Mindgard raises $30M Series A for AI security testing
Lancaster University spinout Mindgard says its $30 million Series A will expand red-teaming tools, while vendor responses to cited flaws remain unclear.
By Ingrid Halvorsen · Venture Capital Reporter
· 3 min read
Mindgard has raised a $30 million Series A to build out its AI security testing business, a round led by Album VC with Karma Ventures and returning backers .406 Ventures, Atlantic Bridge, IQ Capital and Lakestar. The Mindgard $30M Series A brings the company’s total funding close to $42 million, according to the company and reporting by SecurityWeek, but its valuation, revenue, customer count and headcount were not disclosed.
The Boston- and London-headquartered company, founded in 2022 from AI-security research at Lancaster University, says it will spend the new capital on product, engineering, sales and marketing. Its pitch is to test enterprise AI deployments as an attacker would, as companies put models and agents into systems connected to data and software tools.
What does Mindgard’s AI security platform do?
Mindgard describes its product as covering three jobs across AI models, agents and applications: finding unsanctioned or untracked “shadow AI” use, red-teaming systems for weaknesses, and providing runtime protection once they are deployed. In practice, red-teaming means deliberately probing an AI system with adversarial inputs and workflows to identify behavior that could expose data or trigger unintended actions.
The company says the platform draws on more than a decade of research associated with Lancaster University and converts findings by offensive-security researchers into testing capabilities for enterprise security teams. That is a commercial claim, rather than an independently measured assessment of its effectiveness.
What is known about the Cursor, Google and ChatGPT findings?
Mindgard says it has publicly disclosed more than 150 security and safety vulnerabilities in AI products. Its cited examples include a zero-day code-execution issue in Cursor IDE, a trusted-workspace flaw in Google Antigravity, and failures of ChatGPT’s image-generation guardrails.
Those findings are central to the company’s marketing case, but the available reporting does not establish their technical details, affected versions, discovery or disclosure dates, remediation status, real-world exploitation, or user impact. It also contains no responses from Cursor, Google or OpenAI confirming the reports. The financing announcement is therefore distinct from evidence that any of the named products currently remains vulnerable.
Mindgard said it has deployments or is pursuing deployments in financial services, digital services, gaming, healthcare, pharmaceuticals and semiconductors. The company did not disclose the number of customers or any contract or revenue figures.
Why the financing matters
The round gives a research-led security startup capital to turn vulnerability discovery into a repeatable enterprise product. The investment case rests on whether customers will buy continuous testing and protection for AI systems rather than treating model security as a one-off pre-launch exercise.
Album VC partner Ty Boswell said organizations are putting AI into critical operations without security infrastructure designed for how such systems operate. That is an investor view, but it identifies the market Mindgard is targeting: AI applications create potential attack paths through model behavior, prompts and connected tools that conventional application testing may not cover.
This story draws on original reporting from TechFundingNews.
