Open source Envoy AI Gateway has reached production-stable v1.0 with contributions from Bloomberg, Nutanix and Tetrate, bringing vendor-neutral routing, governance and observability to enterprise AI workloads while advancing an open standard for AI traffic.
Envoy AI Gateway has reached production-stable v1.0, marking a major milestone for what is described as the first open-source AI gateway built on the CNCF Envoy Gateway project. Backed by Bloomberg, Nutanix, Tetrate and the broader Envoy community, the release aims to establish an open standard for enterprise AI traffic after 16 months of collaborative development.
Designed to extend Envoy’s proven internet-scale traffic management capabilities to AI workloads, Envoy AI Gateway delivers AI traffic routing, governance, observability, extensibility and compliance controls while promoting vendor neutrality, open governance and enterprise-grade reliability. Notably, the same code available in the public repository is already running in production environments.
Bloomberg initiated the effort through the Envoy Gateway community and currently runs Envoy AI Gateway in production. Nutanix is taking the platform into production and integrating it into Nutanix Agent Gateway and Nutanix Enterprise AI, while LY Corporation uses it to manage multi-tenant self-hosted LLM traffic.
“We see the Envoy AI Gateway as a key element toward standardising how enterprises securely and reliably serve AI workloads,” said Dan Sun, Envoy AI Gateway and KServe co-founder/maintainer.
New v1.0 capabilities include a unified API across major AI providers, a native MCP Gateway, token-aware traffic management, centralised credential management, AI-native observability and a stable v1 API surface. Future development will focus on spend-based governance, deeper MCP security and broader alignment with the agentic AI ecosystem.
