Virtru Data Harbor Platform Launches: AI Agents Get the Data They’re Entitled To, and Nothing More
Data Harbor keeps security policies bound to sensitive data and provides verifiable proof of every access decision.
Press Release Disclaimer: This is a press release distributed through the XPR Media network. It has not been independently verified by our newsroom.

![]()
WASHINGTON, Nov. 17, 2026 (GLOBE NEWSWIRE) — Virtru, the global leader in data-centric security, today unveiled the Virtru Data Harbor Platform at a launch event in Washington, D.C. Built on the open Trusted Data Format (TDF) standard, Data Harbor cryptographically controls how human and non-human identities access sensitive data. With Data Harbor, organizations enforce access policy on the data itself, so people and AI agents can reach only the data they’re entitled to.
AI is accelerating the data control challenge enterprises have grappled with for years. Unlike external attackers, AI agents have authorized access to enterprise systems. They make thousands of requests instantly and operate without continuous human oversight, often accessing more data than intended. A 1Password survey published in July found AI agents regularly access unapproved data at 41% of responding organizations.
In response to this growing challenge, the cyber industry has predominantly focused on adding controls around the agent at the application tier. Nearly 500 companies now sell tools for governing AI agents, and the most crowded categories observe agents, set runtime guardrails, manage agent identities, and route agent traffic through gateways. Data Harbor works alongside these controls and adds a layer that they don’t: policy enforcement on the data itself, which stays in force wherever the data travels.
“Most approaches to governing agents focus on the environment around the data: the endpoint, the network, the access, the app, the model. Those are important control points, but agents cross all of them constantly. The Virtru Data Harbor Platform adds a critical layer. The protection and the policy are bound to the data itself,” said Cole Grolmus, Founder of Strategy of Security.
“Every company has an AI policy. Almost none can actually enforce it, because their agents already hold valid credentials,” said John Ackerly, CEO and co-founder of Virtru. “Data Harbor moves enforcement into the data itself. If an agent isn’t entitled to a file, it stays encrypted, and there is evidence to prove it.”
Verifiable Proof of Every Access Decision
Data Harbor turns every access decision into audit evidence. Each grant and refusal is captured in a signed, tamper-evident record showing who or what requested the data, which policy applied, and the outcome. Auditors and security teams can verify these records independently, so proof of how access policy was enforced doesn’t depend on trusting the system’s own logs. When a partner relationship ends, organizations can revoke access, and previously shared protected files can no longer be opened.
Run It in Your Own Environment, Under Your Own Controls
Organizations choose where Data Harbor runs. Banks, insurers, health systems, and other organizations that prefer to operate their own security infrastructure can build data-centric protection into their applications and AI workflows inside their own environment, under the controls their auditors and regulators already review.
Developers Inherit Data Protection Instead of Building It From Scratch
The Virtru Data Harbor Platform also includes SDKs and APIs that let third-party developers inherit granular policy and cryptographic data-access controls in their own applications. At the launch event, Booz Allen, ThunderCat, Kong, Mattermost, Rocket.Chat, Pexip, and others shared how they are already building on the platform.
“Our customers on air-gapped and classified networks want agents working beside their teams on their most sensitive files, without giving up control,” said Ian Tien, CEO and co-founder of Mattermost. “With Virtru Data Harbor, files in our mission workspace stay protected wherever they go, opening only for those cleared, whether shared with coalition partners or agents.”
Get Started
For developers: Documentation is available for the Virtru Data Harbor Platform and OpenTDF, the underlying open standard. Virtru provides SDKs, training, and support to teams ready to build their own applications on the platform.
For CEOs, CIOs, CISOs, and chief data officers: Leaders with an AI project stalled on a data security question can contact Virtru, and Virtru will work with their teams to get it moving.
You can watch the Virtru Data Harbor Platform launch event in its entirety here.
About Virtru
Virtru is pioneering the shift from network-centric to data-centric security, embedding protection directly into data so mission owners maintain control wherever sensitive information is shared. The Virtru Data Harbor Platform is built on OpenTDF, an open standard evolved from technology developed at the NSA by co-founder Will Ackerly, and supports ACP-240, the Five Eyes-ratified Zero Trust standard for secure coalition operations. Trusted by more than 6,000 public and private-sector organizations, including the U.S. Department of War, JPMorgan Chase, Capital One and Salesforce, Virtru enables secure collaboration across classification boundaries at mission speed, with integrations across leading defense, cloud, and cross-domain solution providers. For more information, visit www.virtru.com.
Contact
Nick Michael
Virtru
nick.michael@virtru.com

Media gallery
