EQTY Lab has introduced its new Verifiable Runtime at NVIDIA GTC 2026, unveiling a solution designed to secure and optimize autonomous AI agents running on the NVIDIA Enterprise AI Factory with NVIDIA OpenShell. With AI agents rapidly expanding from local desktop environments into enterprise server infrastructure, organizations now face a growing trust gap. Therefore, EQTY Lab is positioning this launch as a direct response to one of the biggest barriers to scaling the agent economy: proving that AI agents behave securely, transparently, and as intended.
At the core of this new offering, Verifiable Runtime uses EQTY’s Verifiable Compute Framework to deliver hardware-rooted oversight. As a result, enterprises can verify that autonomous agents remain secure and accountable throughout their lifecycle. In addition, the company revealed a broader vision for agent governance during NVIDIA GTC on March 18, when Founder and CEO Jonathan Dotan shared the roadmap behind the technology.
Furthermore, EQTY Lab has extended NVIDIA Confidential Computing (CC) to the NVIDIA BlueField Data Processing Unit (DPU). By combining this with the NVIDIA DOCA Argus advanced threat detection framework, the solution creates a dedicated silicon-based enforcement layer. This architecture gives enterprises a more dependable way to oversee agent activity at scale while maintaining system performance.
More importantly, the Verifiable Runtime delivers hardware-attested execution, which cryptographically proves three core security properties. First, it supports behavior alignment, ensuring agents are continuously audited during runtime. Trust is anchored in the DOCA Argus Service, while the DPU enforces policies from a separate control plane to maintain alignment across the entire agent lifecycle. Second, it establishes delegation chains, clearly showing who authorized each agent action, what credentials the agent holds, and whether its behavior stays within approved guardrails. Third, it protects untampered security postures, ensuring agents cannot alter the policies or access sensitive credentials through native tools.
“EQTY’s Verifiable Runtime gives enterprises the new governance layer that operates entirely outside the agent’s reach. Hardware-attested, cryptographically proven, and enforced at near-zero cost to the system, this new pairing of GPU and DPU is a game-changer for agentic productivity and security,” says Yurko Jaremko, Co-Founder, Head of Verifiable Compute, EQTY Lab
Additionally, EQTY Lab emphasizes that this approach creates an unprecedented framework for deploying AI agents with a dedicated, silicon-based governance plane. Much like Kubernetes standardized container orchestration, Verifiable Runtime aims to establish a predictable and consistent orchestration layer for agent governance. Consequently, providers and developers can gain access to standardized, hardware-rooted interfaces for real-time oversight and lifecycle management. This means governance no longer needs to be treated as a custom engineering burden. Instead, it becomes a built-in capability of the infrastructure itself.
“As enterprises scale AI factories powered by autonomous agents, establishing verifiable trust, governance, and runtime security becomes essential to safely deploy multi-agentic systems across critical workflows,” said Ofir Arkin, Sr. Distinguished Architect for Cybersecurity, NVIDIA. “Built on NVIDIA BlueField DPUs and the NVIDIA DOCA Argus framework, EQTY Lab’s Verifiable Runtime moves agent governance into a dedicated silicon enforcement layer that continuously observes, verifies, and protects AI agent activity.”
Another major advantage of the solution lies in its efficiency. Since the DPU operates with dedicated ARM-based computing resources and direct memory access to the host runtime, enterprises can enforce governance in real time at near-zero cost to the system. Meanwhile, cryptographically attested agent state and behavior traces provide a universal accountability layer, allowing authorized stakeholders to monitor and verify agent actions without adding performance penalties.
EQTY Lab’s AI Gateway serves as the first reference implementation of this framework. Because these verifiable traces are open by design, organizations can build or integrate their own oversight tools and trust infrastructure as the agent economy continues to expand.
Overall, EQTY Lab’s Verifiable Runtime marks an important step in bringing zero-trust principles to the silicon layer. By shifting agent governance into dedicated hardware, the company is helping enterprises replace static permissions with real-time, hardware-rooted enforcement. In doing so, EQTY Lab is turning AI agent governance from a growing operational concern into a strategic advantage.
Recommended Cyber Technology News:
- Rapid7 Enhances Cloud Security in Exposure Command
- Discern Debuts AI-Native Security Platform for Speed
- Nexusguard and Linknet Enterprise Partner to Deliver World Class DDoS Protection Across Indonesia
To participate in our interviews, please write to our CyberTech Media Room at info@intentamplify.com
