Mindgard Raises $30M for Attacker-Aligned AI Security

Mindgard began with the conclusion that traditional application security was looking at the wrong kind of problem. Research led by Dr. Peter Garraghan at Lancaster University had been examining AI security for more than a decadebefore Mindgard was founded in 2022. The challenge was increasingly clear: AI systems could behave probabilistically and opaquely, introducing exploitable risks that conventional AppSec was not designed to understand.

That research has now attracted $30 million in Series A funding, led by Album VC, with participation from Karma Ventures and existing investors .406 Ventures, Atlantic Bridge, IQ Capital and Lakestar. Mindgard will use the capital to expand product, engineering, sales and marketing while growing its international reach.

Mindgard’s approach starts from an unusual premise: better AI defense begins by learning how the system breaks.

The Boston- and London-headquartered company applies offensive-security thinking across AI models, agents and applications. Its platform discovers AI assets and shadow systems, maps potential exposure, runs adversarial red teaming and assessments, then carries intelligence into runtime protection. Instead of treating a model as an isolated component, Mindgard is built to examine the wider AI system around it.

That philosophy is producing a growing body of real-world research. Mindgard says its technology has helped uncover and publicly disclose more than 150 high-impact security and safety vulnerabilities across widely used AI products. Its disclosures include issues affecting Cursor IDE, Google Antigravity, OpenAI products, Gemini CLI, and other AI infrastructure and developer tools.

The disclosures are more than a research scoreboard. Mindgard says intelligence from those discoveries feeds a proprietary knowledge base that continuously strengthens its security platform. Each attack it understands becomes another lesson its defensive system can carry into the next assessment.

Enterprise demand is beginning to follow. Mindgard reports deployments across Fortune 2000 companies and AI innovators spanning financial services, pharmaceuticals, gaming, digital services, semiconductors and healthcare. Its customer material shows the platform being used for continuous red teaming, AI-system assessment and runtime protection in regulated and large-scale environments.

CEO James Brear, who joined Mindgard’s leadership in 2025, is now scaling a company whose roots remain unusually close to the research lab. As enterprises give AI agents more access to tools, data, and operational decisions, security has to understand behaviors that may change faster than a static checklist can follow.

Mindgard’s $30 million bet is that defending AI will require security that learns from attackers almost as quickly as AI itself learns new capabilities.