Startup Sentra Nabs $50M to Boost AI-Powered Data Protection (2025)

Data Governance , Data Loss Prevention (DLP) , Data Security

A startup founded by an Israeli Military Intelligence colonel raised $50 million to help enterprises secure and capitalize on their data.

New York-based Sentra plans to use the Series B proceeds to build capabilities in enforcement, add native remediation features and label sensitive data to inform access management for AI platforms such as Copilot, according to co-founder and CEO Yoav Regev. He said Sentra's data security platform is built to safeguard structured and unstructured data across IaaS, PaaS, SaaS and on-premises environments.

"We want to have enough money to push it as fast as we can and to accelerate, to be boots on the ground, to be everywhere, to serve our customers, to see the opportunity and to give them the best solution," Regev told Information Security Media Group. "We found that this number they gave us is a good number for the next two years, three years, to push everything at the right pace."

The Importance of Accurately Classifying Data

Sentra, founded in 2021, employs 131 people and has raised $103 million in three rounds of outside funding, with Standard Investments leading the firm's most recent $30 million Series A funding round in January 2023. The company has been led since its inception by Regev, who spent nearly 24 years in Israel's Unit 8200, culminating in a stint as the head of its cyberdepartment.

Enterprises are under more pressure than ever to protect sensitive information due to a relentless stream of data breaches and expanding regulatory requirements, Regev said. At the same time, companies are eager to monetize their data by adopting technologies such as AI, which requires deep, accurate visibility into data assets. Sentra's Series B funding round was led by Key1 Capital.

Regev wants Sentra to maintain high classification accuracy, extending coverage across IaaS, PaaS, SaaS and on-prem environments and enhancing remediation and enforcement. These components are not isolated features but interlinked parts of a larger system that provides data visibility, contextual intelligence and actionable insights, he said. He wants to shift from observing and classifying data to automating responses.

"Classification is a core capability for a data security platform," Regev said. "We want to support every platform and every place the customers have data - from IaaS, PaaS, SaaS and on-prem. We are taking our platform to the next level, from the remediation and enforcement part - DLP, permissions. And of course AI - we keep inputting more and more AI capabilities to support AI use cases for the customers."

Achieving more than 95% classification accuracy across structured and unstructured data without removing that data from the customer’s environment is made possible by Sentra’s use of AI and LLMs, according to Regev. The classification engine must continuously evolve to support new data types, more platforms and emerging use cases such as AI governance, Regev said.

"To do 95% highly accurate classification without taking any piece of data outside of the customer environment, this is very unique," Regev said. "Most other capabilities and companies cannot meet those standards."

How Sentra Approaches Securing On-Prem Settings, AI Models

While cloud platforms offer standardized APIs and unified architecture, on-prem environments are inconsistent and outdated, requiring entirely separate engineering efforts, according to Regev. As such, Regev said Sentra is investing in two parallel development paths, one for cloud and one for on-prem. In the long term, Sentra is aiming for 99% coverage of where enterprise data lives, Regev said.

"You can support each place of on-prem, but to do it in a very efficient and accurate way, it's kind of a different branch," Regev said. "There is one branch from the cloud and one branch from the on-prem. This is kind of different technology; you have to put to double the efforts to support both."

Sentra wants to control which data enters AI models, manage permissions and access governance for AI tools, and enable data labeling for tools such as Microsoft Copilot and AWS Bedrock to enforce access policies, Regev said. Sentra is stepping in to address accidental exposure of sensitive data by offering context-aware classification and enforcement that travel with the data as it flows into AI environments.

"We just started to see AI use cases," Regev said. "It's pretty new, six months, 12 months, something like that. I couldn't tell you a year ago those are the major AI use cases. So we want to support those three for almost every service that's important to our customers."

Sentra will develop native enforcement capabilities, allowing the platform itself to take intelligent action. This includes permission reduction, encryption, masking and even data deletion, according to Regev. Over time, Sentra will evolve into a fully autonomous system that not only detects data risk but actively remediates it, closing the loop on security, Regev said.

"Now, we push more and more to act by ourselves, to help you mask the data, delete the data, encrypt the data, reduce encryption," Regev said. "All those actions that we can do on top of your data by ourselves. This is what we push in the next couple of months based on that investment."