AI Agent Data Security: Sentra Raises $150 Million

Sentra, a data security posture management (DSPM) company, raised $150 million in a strategic investment led by Capital One Ventures, it announced on November 4, 2026, to extend AI agent data security across enterprise systems.

Bessemer Venture Partners, Zeev Ventures, Standard Investments, Munich Re Ventures and Key1 Capital returned as investors. Capital One will take a seat on Sentra’s board.

For compliance and risk teams at banks and large companies, the round matters because of what it points to: autonomous software that reads regulated data under its own identity. Each AI agent is a new identity that pulls data continuously, so its data security is a control that auditors, legal teams and boards will ask about. Sentra’s product is data security tooling rather than compliance software, but the evidence it produces about who and what could reach sensitive records is the kind compliance functions need.

Sentra said its growth in the past year exceeded 300% compared with the year before. It has not disclosed a valuation or a customer count in the announcement. It also did not name any bank customers; Capital One is named as the lead investor, not as a customer.

What the $150 million round funds

Sentra said it will use the money to broaden visibility into agents, copilots and models, maintain continuous least privilege and speed up containment when something goes wrong. It will also scale its field and channel programs, invest in brand and marketing, and expand into new regions.

The round is the company’s largest. Sentra raised a $30 million Series A in January 2023 and a $50 million Series B led by Key1 Capital on April 22, 2025. At the time of the Series B, its total funding passed $100 million.

Sentra was founded in 2021 by Asaf Kochan, Yoav Regev, Ron Reiter and Yair Cohan. Regev is co-founder and chief executive.

“AI agents are now the busiest consumers of enterprise data, and many of them inherit access to data that hasn’t been fully classified or governed.”

Yoav Regev, Co-founder and CEO, Sentra

How AI agent data security works at Sentra

Sentra argues that data security was designed around human users. It says three pressures have converged. Businesses are pushing copilots, assistants and cloud AI services into production faster than governance frameworks can adapt. Sensitive records sit across cloud, SaaS and on-premises systems, which leaves older visibility tools unable to track where the data is or who can reach it. And each agent adds an identity with no central point of control.

The company’s approach focuses on data at rest rather than inspecting each agent request in real time. It works in five steps:

Before deployment, the platform locates and classifies sensitive data.
It maps which people, service accounts and agents can reach that data.
It enforces least privilege for non-human identities, meaning each identity keeps only the access it needs.
It deletes obsolete or high-risk data.
When an agent behaves unexpectedly, data-aware detection flags the activity and automated remediation revokes access, escalating serious issues to data owners.

Afterward, Sentra reconstructs what an agent could have accessed. The company says this gives security, legal and compliance teams evidence they can defend.

Sentra’s analysis runs inside the customer’s own environment, so data is never copied out. The company says this avoids the retention, residency and compliance burdens that come with rival tools that replicate customer data. It adds that the design lets it scale across clouds, SaaS, warehouses and on-premises systems at lower cost.

Agent controls announced in April 2025

The funding builds on a product Sentra launched on April 28, 2025, called Data Security for AI Agents. It automatically finds AI agents and the models and knowledge bases they connect to, and shows the sensitive data they can reach and the users who interact with them.

The product labels files in data sources before agents use them. It applies identity-based access rules so that an agent operates within the roles and permissions of the user it serves, which is meant to stop AI-generated responses from revealing information that user is not authorized to see.

Continuous monitoring flags unauthorized access or possible leakage and offers remediation recommendations. “AI Data Exposure Insights” show who accessed what through AI interactions, including the AI-generated responses, to support investigations. The launch covered Microsoft Copilot Studio, Amazon Bedrock and OpenAI ChatGPT Enterprise.

What it means for compliance teams

Neither the funding announcement nor the product launch names GDPR, DORA or the EU AI Act. Both refer only to compliance and regulatory compliance in general terms. Any link to a specific rule is therefore the buyer’s to make, not Sentra’s.

The connection is practical. Data protection rules and operational resilience rules both rest on knowing where sensitive data sits and who can reach it. An access map that includes service accounts and agents, and a record of what an agent could have read after an incident, bear directly on that. So does the board reporting that follows an incident involving automated systems.

The same data-exposure question drove the fallout from the Dentsu breach, where employee bank, payroll and social security data was exposed. For AI agent data security, the risk is higher because agents add speed and volume and act continuously rather than when a person logs in.

The wider field for AI agent data security

Sentra competes with other DSPM vendors and with a growing group of startups that govern what agents can reach. Cygrid raised €2.66 million to control AI agent data access, and Opsin raised $21 million for enterprise AI agent protection. Their products differ in where they sit, but they target the same gap between agent deployment and governance.

Sentra’s pitch separates it from tools that sit in the request path. By securing data before an agent reaches it, the company puts the control in the data layer. That design suits teams that already hold data inventories and classification schemes, and it connects to the golden source data discipline that financial institutions apply to regulatory reporting.

Capital One’s role is notable because it is a large US bank taking a board seat at a vendor that sells AI agent data security. The announcement does not say whether Capital One uses Sentra’s platform.

What comes next

Sentra said the new funding will go first to agent, copilot and model visibility and to expansion into new regions. Its customers will test the evidence claims when regulators or auditors ask them to show what an agent could access. The next marker is whether the company discloses named financial services customers and the rules they map the product to as the regional expansion begins.