DataHub

Acryl Data is the company behind DataHub, an open source metadata platform originally created at LinkedIn. Its commercial product, DataHub Cloud, is a fully managed enterprise SaaS offering that layers discovery, observability, data quality monitoring, lineage, and governance on top of the open source core. The platform positions itself as a "context management platform" organized into a context store (metadata, knowledge, embeddings, memory), a context layer (lineage, impact analysis, metrics, data contracts, data quality, access and policy) and a context activation tier exposed through MCP, APIs/SDKs and a conversational "Ask DataHub" interface, so that both human data teams and AI agents can query governed definitions and lineage.

The platform ingests structured signals (usage, schemas, freshness), unstructured material (SOPs, runbooks, code), business application state and semantic models such as ontologies. Features include automatically derived intent, joins and definitions from query logs and usage; freshness, volume and column-level checks surfaced at query time; and workflows in which agents propose metric definitions or business context that human experts review and approve. Earlier product descriptions covered a developer-friendly data catalog, an observability module named Observe for real-time data quality monitoring, and configurable metadata tests that flag expensive or underused data assets.

In May 2025, alongside a $35 million Series B, the company announced it was rebranding to operate simply as DataHub, describing itself as "DataHub, by Acryl Data."

Founding story

Acryl Data was founded by Shirshanka Das and Swaroop Jagadish. Das was lead architect for the Big Data team at LinkedIn, where he founded the open source projects DataHub and Apache Gobblin, and previously worked at PayPal and Yahoo!. Jagadish was head of Data and Search Infrastructure at Airbnb, where his team built the Dataportal data discovery tool, and had also worked at LinkedIn and Yahoo!. The founders' shared thesis was that the fragmentation of the modern data stack left discovery, quality and management unsolved, and that a common streaming metadata substrate could unify tools across the ecosystem. At founding, Das was CEO and Jagadish CTO; by 2023 Jagadish was described as CEO, and in the 2025 Series B announcement Jagadish is CEO and co-founder and Das is CTO and co-founder.

Business model

Open core: the DataHub metadata platform is developed as an open source project used by more than 3,000 organizations, while the company monetizes DataHub Cloud, a fully managed enterprise SaaS offering with AI-powered discovery, observability and governance capabilities. Deployment options span single-node, cloud-hosted, hybrid and decentralized configurations.

Revenue comes from the enterprise managed service DataHub Cloud, sold to enterprise data organizations; the company reported 6x growth in selling this managed service over the two years preceding May 2025.

Traction

More than 3,000 organizations use the open source platform, and the community reached 13,000+ members by May 2025 (reported as 50x growth) and 15,000+ members per the company site. DataHub Cloud sales grew 6x over the two years to May 2025. Published deployments include Pinterest, where DataHub context powers an analytics agent covering 400,000 previously ungoverned tables and became the company's most-used internal agent; Netflix, which built a unified global catalog for data, ML and software entities; Optum, which built a decentralized 'Data Forge' data mesh over petabyte-scale healthcare data; and Super Technologies, which deployed DataHub Cloud with 600+ custom SQL assertions across 800+ professionals and four markets.

Latest developments

The May 2025 Series B was accompanied by the rebrand to DataHub and a stated focus on AI governance and context management. The company subsequently announced support for the Model Context Protocol and released DataHub Cloud v0.3.17, adding native Microsoft Fabric connectors for cross-platform lineage, Ask DataHub Plugins for multi-tool context, and improved data quality monitoring. It also published survey research on the state of context management drawing on 250 IT and data leaders, and a customer story on Netflix's discovery and governance program.