OPAQUE Joins the Linux Foundation's Appia Foundation and Agentic AI Foundations to Advance Open Standards for Verifiable AI

The memberships bring OPAQUE's open AI trust standards to the industry efforts shaping how AI is verified, and autonomous agents are governed

SAN FRANCISCO, July 16, 2026 /PRNewswire/ -- OPAQUE, the Confidential AI company, today announced it has joined two Linux Foundation initiatives, the recently announced Appia Foundation and the Agentic AI Foundation (AAIF). Through these memberships, OPAQUE aims to contribute its work on AI identity, runtime governance, and cryptographic evidence to the open standards shaping how trustworthy AI systems are assessed and autonomous agents are governed.

The announcement comes on the heels of the Linux Foundation's and OPAQUE's Confidential Computing Summit, where leaders from across the AI ecosystem echoed the same message: the next phase of enterprise and sovereign AI will require organizations to prove, not simply claim, how AI systems are governed. As AI agents become more autonomous and regulatory expectations continue to evolve, industry standards for identity, governance, and verifiable runtime evidence are becoming foundational infrastructure. Reflecting that shift, the Linux Foundation launched the Appia Foundation, expanded the Agentic AI Foundation to more than 200 members, and introduced new open AI infrastructure projects, signaling a broader industry move from building more capable AI systems to building the trust infrastructure those systems require. Together, these efforts signal that the industry is beginning to standardize how trust is defined, implemented, and independently verified, much as earlier generations of the internet standardized security and interoperability.

The two foundations address different parts of that challenge. The Appia Foundation delivers an open connecting layer providing testing criteria, evaluation guidelines, and component typologies to verify and audit safe, trusted AI models, systems, and applications. The AAIF provides a neutral, open foundation to ensure agentic AI evolves transparently and collaboratively.

Today, enterprises lack a standard way to independently verify how AI systems actually behave once they're running. OPAQUE's work helps close that gap by contributing cryptographic identity, runtime policy enforcement, and verifiable evidence that is hardware-enforced with a silicon encryption key built in NVIDIA, AMD, and others. This lets organizations independently prove, not simply trust, that AI systems operated according to policy.

"AI is becoming part of the world's critical infrastructure, and critical infrastructure can't run on trust alone, it needs proof," said Aaron Fulkerson, CEO of OPAQUE. "The challenge of obtaining that proof shows up concretely today: enterprises buying AI have no standardized way to verify how a model or agent actually behaved short of trusting the hyperscaler's word for it, and no way to run models independently of the handful of providers currently gatekeeping access to them. That's why we're putting our work on agent identity and runtime governance into the open, through the Appia Foundation and Agentic AI Foundation. We're giving enterprises, regulators, and auditors the ability to independently verify how AI systems behave."

OPAQUE's contributions are built on the open-source Agent Governance Toolkit (AGT), created by OPAQUE Chief Platform Officer Imran Siddique, formerly Agentic AI Architect and Engineering Leader at Microsoft Azure. With close to 5,000 GitHub stars and over 100 contributors, AGT has grown into something larger than a single project: it functions as a lab, a proving ground where focused new standards are incubated, hardened alongside early adopters, and matured until they can stand on their own. OPAQUE aims to bring the standards that graduate from that work to the AAIF, supporting new agentic governance standards flowing into the AAIF.

Through AGT, OPAQUE aims to contribute several emerging specifications to the AAIF, including:

• Agent Manifest gives every AI agent a verifiable identity.
• TRACE (Trust Runtime Attestation and Compliance Evidence), an open specification for hardware-attested trust records that provide cryptographic evidence of AI execution and policy enforcement.
• Confidential MCP (cMCP) extends the Model Context Protocol (MCP) with hardware-enforced policy and verifiable runtime evidence.
• Confidential A2A (cA2A) enables verifiable identity and provenance across communication between AI agents.

OPAQUE is developing these specifications in the open with partners across the ecosystem. Key thoughts from TRACE have already been contributed through CoSAI's Working Group 4, while Agent Manifest, Confidential MCP, and Confidential A2A are expected to be contributed to the AAIF in the coming months as they mature into vendor-neutral community standards.

"Verifiable trust in AI cannot be something any one company owns. It has to be an open standard the whole industry can build on and independently check. I created the Agent Governance Toolkit to give teams a common language for what an agent is allowed to do, and Agent Manifest, TRACE, Confidential MCP, and cA2A extend that from written policy into cryptographic proof. Bringing this work into the Appia Foundation and Agentic AI Foundation is exactly where it belongs: a neutral home where enterprises, regulators, hardware makers, and model providers shape it together. I've been contributing to both communities, and joining as a member is our commitment to build this in the open, with everyone, not around anyone," said Imran Siddique, Chief Platform Officer, OPAQUE.

This news builds on OPAQUE's recent momentum with the launch of OPAQUE 3.0 and longstanding collaboration with the Linux Foundation. OPAQUE is a member of the Confidential Computing Consortium and founded the Confidential Computing Summit, now co-hosted with the Linux Foundation. Alongside founding contributors, including the Technology Innovation Institute (TII), and a growing community of open-source collaborators, OPAQUE is helping advance the standards that underpin trustworthy AI. Joining the Appia Foundation and the AAIF extends that collaboration from confidential computing into the open standards shaping the next generation of enterprise AI.

Availability: The Agent Governance Toolkit, Agent Manifest, TRACE, Confidential MCP, and Confidential A2A are available today as open-source projects at github.com/agentrust-io.

About OPAQUE

OPAQUE is the Confidential AI company. Born from UC Berkeley's RISELab (now the Sky Compute Lab), OPAQUE lets organizations run AI models, agents, and workflows on their most sensitive data with hardware-rooted isolation and verifiable evidence that approved governance policies were actually enforced. Founded by Dr. Ion Stoica (co-founder of Databricks; co-director, UC Berkeley Sky Compute Lab), Dr. Raluca Ada Popa (ACM Grace Hopper Award winner; Senior Staff Research Scientist at Google DeepMind, where she leads AGI security research), and Rishabh Poddar (CTO); Imran Siddique, creator of the open-source Agent Governance Toolkit (AGT), is Chief Platform Officer. OPAQUE created the Confidential Computing Summit, now co-hosted with the Linux Foundation.

Media contact: [email protected] and [email protected]

SOURCE OPAQUE