AI infrastructure isn’t static—and Mirantis is planning for what comes next. In this clip, Mirantis CTO Shaun O’Meara and VP Randy Bias discuss how their AI Factory Reference Architecture will evolve to support secure, scalable agentic systems in the enterprise.

From Static Infra to Smart Contracts with Agents

“Agentic workloads will need to create contracts with infrastructure,” says O’Meara. These contracts could govern resource guarantees, performance SLAs, and most critically—security boundaries.

As enterprises deploy autonomous agents, these agents will need secure isolation, identity enforcement, and encryption by default. “It’s about ensuring agents only talk to what they’re allowed to,” adds Bias.

Building Policy-Driven Guardrails

Bias points out that compliance, governance, and data privacy are often left behind in fast-moving AI rollouts. Mirantis aims to change that by introducing a policy-enforced agentic management system—on top of the AI Factory’s GPU platform and orchestration layer.

This next layer will offer enterprise-grade guarantees around identity, encrypted communication, and auditable interaction across AI components.

Why It Matters

Mirantis is applying lessons from running highly regulated, mission-critical systems—like global financial infrastructure—to modern AI. The result? A future where intelligent agents are secure by design, infrastructure is composable, and governance is enforced through open standards.