GitGuardian Doubles Down on AI Agent Defense With $50M Raise

A non-human identity security platform raised $50 million to detect, remediate and govern secrets across modern enterprise environments.

See Also: Proof of Concept: Machine Identities Fuel Rising IAM Crisis

The Insight Partners-led Series C funding round will help New York-based GitGuardian to address the growing risk associated with non-human identities and secrets as artificial intelligence agents spread across enterprises, said co-founder and CEO Eric Fourrier. He said the dramatic growth in non-human identities has led to more organizations exposing sensitive data through credential misuse and secret sprawl.

“If you want to get ROI from AI agents and really want them to do great work, they actually need to have access to data,” Fourrier told Information Security Media Group. “And to get access to data, you need to provide them with non-human identity and secrets.”

GitGuardian, founded in 2017, employs 176 people and has raised $106 million, having last completed a $44 million Series B funding round in December 2021 led by Eurazeo. The company has been led since inception by Fourrier, who was previously a co-founder at Quantiops, which applies data-driven strategies to solve business problems (see: How to Snare Software Supply Chain Hackers With Honeytokens).

Why Secrets Leakage Extends Beyond Source Code

What was once a ratio of roughly 10 non-human identities per human has now expanded to potentially 100-to-1 through the rapid proliferation of AI agents and automation systems, Fourrier said. This explosion of credentials, tokens and machine identities has created an exponentially larger attack surface, Fourrier said.

“The problem is getting really huge, and we actually need to solve it,” Fourrier said.

AI and LLMs are evolving rapidly with new protocols, agent frameworks and integration patterns emerging continuously, and Fourrier said GitGuardian must stay technologically agile to keep pace with new systems where secrets may be exposed. He said GitGuardian plans to invest heavily in product engineering to remain competitive in a rapidly evolving environment shaped by AI advancements.

“The AI and LLMs are moving very quickly with new releases from opening up big new protocols,” Fourrier said. “We had MCP protocol last year, and now we have a swarm of agents. So we have more and more things to secure. So we need to stay very dynamic and very innovative.”

GitGuardian originally focused on spotting hardcoded credentials in repositories and helping developers correct them, but over time recognized that secrets leakage extends far beyond source code, Fourrier said. Firms leak credentials across collaboration tools like Microsoft Teams and Slack, ticketing systems like JIRA, documentation platforms like Google Drive and Confluence, and no- automation platforms.

“Detection without remediation is just noise,” Fourrier said. “So, we need to make sure our customers can fix the incident. We still have a big, big part of our job to do here.”

How GitGuardian Can Help Reduce Over-Privileging

Organizations need full visibility into both secrets shared in Slack or embedded in code and storage locations like vaults, and Fourrier said governance involves understanding how these credentials are used. Many credentials grant full administrative access to cloud infrastructure or CRM systems, and Fourrier said GitGuardian seeks to introduce policies and controls that reduce over-privileging.

“You cannot secure what you don’t see,” Fourrier said. “So, we really need to be able to have a full view of the non-human in the secrets. You need also the visibility and, ‘Okay, how do you use credential the right way?'”

AI agents function as autonomous systems that access data, execute actions and operate with delegated privileges, and Fourrier said they resemble workloads, service accounts or automation scripts with broader and more dynamic permissions. With AI agents integrated into workflows, attackers can exploit agents through malicious injections, compromised dependencies or by extracting secrets, he said.

“Attackers are looking for the easiest path to damage companies, and definitely stealing and abusing secrets is one of the easiest ways to attack and hack a company, steal some data and move laterally,” Fourrier said.

Competitive pressure primarily comes from Microsoft’s GitHub Advanced Security and Wiz since both companies operate across broader security domains and may include secret scanning features within larger suites, Fourrier said. GitGuardian identifies leaked or misused secrets, while vaulting solutions including CyberArk and HashiCorp help securely store credentials, making them partners rather than competitors, he said.

“GitHub advanced security is working on securing cloud and securing the code,” Fourrier said. “We don’t see a lot of competition from early players. It’s mostly the big guys.”