Mindgard, a startup focused on defending artificial intelligence systems, announced a $30 million Series A funding round. With this investment, the company says it is nearing $42 million in total funding to date.
The round was led by Album VC, with additional backing from Karma Ventures and earlier investors including .406 Ventures, Atlantic Bridge, IQ Capital, and Lakestar. Founded in 2022, Mindgard spun out of Lancaster University and is headquartered in London and Boston.
At the center of the news is what Mindgard describes as an automated AI security and red-teaming platform designed to help organizations identify and respond to AI risks. Instead of treating AI risk as something you handle once at deployment, the platform is positioned as a continuous defense layer for AI models, agents, and applications.
What Mindgard says its AI security platform does
Mindgard’s platform focuses on attacking and analyzing AI systems in a structured way, with the goal of finding weaknesses before they can be exploited. The company states that it can capture and exploit the psycho-technical attack surface within AI agents, models, and AI applications.
In practical terms, Mindgard presents itself as a red-teamer that helps organizations identify exploits, assess risk, and ultimately defend AI systems. The platform aims to map an organization’s relevant attack surface, then continuously analyze it over time.
Beyond analysis, Mindgard also claims to provide runtime protection, with the intention of stopping attacks as they occur. This moves the work from purely identifying issues during testing to responding during operation.
Scale built on automated red-teaming
Many security programs struggle to keep pace with how quickly AI tooling evolves. Mindgard’s approach is rooted in automation: the platform is designed to streamline red-teaming activity so that teams can operationalize expertise more consistently.
The company frames its work as more than just running attacks. Mindgard says it turns the knowledge of leading AI security researchers and offensive security practitioners into capabilities that enterprises can use as part of their day-to-day defenses.
CEO James Brear emphasized the idea that AI introduces a fundamentally different kind of risk environment. In his view, the rise of AI systems creates an entirely new attack surface, and organizations need a security method that matches that reality.
Vulnerabilities uncovered across popular AI products
Mindgard also pointed to real-world findings to illustrate what it claims its platform can discover. The company says its solution has already identified over 150 vulnerabilities across widely used AI products.
Among the examples cited are a zero-day code execution flaw in Cursor IDE. Mindgard also reported security issues in Google Antigravity and ChatGPT, according to the company’s announcement.
These disclosures are positioned as evidence that vulnerabilities can exist not only in traditional software components, but also in how AI features behave—especially when systems are designed to follow complex instructions or interact with users in flexible ways.
How Mindgard plans to use the new funding
With the additional capital from this round, Mindgard says it will expand key areas of the business. The company plans to scale its product, engineering, sales, and marketing teams.
The goal is to accelerate deployments and broaden adoption across multiple sectors that are actively evaluating AI while also facing strict security and compliance expectations.
Mindgard specifically mentioned interest in scaling deployments across financial services, digital services, gaming, healthcare, pharmaceutical, and semiconductor sectors. For organizations in these industries, protecting AI workloads can be critical—not only for preventing breaches, but also for maintaining trust and minimizing operational disruption.
Why AI security is evolving toward runtime defense
A key theme in Mindgard’s positioning is the difference between testing and ongoing protection. Many teams run red-team exercises or security assessments to find problems, but AI systems can change in behavior over time as models, prompts, integrations, and workflows evolve.
Mindgard’s description of continuous analysis and runtime protection suggests a shift toward security that operates while AI systems are actively being used. That focus aligns with the idea that attacks may attempt to exploit weaknesses during real interactions, not just during scheduled testing.
By mapping the attack surface and continuously analyzing it, Mindgard is effectively aiming to provide visibility and controls that can adjust as the environment changes.
Automating expertise for enterprise AI risk
Another part of the company’s narrative is operationalization. Security teams often rely on specialized knowledge that can be hard to scale across an entire organization.
Mindgard argues that its platform helps make offensive security know-how actionable at enterprise level. Instead of treating AI security as a one-off project, the company emphasizes turning researcher and practitioner expertise into repeatable capabilities—something that can support ongoing risk management.
This approach is particularly relevant for organizations deploying AI agents or AI-enabled applications where behavior can be influenced by inputs, context, and interaction patterns.
Context: more funding for defensive cybersecurity AI
This announcement follows a broader trend in cybersecurity investment, including other companies working on defensive or runtime-oriented security for AI models. Mindgard’s funding round sits alongside similar market signals that suggest continued investor interest in AI-driven security capabilities.
While each company’s technology differs, the shared direction is clear: security teams are looking for tools that can identify risks and respond faster, especially as AI systems introduce new ways to fail.
Bottom line
Mindgard’s $30 million Series A highlights growing momentum behind practical defenses for AI systems. The company’s AI security platform is built around automated red-teaming, continuous attack-surface analysis, and runtime protection—along with a track record that it claims includes discovering a large number of vulnerabilities.
With plans to scale engineering, product, and go-to-market efforts, Mindgard is positioning itself to accelerate deployments across multiple high-stakes industries where AI adoption needs to be matched with stronger security controls.
Source: https://www.securityweek.com/mindgard-raises-30-million-to-protect-ai-systems/
