Sentra Raises $150M Series C to Expand Agentic Data Security Capabilities

Data security platform provider Sentra announced a $150 million Series C funding round led by Capital One Ventures, bringing the company’s total capital raised to over $260 million. Co-headquartered in New York City and Tel Aviv and founded in 2021 by CEO Yoav Regev and his co-founders, Sentra develops cloud-native software aimed at protecting corporate data assets as enterprise workflows increasingly rely on autonomous AI agents and complex cloud architectures.

The investment highlights a growing security priority: as organizations rapidly deploy artificial intelligence workloads, managing which identities—human, programmatic, or autonomous AI agents—can reach sensitive internal data has become a critical operational friction point. Sentra plans to use the capital injection to accelerate its agentic data security capabilities, enhance visibility into AI data pipelines, scale its go-to-market organization, expand global brand footprint, and enter new regional markets.

The Threat Vector: Machine-Speed AI Agents and Data Sprawl

Enterprise adoption of agentic AI—autonomous software entities that interact with databases, execution environments, and cloud services to execute multi-step workflows—fundamentally changes the speed and scope of data access. Traditional Data Security Posture Management (DSPM) tools were built to discover, classify, and track static data across cloud repositories such as AWS S3, Snowflake, Google BigQuery, and enterprise SaaS applications. However, the rise of Retrieval-Augmented Generation (RAG) architectures and autonomous agents introduces dynamic identities that traverse knowledge stores and file shares at machine speed.

When an AI agent is given broad API permissions or service account credentials, its blast radius expands significantly beyond that of a human operator. If an agent retains excessive read privileges, a prompt injection attack, flawed tool-calling instruction, or credential compromise can allow the agent to query, extract, or aggregate sensitive regulated data—such as personally identifiable information (PII), proprietary source code, or financial records—and expose it through LLM outputs, secondary caches, or external third-party tools. Traditional perimeter firewalls and identity governance frameworks are often blind to these machine-to-machine micro-transactions, creating blind spots where toxic combinations of excessive access permissions and sensitive shadow data intersect.

How Sentra Addresses Agentic Risk and DSPM

Sentra’s platform provides unified visibility across cloud environments, SaaS platforms, data warehouses, vector databases, and on-premises file systems to continuously locate and classify sensitive data at rest. By correlating data sensitivity classifications with exact permission maps, the platform tracks which human users, service accounts, and AI agents have actionable pathways to specific assets.

Key technical capabilities embedded within the platform include:

• Continuous Asset & Identity Mapping: Graphing access pathways across structured and unstructured data repositories to identify over-privileged AI agents and service accounts before exploitation occurs.
• Least Privilege Enforcement: Identifying and eliminating unnecessary data access pathways that agents do not require for their designed business tasks, shrinking the potential blast radius.
• Automated Remediation & Threat Containment: Monitoring for anomalous data access behavior across cloud workloads, triggering automated access revocations, and blocking high-risk data queries in real time.
• Contextual Audit Reporting: Generating detailed reports that outline exactly which data assets an agent could access, complete with business context and risk metrics for compliance and security engineering teams.

By enforcing continuous least privilege at the data layer rather than relying purely on prompt filters or model guardrails, the platform prevents AI workflows from indexing or outputting data that should remain restricted.

Architectural Requirements for AI Data Governance

As enterprises expand their reliance on external model providers and internal AI agent networks, securing data at rest and in transit requires a shift toward data-centric security controls. Security leaders evaluating AI enablement must address several core architectural requirements:

1. Explicit Identity Scoping for Autonomous Tools: Service accounts assigned to AI frameworks (such as LangChain, AutoGPT, or internal agent execution engines) should strictly enforce transient, scoped tokens mapped to the specific user session initiating the request, rather than utilizing broad, permanent cloud credentials.
2. Shadow Data & Knowledge Store Mapping: Unstructured knowledge repositories used for vector embeddings often collect unvetted documents containing credentials, API keys, or unredacted enterprise data. Security teams must enforce automated scanning on vector databases and knowledge bases to ensure sensitive records are sanitized prior to ingestion.
3. Automated Risk Isolation: Remediation workflows must operate without human intervention when high-volume data exfiltration or anomalous cross-tenant querying is detected, automatically severing service account access or revoking OAuth tokens to limit operational damage.

Sentra’s funding round reflects a broader industry movement toward securing the data layer directly, ensuring that regardless of how rapidly machine-driven agents operate, organizational governance and least-privilege boundaries remain intact.