Higress Joins CNCF: Delivering an enterprise-grade AI gateway and a seamless path from Nginx Ingress

What is Higress? Customer adoption Why CNCF? Roadmap: Long-term compatibility and ecosystem evolution Sandbox is just the beginning Posted on March 25, 2026 by epower CNCF projects highlighted in this post We are thrilled to announce that Higress has officially passed the Technical Oversight Committee (TOC) vote to join the Cloud Native Computing Foundation (CNCF) as a Sandbox project, becoming a proud member of the CNCF ecosystem. We are thrilled to announce that Higress has officially passed the Technical Oversight Committee (TOC) vote to join the Cloud Native Computing Foundation (CNCF) as a Sandbox project, becoming a proud member of the CNCF ecosystem. Higress is an AI-native, high-performance API gateway built on Envoy and Istio. It conceptually unifies the traffic gateway, microservices gateway, and AI gateway into a single control plane, drastically reducing operational complexity for both cloud-native and AI workloads. Born out of Alibaba’s massive-scale internal production environments, Higress brings enterprise-grade stability and WebAssembly (Wasm) extensibility to the CNCF ecosystem. Today, its core capabilities are built upon two main pillars: Ingress controller and gateway capabilities: Higress serves as a highly mature Kubernetes Ingress Controller, offering full support for the Gateway API and its Inference Extension. With Nginx Ingress scheduled for retirement in 2026[1], Higress provides a secure, drop-in replacement for Nginx Ingress. It remains fully compatible with mainstream Nginx Ingress annotations while replacing the vulnerable configuration-injection model with a robust xDS control plane and Wasm sandbox, effectively eliminating the security risks inherent to legacy architectures. Whether organizations choose to maintain existing Ingress setups or migrate fully to the Gateway API, Higress delivers unified, highly scalable traffic governance. AI-native gateway capabilities: Higress treats AI traffic as a first-class citizen. It provides native support for Large Language Model (LLM) invocations, the Model Context Protocol (MCP), and diverse AI inference scenarios. By offering token-based rate limiting, multi-model fallback, Retrieval-Augmented Generation (RAG) integration, model-aware routing, and intelligent load balancing, Higress standardizes how cloud-native applications consume LLMs.