Nutanix adds MCP gateway and Kubernetes tools for agentic AI
Nutanix has announced general availability of Nutanix Enterprise AI (NAI) 2.8 and previewed the forthcoming Nutanix Kubernetes Platform (NKP) 2.19, two releases designed to let enterprises deploy and govern production agentic AI without rearchitecting the infrastructure that runs their existing applications and data.
The NASDAQ-listed hybrid cloud vendor says the updates extend its Nutanix Cloud Platform (NCP) with what it calls a dual-native architecture, treating virtual machines and containers as equally supported infrastructure rather than forcing workloads into a single model. Nutanix argues this lets organisations bring AI to where their data already lives, reducing the need to create new infrastructure silos or move data across environments.
What is shipping
NAI 2.8 is generally available now. Its headline addition is a generally available Model Context Protocol (MCP) Gateway within the Nutanix Agent Gateway, providing a standardised connection layer between AI agents and enterprise tools and data. The release also introduces multi-GPU inference via tensor parallelism for large language model serving, speculative decoding that Nutanix says can accelerate token generation by up to 2.5x using lightweight draft models, and parameter-efficient fine-tuning via Low-Rank Adaptation (LoRA) for models below eight billion parameters. Enhanced identity and access management, custom roles, and support for air-gapped NVIDIA NIM deployments round out the security additions.
NKP 2.19 does not yet have a firm release date. The forthcoming release will introduce NKP Metal, which aims to bring hyper-converged simplicity to bare-metal Kubernetes deployments with automated OS and firmware provisioning, and an AI Applications Catalog offering one-click deployment for validated open-source AI and machine learning tools including Kubeflow, Milvus and Slurm. NKP has also attained CNCF Certified Kubernetes AI Conformant Platform status, a formal validation of its standards compliance for enterprise AI workloads. Nutanix Unified Storage has separately received NVIDIA-Certified Storage validation at the enterprise level.
"Enterprise AI should not require customers to rebuild the systems that already run their business," said Thomas Cornely, Executive Vice President, Product Management at Nutanix. "This gives organisations a practical path to production without creating new silos or limiting future choice."
Market context
The releases land in a competitive segment where Nutanix faces pressure from VMware (now under Broadcom), Red Hat OpenShift, and a range of cloud-native platform vendors. A Gartner forecast cited in the release projects that 55 percent of enterprises will migrate all workloads away from VMware by 2029, a transition Nutanix is explicitly positioning itself to capture with its partner programmes, including the newly launched Powered by Nutanix: Verified Services scheme and the general availability of Service Provider Central, a multitenant control plane aimed at displaced VMware service provider partners.
MCP as a protocol for connecting AI agents to external tools has gained considerable traction in 2025 and 2026, with hyperscalers and specialist vendors alike incorporating support. Nutanix's generally available MCP Gateway is a meaningful production milestone, though enterprise buyers will want to evaluate how it compares with managed agentic orchestration layers offered by AWS, Microsoft Azure and Google Cloud, all of which are building native agent connectivity into their respective platform stacks.
Regulatory read-across
Agentic AI deployments are attracting growing scrutiny under the EU AI Act, particularly where agents interact autonomously with enterprise data and trigger consequential actions. Nutanix's emphasis on fine-grained IAM, least-privilege security, and air-gapped deployment options is consistent with the governance posture the Act encourages, though compliance obligations will ultimately depend on the risk classification of the specific use cases customers deploy on the platform. Enterprises in regulated industries should note that NKP's network-level sandboxing via Nutanix Flow is framed specifically as a mitigation for rogue-agent lateral movement, a threat model that is becoming a formal consideration in AI security frameworks.