Connect with us

Hi, what are you looking for?

SecurityWeekSecurityWeek

Artificial Intelligence

Offensive AI Startup Dreadnode Secures $14M to Stress-Test AI Systems

Dreadnode is building “offensive machine learning” tools to safely simulate how AI models might be exploited in the wild.

Dreadnode, an early stage startup specializing in offensive AI security, has raised $14 million in a funding round from an investment group that includes Decibel, Next Frontier Capital, In-Q-Tel (IQT), Sands Capital, and Indie VC. 

The Dreadnode Series A investment lands amidst heightened interest in the security of AI technologies being deployed at scale around the world and signals a bet on the need for new products aimed at stress-testing artificial intelligence systems for vulnerabilities.

The brainchild of former NVIDIA AI red-team lead Will Pearce and ex-NetSPI VP of Research Nick Landers, Dreadnode styles itself as a pioneer in “offensive machine learning” with tooling for enterprise defenders to safely simulate how cutting-edge AI models might be exploited in the wild.

Dreadnode said the new capital will be used to drive innovation and go-to-market initiatives for two new products — Strikes and Spyglass — that form the core of its platform. 

The company describes Strikes “an AI agent training ground focused on offensive cyber security” that provides “real-world scenarios to test your agent.”. In practice, Strikes is serving up a simulated battleground for AI-driven agents and models, where they can be trained and evaluated against realistic attack scenarios.

The company and its investors believe this approach allows AI developers and red teams to iteratively improve their models’ resilience by repeatedly exposing them to diverse offensive techniques in a safe environment.

Spyglass, on the other hand, is Dreadnode’s product for testing AI systems that have already been deployed. The company calls Spyglass “an AI red teaming product for attacking, evaluating and optimizing AI applications and gaining actionable insights” and said it enables organizations to continuously audit and probe their live AI models (whether language models, computer vision systems, or others) for weaknesses. 

The Spyglass product integrates the latest adversarial machine learning research into practical tests, helping users identify vulnerabilities like prompt injection susceptibility, model bypasses, or data poisoning risks in their AI services. 

Advertisement. Scroll to continue reading.

Together, Dreadnode is positioning Strikes and Spyglass to cover the lifecycle of offensive AI security: from training-phase stress tests to deployment-phase red-team evaluations.

Dreadnode also offers Crucible, an AI hacking sandbox where security practitioners test, learn, and advance their AI red team skills.

 “We’re equipping practitioners with the tools required to explore what’s possible when AI is applied — at scale — to offense,” Pearce said in a note announcing the new cash raise. “We’ve seen what AI is capable of across other domains—why should offensive security be different?”

Related: AI Can Supercharge Productivity, But We Still Need a Human-in-the-Loop

Related: SplxAI Raises $2 Million to Protect AI Chatbot Apps

Related: Ex-GitHub Engineers Raise $20M for Pen-Testing with AI-Powered XBOW

Related: Can AI Early Warning Systems Reboot the Threat Intel Industry?

Written By

Ryan Naraine is Editor-at-Large at SecurityWeek and host of the popular Security Conversations podcast series. He is a security community engagement expert who has built programs at major global brands, including Intel Corp., Bishop Fox and GReAT. Ryan is a founding-director of the Security Tinkerers non-profit, an advisor to early-stage entrepreneurs, and a regular speaker at security conferences around the world.

Trending

Daily Briefing Newsletter

Subscribe to the SecurityWeek Email Briefing to stay informed on the latest threats, trends, and technology, along with insightful columns from industry experts.

Join this event as we dive into threat hunting tools and frameworks, and explore value of threat intelligence data in the defender’s security stack.

Register

Learn how integrating BAS and Automated Penetration Testing empowers security teams to quickly identify and validate threats, enabling prompt response and remediation.

Register

People on the Move

SplxAI, a startup focused on securing AI agents, has announced new CISO Sandy Dunn.

Phillip Miller is joining tax preparation giant H&R Block as VP and CISO.

Linx Security has appointed Sarit Reiner Frumkes as Chief Technology Officer.

More People On The Move

Expert Insights

Daily Briefing Newsletter

Subscribe to the SecurityWeek Email Briefing to stay informed on the latest cybersecurity news, threats, and expert insights. Unsubscribe at any time.