OpenCost Introduces MCP Server For AI Costing

The CNCF-hosted OpenCost project published a 2025 year-in-review and 2026 priorities, reporting 11 releases that expanded usability and multi-cloud cost tracking. Notable 2025 work includes a Prometheus-less mode, generic export framework, diagnostics, and an AI-ready MCP server that lets agents query cost data in natural language. The project plans AI-usage cost tracking, supply-chain security, and KubeModel refinement for 2026.
Key Points
- 1Delivered 11 releases adding Prometheus-less operation, generic export framework, diagnostics, and multi-cloud cost tracking
- 2Introduced MCP server enabling AI agents to query cost data in natural language for real-time insights
- 3Plans to track AI usage costs and strengthen cost-data supply chain security for actionable FinOps workflows
Scoring Rationale
Strong product innovation with official CNCF backing, but its impact is concentrated in cloud-native FinOps rather than broader IT.
Sources
Public references used for this report.
Practice with real Ride-Hailing data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Ride-Hailing problems

