The security of an AI workload depends on more than the applications and models running on it. Beneath every deployment sits an extensive infrastructure environment responsible for processing, networking, storing, and managing the data required by artificial intelligence systems.

As GPU cloud capacity becomes increasingly important to organizations running AI workloads, the security of that underlying environment deserves greater attention. Neocloud providers and AI data center operators must account for vulnerabilities involving hardware, firmware, networking technologies, management interfaces, and shared operational systems.

Lava Security is bringing these issues into focus through its Top 10 Neocloud and AI Data Center Security Risks framework, which has become an official OWASP project.

The initiative was developed by Yakir Kadkoda, CTO and Co-Founder of Lava Security, and Michael Katchinskiy, Head of Security Research. It is designed to help infrastructure providers, operators, and organizations purchasing GPU cloud capacity better understand the security challenges affecting AI environments.

Modern AI data centers combine thousands of GPUs with high-speed networking, specialized management controllers, shared storage, and complex orchestration systems.

Many of these technologies were originally designed for trusted high-performance computing environments. Their growing use in commercial neocloud environments, where unrelated organizations may share physical infrastructure, introduces additional security risks.

Bare-metal neocloud deployments illustrate the issue. Customers may receive root access to rented servers, giving them substantial control over the computing environment. However, if management interfaces are not adequately restricted, that access could allow changes to hardware boot configurations.

Those configurations may persist even after an operating system reinstall. If a server is subsequently reassigned without resetting the affected settings, another customer could inherit a compromised environment.

This is one example of why infrastructure security requires visibility into components that conventional operating system and application protections may not cover.

"AI infrastructure introduces attack surfaces that traditional security approaches may not adequately address. Organizations need to understand what happens below the operating system, particularly when hardware, firmware, and management interfaces are shared across customers," said Michael Katchinskiy, Head of Security Research at Lava Security. "This framework helps make those risks easier to identify and provides practical guidance for addressing them."

A Wider Definition of AI Infrastructure Security

The OWASP framework examines vulnerabilities across multiple layers of the infrastructure stack.

Its coverage includes compromised server firmware, weaknesses in high-speed networking technologies such as InfiniBand and RDMA, inadequate tenant isolation, and insecure out-of-band management interfaces.

It also considers supply chain threats, data center facility systems, shared storage and AI artifacts, incomplete security certifications, centralized operational services, and delays in vulnerability patching.

These categories are intended to provide a broader picture of how infrastructure vulnerabilities can affect an entire environment.

Each category includes potential attack scenarios, real-world implications, and practical mitigation recommendations. Infrastructure providers and operators can use this material to evaluate their environments and prioritize security improvements.

For organizations purchasing GPU cloud capacity, the framework also offers a reference for understanding the controls protecting their workloads.

Built on Research and Practitioner Feedback

According to Lava Security, it initially published the framework in July 2026. Its development drew on disclosed vulnerabilities, incident reports, technical documentation, and the company's research into AI infrastructure security.

More than 20 security and infrastructure practitioners also provided feedback during its development. Their backgrounds included neocloud providers, hardware OEMs, security companies, universities, and research institutions.

The framework's transition to OWASP gives the initiative a structure for continued collaboration, allowing contributors from different parts of the infrastructure ecosystem to participate in its evolution.

Kadkoda believes that this broader perspective is essential to addressing the security challenges affecting AI environments.

"Securing AI infrastructure requires looking at the entire environment, not just individual components. Hardware, networking, management systems, and workloads are interconnected, and risks in one area can affect the security of the broader infrastructure," said Yakir Kadkoda, CTO and Co-Founder of Lava Security. "Developing this framework within OWASP gives the community an opportunity to contribute its expertise and continuously improve how these risks are understood and addressed."

Keeping Up With the Next Generation of Infrastructure

The framework will continue evolving through an open development process on GitHub. Security researchers, infrastructure engineers, hardware vendors, and operators can review its contents, challenge existing assumptions, propose changes, and contribute additional research or mitigation recommendations.

Contributors will receive credit in future releases.

The project is designed to develop alongside the infrastructure itself. New GPU architectures, networking technologies, management systems, and deployment models may introduce additional security considerations that the framework can address through future updates.

Those interested in exploring the initiative can visit the official OWASP project page or access the GitHub repository to participate in its development.

As the infrastructure powering AI continues to evolve, the ability to identify vulnerabilities across the entire technology stack will remain an important consideration for providers and organizations relying on GPU cloud environments. Lava Security's OWASP initiative provides an open foundation for bringing that discussion into a shared industry effort.

This article was published under HackerNoon's Business Blogging program.