Microsoft Acquires Osmos for Agentic AI Data Engineering
Microsoft has announced the acquisition of Seattle-based startup Osmos,
specializing in agentic AI for data engineering, in a strategic move to reduce time-consuming manual data preparation. The integration aims to enhance Microsoft Fabric's analytics capabilities by automating complex ingestion and transformation tasks that typically drain IT teams' time and resources. Deploying autonomous AI agents will minimize human intervention.
Agentic AI for Data Engineering
Osmos' AI agents are designed to autonomously manage end-to-end data workflows, from intake to validation, allowing teams to focus on generating insights rather than cleaning data. The acquisition signals Microsoft's growing investment in simplifying big data operations and closing the gap with rivals like Databricks on the Azure platform.
"Organizations today face a common challenge: Data is everywhere, but making it actionable is often manual, slow and expensive," wrote Bogdan Crivat, corporate vice president of Azure Data Analytics at Microsoft, in a blog post announcing the acquisition. "Many teams spend most of their time preparing data instead of analyzing it."
Osmos Background
Founded in 2019, Osmos raised $13 million in 2021 from investors including Lightspeed Venture Partners, CRV, Pear and SV Angel.
Osmos will be folded into Microsoft Fabric, the company's unified data and analytics platform. The integration will center on OneLake, Fabric's unified data lake that serves as the foundation for the platform.
The Osmos team already built native integrations with Fabric through Microsoft's Workload Hub extensibility platform. Those products included an AI data wrangler and AI data engineering agents that generate production-grade PySpark notebooks directly within Fabric environments.
According to Roy Hasson, Microsoft senior director of product, the existing Osmos products demonstrated strong customer adoption. "We quickly realized that customers loved using Osmos on top of Fabric Spark and it reduced their dev and maintenance efforts by 50%," Hasson said in a LinkedIn post.
How the Technology Works
The technology creates Fabric-native notebooks with built-in validation, metric logging and version control. These notebooks support multiple data formats including CSV, Excel, JSON, Parquet and text files.
At the core of Osmos is what the company calls an "AI Data Engineer" — an agentic system designed to function like a human data engineer. The technology interprets user requirements, writes code, performs validation and deploys to production tables pending human approval.
Competitive Landscape
The acquisition positions Microsoft's Fabric more directly in competition with Databricks, which also provides automated ETL tools on Azure. Both platforms are built around Apache Spark for data processing.
Microsoft Fabric launched in May 2023 and has received steady updates, including the Real-Time Intelligence module announced at Build 2024 and the general availability of Copilot for Fabric.
What's Next
The Osmos team will join Microsoft's Fabric engineering organization. Microsoft said it will share integration timelines and product updates through the Microsoft Fabric Blog. For more information, visit the Microsoft site.