The era of rapid, ungoverned AI deployment is officially over. Bedrock Data, an AI-native platform focused on data security and governance, announced it has closed a $25 million Series A round led by Greylock Partners, capitalizing on the urgent enterprise need to control the data fueling their models.
The funding arrives as regulatory bodies, including the EU and SEC, increasingly demand documentation on the data used in high-risk AI systems. According to Bedrock Data’s own research, 82 percent of cybersecurity professionals struggle significantly with finding and classifying organizational data, a problem compounded by the fact that 60 percent of security teams now have expanded responsibilities that include AI governance.
This shift has moved the data visibility problem from a theoretical concern to a critical operational priority, according to CEO Bruno Kurtic. Enterprises are no longer just asking how to deploy AI, but how to govern the systems and agents already in use.
“Heading into 2026, the priority is operationalizing governance at the data layer,” Kurtic said. “Leaders need to build systems that maintain continuous visibility into their data sensitivity and exposure, trace how data moves through analytics and AI workflows and understand which models or agents interact with it.”
Mapping the AI Data Supply Chain
Bedrock Data’s core innovation is ArgusAI, launched late in 2025. ArgusAI directly addresses the governance gap by automatically creating a Data Bill of Materials (DBOM) that links specific AI models to their underlying datasets. This allows security teams to perform guardrail gap analysis, validating whether existing controls actually prevent sensitive data from being accessed by models during training or inference.