Prefect announced that it has agreed to acquire Dagster Labs, the creator of the Dagster orchestrator. The transaction brings together two major modern workflow orchestration platforms that emerged as successors to Apache Airflow.
The combination creates a broader automation software platform for data pipelines, machine learning operations, and AI agent infrastructure. Prefect said the combined suite of products serves thousands of data teams running production workloads.
Prefect develops AI and data automation software used by engineering teams to run and observe mission-critical workflows. Dagster is known for its declarative, asset-based orchestration model, which helps teams define what data and results should exist and confirm that work has produced them.
The acquisition reflects how production workloads are evolving from traditional data pipelines into more complex agentic workflows. Prefect said modern automation requires three core capabilities: defining what work should produce, determining how that work runs, and governing how agents access tools and data.
Prefect contributes runtime execution capabilities, while FastMCP provides governed access for AI agents through the Model Context Protocol. Dagster adds declarative outcomes and asset-based orchestration, giving engineering teams a more structured way to manage data dependencies, lineage, and production results.
FastMCP has become an important part of Prefect’s agent infrastructure strategy. The project has been downloaded more than 92 million times in the last month and has earned more than 26,000 GitHub stars.
Prefect said FastMCP is now the default way to connect AI agents to external tools and data and is the basis for most MCP servers in production. Prefect built FastMCP shortly after Anthropic announced the Model Context Protocol, and Anthropic later adopted it as the official MCP SDK.
Dagster founder Nick Schrock and CEO Pete Hunt will serve as strategic advisors to Prefect and remain active in the open-source community. Prefect said Dagster and Dagster+ will keep their names, open-source license, and roadmap.
Prefect also committed to long-term support and continued investment for both products. Dagster and Prefect will continue as independently supported products, and teams running either platform in production can remain on their current systems for as long as they choose.
The company said current features and support will carry over for both products. Pricing for Prefect Cloud and Dagster+ will remain unchanged, and open-source Prefect and open-source Dagster will continue receiving maintenance releases, new features, and security patches.
The combined company plans to invest in full-stack agentic automation, reliability, and a tighter connection between flexible execution and declarative structure. Prefect said the goal is to let teams automate agentic workflows the same way they automate pipelines: defining expected outcomes, running the work, and keeping it under control.
Reliability will also be a major focus. Dagster’s materialization tracking and freshness policies can detect data problems at the point of computation, while Prefect’s event-driven execution can react when state changes. Together, the platforms are designed to help teams catch issues earlier in the workflow lifecycle.
Prefect said pairing its Python-native execution model with Dagster’s declarative structure gives engineers both flexibility and rigor. This combination is intended to let teams write code naturally while still having a clear way to reason about outcomes, dependencies, and production reliability.
Prefect has operated profitably for the past year. The company said that profitable foundation gives users confidence that Prefect, Dagster, and FastMCP will be supported by a self-sustaining business over the long term.
KEY QUOTES:
“For years, Prefect and Dagster raised the bar for each other, and as a result, we built two of the most popular tools in the data ecosystem. Together, we’re consolidating the category around our shared vision while also building the best foundation for the next generation of AI automation.”
Jeremiah Lowin, Founder and CEO of Prefect
“Prefect and Dagster are the two orchestrators the market turned to as it moved past Airflow, and putting them under one roof creates the clear leader in next-generation orchestration. That is what got me excited about this. Prefect is one of the only companies that could credibly keep building Dagster and supporting Dagster customers for the long term, and that mattered more to us than anything.”
Pete Hunt, CEO of Dagster
“As WHOOP builds more AI-driven experiences for our members and across our internal platforms, we need infrastructure that can govern automated decisions with real confidence. Prefect’s move to steward Dagster alongside FastMCP gives us the right tool for each job, with asset-aware orchestration for our data pipelines, flexible Python-native execution for dynamic workflows, and a modern protocol layer for connecting AI agents to the systems they depend on. That’s the kind of clarity engineering teams actually want.”
Carlos Peralta, Head of Data Platforms and ML Ops at WHOOP
