Operations | Monitoring | ITSM | DevOps | Cloud

During the Toronto KCD (Kubernetes Community Days), I attended an insightful talk on AI resource optimization that highlighted a staggering Gartner study: “AI infrastructure is adding $401 billion in new spending this year alone. Yet, real-world audits tell a much darker story, revealing that average GPU utilization in the enterprise is stuck at a dismal 5%”. While many people in the audience were shocked by that number, the data didn’t come as a surprise to us.

AI is changing how teams work. Developers are generating code faster, security teams are automating investigations, and employees across the business are using AI tools to accelerate research, content creation, and decision-making. But this adoption comes with a catch. As usage explodes, it introduces a new set of security risks: a rapidly expanding attack surface, faster attack timelines, potential data exposure, and an alarming lack of visibility into how these tools are being used.

You know every service in your cluster by name. You know which team owns each one, what it talks to, how it scales, where its logs go. The agents are a different story. That’s not a criticism, it’s an observation, and it’s one we keep running into. Every company we talk to is shipping agents of some kind, from scales of 10s to 1000s. Customer service bots that field tier-one tickets. Internal copilots that draft emails and summarise meetings and write the boring half of every PR.

The first post in this series argued that AI agent governance hasn’t kept pace with deployment. The second laid out the five pillars of accountability, and what is required. The third walked through why network policies, API gateways, MCP/A2A protocols, DIY security patterns, and Role-based Access Control (RBAC) each leave critical accountability gaps. So what does good look like? The five pillars define what AI agent accountability requires.

Claude Mythos is now available to the public through Claude Fable 5, released June 9, 2026. Claude Fable 5 pricing is $10 per million input tokens and $50 per million output tokens, exactly 2x Claude Opus 4.8 ($5/$25). Claude Mythos 5 (the restricted Project Glasswing version) has identical pricing. Prompt caching cuts input spend by 90%. Batch API pricing is $5/$25 (50% off). In April 2026, Anthropic announced a model it said was too dangerous to release.