The OpenCost project, supported by the Cloud Native Computing Foundation (CNCF), recently shared its accomplishments from 2025 and outlined its goals for 2026. Last year, the community rolled out 11 releases that improved usability and functionality. Key updates included running OpenCost without Prometheus using environment variables, the introduction of a beta Collector Datasource for cost data exports, and a diagnostics system that tracks health and exports data. Enhanced multi-cloud cost tracking, aided by contributions from Oracle and DigitalOcean, aimed to make cost transparency more actionable within Kubernetes environments. A standout development was the launch of the OpenCost MCP server, which allows AI agents to query cost data in real-time using natural language. This feature automates the analysis of spending patterns across various Kubernetes resources, enabling teams to generate cost reports and optimization recommendations without manual input. The MCP server has become a default component, making it easier for teams to manage costs while integrating with AI tools for automated financial operations. Looking towards 2026, OpenCost plans to enhance tracking for AI usage costs, especially as machine-learning workloads become more prevalent in cloud environments. There's a strong focus on improving supply chain security for cost data and refining the KubeModel framework to better capture the complexities of Kubernetes resource behavior. Participation in future KubeCon + CloudNativeCon events is also on the agenda, aiming to boost awareness and engagement among cloud-native users. The project's emphasis on standardizing Kubernetes cost reporting and integrating AI tools positions it as a relevant player in managing cloud costs and resource allocation.