CloudBolt Connects Infrastructure, AI and FinOps
Alan Shimel talks with Yasmin Rajabi, Chief Operating Officer at CloudBolt, about the changing role of infrastructure management as enterprises adopt Kubernetes, AI tools and cloud cost controls. Rajabi explains that CloudBolt helps organizations manage infrastructure across their cloud journey, from traditional VMs and VMware migrations to Kubernetes platforms and AI-enabled environments.

The conversation focuses on AI infrastructure management as companies give developers and teams more ways to provision resources. Natural language tools and MCP integrations can make infrastructure easier to access, but they also create new governance challenges. Rajabi says enterprises still need auditability, RBAC, security controls and clear visibility into what teams are using.
Kubernetes Cost Allocation Requires Accuracy
Rajabi explains why basic showback is not enough when organizations need true chargeback. Splitting cloud costs evenly across teams or namespaces may offer a rough estimate, but it does not build trust. If finance and engineering teams see different numbers, even small inaccuracies can turn cost allocation into a political problem.

CloudBolt has had to solve this challenge internally because its own platform runs on Kubernetes. Rajabi says accurate allocation requires real billing data, discounts, savings plans and granular per-container usage. Without that level of accuracy, teams may question the bill and resist taking ownership of optimization work.
AI Adds a New Layer of Cost Complexity
The discussion also explores how AI infrastructure management is making cost visibility harder. GPU usage can be expensive and difficult to divide efficiently. AI agents can also trigger many sub-agents, each using different models for different amounts of time. That makes it harder to map usage back to a specific team, workflow or business outcome.

Rajabi says CloudBolt is working on more granular cost allocation by pulling usage data from public cloud billing standards and mapping it back to application activity. The goal is to help teams understand which models, agents and workloads are driving spend. That insight can help organizations decide when to use premium models, when to delegate to cheaper models and where optimization makes sense.
Governance Becomes a Team Discipline
The episode closes with a practical look at how companies are using AI inside their own teams. Rajabi notes that adoption is not just a technology problem. It also requires enablement, trust and new habits across engineering, product, sales and operations teams.

For IT and business leaders, the message is clear. Kubernetes, AI and cloud infrastructure are becoming more connected, and cost accountability needs to keep up. CloudBolt is positioning AI infrastructure management as a way to help enterprises govern that complexity while still giving teams room to innovate.