# AI Security Startup Mindgard Gains Direct Access to Five Major AI Platforms > Source: > Published: 2026-09-09 14:04:56+00:00 Mindgard, the AI security company built on more than a decade of research out of Lancaster University, has expanded its position inside the technology ecosystems of Anthropic, NVIDIA, Microsoft, Google Cloud and Amazon Web Services. The company confirmed it now holds a place in a program run by each of the five providers, giving it earlier and closer access to the models and infrastructure that enterprise customers are already deploying. The move matters now because AI security vendors are increasingly judged on how quickly they can adapt to new models and agent frameworks, not just on the strength of their existing tooling. Each relationship comes through a different door. [Mindgard](https://www.kobaran.com/tag/Mindgard) is working with Anthropic through its Cyber Verification Program, with NVIDIA through NVIDIA Inception, with Microsoft through the Microsoft Founders Hub, with Google Cloud through its track for high growth AI companies, and with AWS through AWS Activate. Anthropic’s program in particular is notable because it is a vetting process that lets approved security teams use Claude’s full reasoning on dual use defensive work, such as exploit analysis and adversarial simulation, that the model would otherwise decline by default. The announcement lands a few weeks after Mindgard closed a $30 million Series A round led by Album VC, which pushed the company’s total funding to nearly $42 million. That capital is earmarked for scaling product, engineering, sales and marketing as enterprise AI environments grow more distributed, combining multiple models, autonomous agents, cloud services and outside tools in ways that create new and harder to track attack surfaces. ## What Each Partner Program Gives Mindgard The five relationships are not identical. Some are primarily about early technical access, while others are geared toward go to market support or startup resourcing. The table below breaks down what is publicly known about each. | Partner | Program | Primary Focus | |---|---|---| | Anthropic | Cyber Verification Program | Vetted access to Claude’s full reasoning for legitimate offensive-style security research | | NVIDIA | NVIDIA Inception | Support for AI startups building on NVIDIA’s compute and software stack | | Microsoft | Microsoft Founders Hub | Resources and credits for scaling startups on Microsoft’s cloud and AI tools | | Google Cloud | High-growth AI company program | Support track aimed at fast-scaling AI companies on Google Cloud | | Amazon Web Services | AWS Activate | Credits and technical resources for startups building on AWS | ### Why the Anthropic Relationship Stands Out Anthropic’s Cyber Verification Program is a relatively new addition to its ecosystem of trust programs, built specifically for security firms whose legitimate defensive work looks, on its surface, similar to offensive hacking. Vulnerability research, threat modeling and adversarial simulation all require reasoning through how an attacker would operate, which frontier models are designed to resist by default. Acceptance into the program is scoped narrowly to defensive use and remains subject to Anthropic’s security and responsible use requirements. ## The Business Case Behind the Expansion ### A Widening Attack Surface Mindgard’s pitch has consistently centered on the idea that AI risk no longer sits inside a single model or a single cloud. As organizations move past isolated experiments with individual models, their production environments increasingly stitch together several vendors’ tools at once, each with its own behaviors and potential failure modes. “The enterprise AI stack will not belong to one model or cloud,” said Peter Garraghan, founder and chief science officer at Mindgard. “It shall span providers, agents, tools, and data each introducing different behaviors and potential attack paths. Working with these technology ecosystems helps us sharpen our attacker-driven research and translate what we learn into stronger protection for customers.” ### Recent Momentum The ecosystem expansion follows a period of rapid growth for the company. Mindgard has said its research has uncovered more than 150 publicly disclosed security and safety vulnerabilities across widely used AI products, including flaws affecting OpenAI’s ChatGPT, Google’s Antigravity coding environment and the Cursor IDE. That research pipeline is part of what Mindgard says feeds directly into its commercial platform, which performs Shadow AI discovery, AI red teaming and run-time protection for models, agents and applications already in production. #### Background on the Company Mindgard was founded in 2022 by Peter Garraghan and James Brear and is built on AI security research that predates the company by more than a decade at Lancaster University in the United Kingdom. It is headquartered across Boston and London and counts Fortune 2000 enterprises among its customers. The company’s Series A, closed in August, was led by Album VC with participation from Karma Ventures alongside earlier backers .406 Ventures, Atlantic Bridge, IQ Capital and Lakestar. What happens next will depend largely on how these five relationships translate into product. Mindgard has said insights from working closer to the frontier models and infrastructure will feed into new platform capabilities meant to broaden coverage as customer environments keep evolving, though the company has not detailed a specific roadmap or timeline for those releases.