The Linux Foundation has joined NVIDIA and leading technology companies to launch the Open Secure AI Alliance, advancing open tools, transparent governance, and stronger security for AI systems worldwide.
The Linux Foundation has announced it is joining NVIDIA and a broad coalition of technology leaders as an inaugural partner in the Open Secure AI Alliance, a new initiative designed to strengthen the security of artificial intelligence (AI) through open collaboration.
The Alliance aims to develop open tools, standards, and techniques that help organisations identify vulnerabilities, improve governance, and build more secure AI systems as adoption accelerates across industries.
According to the Linux Foundation, open source has become the backbone of the modern economy because it allows software to be inspected, improved, and secured by a broad community. The organisation argues the same principles should underpin AI development, particularly as businesses increasingly rely on advanced models and AI agents for everyday operations.
Central to the Linux Foundation’s position is the importance of open models and open weights, which enable organisations to download, inspect, modify, and deploy AI models on their own infrastructure.
The Foundation believes this approach expands access to advanced AI capabilities, promotes greater competition across the technology ecosystem, and allows businesses to retain control over their data, models, and long-term investments rather than becoming dependent on a single provider.
While acknowledging that open models present unique security challenges, the organisation maintains that transparency ultimately creates stronger defences. Rather than relying on closed systems, a wider community of developers and security experts can identify vulnerabilities, benchmark performance, conduct red-team testing, and develop shared safeguards.
The Linux Foundation also highlights the growing importance of AI agent harnesses – the software that enables models to interact with their environment and perform tasks. Making these frameworks open, testable, and auditable is viewed as essential to building trustworthy AI systems.
The Open Secure AI Alliance brings together organisations from across the technology sector, including Adobe, Cisco, Dell Technologies, IBM, Meta, Microsoft, NVIDIA, Red Hat, Salesforce, SAP, Siemens, Snowflake, and many others.
As part of the initiative, NVIDIA is contributing the open source NVIDIA Labs Object-Oriented Agent (NOOA) framework, making it available through GitHub. The research framework is designed to help organisations test, trace, audit, and govern AI agent behaviour more effectively by improving how agent harnesses integrate with AI models.
The Linux Foundation sees this contribution as an example of practical collaboration that goes beyond discussion by delivering working tools that can be adopted across the industry.
The initiative also builds on the Foundation’s existing work through the Open Source Security Foundation (OpenSSF), which develops security tools, best practices, and education programmes to strengthen open source software security.
Commenting on the announcement, Jim Zemlin, CEO of the Linux Foundation, said: “Open source became the backbone of modern computing because it let everyone see, improve, and secure the technology they rely on. AI deserves the same foundation.”
Alongside the Alliance launch, the Linux Foundation is encouraging policymakers to distinguish between transparent, secure AI and opaque systems rather than framing the debate as open versus closed.
The organisation argues that open models and open tooling should be viewed as defensive assets that enable independent evaluation, collaborative security improvements, and faster responses to emerging threats.
While openness alone is not sufficient to guarantee trustworthy AI, the Foundation believes combining transparent development with rigorous testing, secure infrastructure, strong governance, and human oversight provides the most resilient foundation for the next generation of AI innovation.
