGPU, Transformers.js, and DuckDB to improve performance, reduce privacy risk, and support more rigorous evaluation of browser inference.

• 8Hacker News·8/26/2026funding
  AWS acquires DuckLabs

  DuckLabs will join Amazon Web Services in early September, while its Amsterdam team continues working on DuckDB, DuckLake, Quack, and related community efforts. The companies say the core projects will remain open source under the MIT license, with the DuckDB Foundation keeping stewardship.

• 7The New Stack·8/25/2026funding
  MotherDuck acquires Tower to secure its AI pipeline runtime

  MotherDuck has acquired Tower, the data infrastructure startup whose runtime was already powering MotherDuck’s AI-built data pipelines. The deal brings Tower’s team and managed Python pipeline execution layer in-house as MotherDuck expands its AI agent-driven pipeline product, Flights.

• 8InfoQ·8/17/2026framework_update
  DuckDB Previews v2.0 with Server Mode, VARIANT, and New Engine Features

  DuckDB has previewed version 2.0, codenamed Cyanoptera, which introduces a new SQL parser, a new default storage format, a reworked C API, and several breaking changes. The release also emphasizes DuckDB as a server, adding Quack-based client/server connectivity, CONNECT for remote execution, triggers, VARIANT enhancements, asynchronous I/O, and improved observability.

• 5Simon Willison·8/12/2026open_source
  Alchemy Utils 0.1a0 brings SQLite-utils style APIs to multiple databases

  Simon Willison released an alpha of alchemy-utils, a new database-agnostic library and CLI inspired by sqlite-utils. The project uses SQLAlchemy to support multiple engines, including PostgreSQL, SQLite, and DuckDB, while preserving familiar methods such as insert, upsert, create, update, and table introspection.

• 8The New Stack·7/19/2026framework_update
  Spark 4.2 adds native vector search and AI pipeline features

  Apache Spark 4.2 expands the platform with native vector search, governed metrics, improved Python interoperability, and stronger streaming support for production AI workloads. The release lets teams keep more retrieval and data-processing steps inside Spark, potentially reducing the need for separate systems such as vector databases.