Security research often begins with identifying a vulnerability, understanding how it could be exploited, and determining what organizations can do to reduce exposure. But when the technology involved spans hardware manufacturers, infrastructure providers, operators, and customers, turning research into consistent industry practices can become considerably more complicated.
That challenge is becoming increasingly relevant in artificial intelligence infrastructure. Modern AI data centers depend on interconnected systems, including GPUs, specialized management controllers, high-speed networking, shared storage, and orchestration technologies. As these environments expand into commercial neocloud deployments, the security implications are becoming more difficult to address through conventional approaches alone.
Lava Security is taking its research into this area to a broader industry audience. The company announced that its Top 10 Neocloud and AI Data Center Security Risks framework has become an official OWASP project, creating an open initiative for the continued examination of vulnerabilities affecting AI infrastructure.
The framework was developed by Yakir Kadkoda, CTO and Co-Founder of Lava Security, and Michael Katchinskiy, Head of Security Research.
From Research to an Industry Framework
Initially published in July 2026, the framework draws on disclosed vulnerabilities, incident reports, technical documentation, and Lava Security’s research into AI infrastructure security.
Its development also incorporated feedback from more than 20 security and infrastructure practitioners representing neocloud providers, hardware OEMs, security companies, universities, and research institutions.
The resulting framework examines risks across the hardware, firmware, networking, management, and operational layers of AI infrastructure.
Its coverage includes compromised server firmware, vulnerabilities in InfiniBand and RDMA networking technologies, inadequate tenant isolation, insecure out-of-band management interfaces, supply chain threats, and data center facility systems.
Other categories address shared storage and AI artifacts, incomplete security certifications, centralized operational services, and delays in vulnerability patching.
The framework provides potential attack scenarios, real-world implications, and practical mitigation recommendations for each category.
Rather than focusing exclusively on individual technologies, it considers how weaknesses across interconnected infrastructure components can affect entire environments.
Why a Shared Reference Is Needed
AI infrastructure introduces security considerations that are not always covered by traditional approaches. One example involves bare-metal neocloud environments, where customers may receive root access to rented servers.
If management interfaces are not adequately protected, a tenant could modify hardware boot configurations that persist beyond an operating system reinstall. Unless those settings are reset before the server is reassigned, the next customer could inherit a compromised environment.
This illustrates the importance of understanding how infrastructure components behave beyond the software layer.
“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.”
The OWASP transition provides a structure for making the research accessible to a wider community and encouraging continued contributions from the people responsible for building, operating, and securing these environments.
The Role of Industry Collaboration
The security of AI infrastructure involves stakeholders with different areas of expertise. Hardware manufacturers develop the technologies powering servers and networking systems, while neocloud providers and data center operators manage the environments in which those technologies are deployed.
Security researchers and infrastructure engineers contribute additional perspectives on vulnerabilities, attack scenarios, and mitigation strategies.
Bringing these groups into a shared initiative can help establish a more consistent understanding of the risks involved.
“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.”
An Open Process for Future Development
As an OWASP project, the framework will continue evolving through an open development process on GitHub.
Security researchers, infrastructure engineers, hardware vendors, and operators can review its content, challenge existing assumptions, propose changes, and contribute additional research or mitigation recommendations. Contributors will be credited in future releases.
The framework is also intended to evolve alongside changes in AI infrastructure, including new GPU architectures, networking technologies, management systems, and deployment models.
Those interested in participating can visit the official OWASP project page and access the GitHub repository to review the initiative and contribute to future updates.
For Lava Security, the transition represents an opportunity to turn research into a continuing community effort, giving the AI infrastructure industry a common reference for understanding and addressing emerging security risks.