Nutanix presents full-stack solution for enterprise AI

Nutanix, a specialist in hybrid multicloud computing, has introduced the Nutanix Agentic AI solution. The full software stack has been specially developed to help customers accelerate the introduction of Agentic AI and use it to transform their business.

With the development towards agentic AI, the spread of AI in companies has reached a turning point. The hurdles to success are no longer the model or the development of individual agents. Rather, the obstacle is the complexity of managing the infrastructure required to securely operate and scale thousands of AI agents.

New requirements for AI infrastructures

Against this backdrop, infrastructure and platform teams need better solutions to build and operate AI factories to enable universal access to their resources, maximize performance and security, and meet digital sovereignty requirements. In addition, data scientists and agentic AI developers expect easy access to tools and services to run and fine-tune models, develop AI agents and connect them to enterprise data.

«AI infrastructures for training models are optimized to handle a single large task. Production AI agent infrastructures, on the other hand, must cope with the scale and high rate of change of thousands of AI services, agents, concurrent users and developers,» said Thomas Cornely, Executive Vice President of Product Management at Nutanix. «Nutanix Agentic AI extends the AHV hypervisor, Flow Virtual Networking, Nutanix Kubernetes Platform and Nutanix Enterprise AI. The result is a cloud operating model for enterprise AI factories. It makes it easier for infrastructure and platform teams to build, operate and govern AI factories. At the same time, Agentic AI developers benefit from the level of performance, rich set of models and AI platform services they need.»

Integration with NVIDIA AI Enterprise

The solution is integrated with NVIDIA AI Enterprise on the Agent Builder layer and orchestrates the NVIDIA-certified AI factory ecosystem for supported configurations. Customers benefit from dynamic multi-user AI environments and can build, run and protect Agentic AI applications. This is supported by a comprehensive software suite for infrastructure orchestration and security, combined with AI platform services and a Model-as-a-Service offering for data scientists and Agentic AI developers.

In addition, Nutanix and NVIDIA are working together on integration with the NVIDIA Toolkit, including the open-source NVIDIA OpenShell runtime environment, to create the basis for autonomous AI agents in enterprise use.

Optimization of performance, safety and costs

Nutanix Agentic AI reduces complexity and offers optimized performance and security. The design serves to reduce token costs and make them predictable. The latest version 2.6 of Nutanix Enterprise AI includes a new AI gateway service for unified rule control across private LLMs and those in the cloud. Also new is support for Model Context Protocol Server and Fine Tuning. It extends the existing robust Model-as-a-Service capabilities to enable AI agents to securely connect to enterprise tools and data sources.

In addition, Nutanix Enterprise AI now supports the NVIDIA Nemotron family of open source AI models, datasets and training tools. With their help, developers can create Agentic AI systems that can draw conclusions on their own, access tools securely and perform complex, multi-part tasks.

Open Kubernetes platform and infrastructure optimization

Nutanix simplifies the path to Agentic AI by extending the CNCF-compliant Nutanix Kubernetes Platform with an extensive catalog of pre-built, open source AI developer tools including notebooks, Vector databases, MLOps workflow engines, and AI agent frameworks. Full integration with NVIDIA AI enterprise software enables developers to deploy NVIDIA NIM microservices, including Nemotron, at the push of a button, accelerating the development of high-performance AI applications for production use.

In the early access version, which recognizes NVIDIA topologies, the Nutanix AHV hypervisor is able to optimize the allocation of physical resources for virtual machines on servers with high GPU density and help maximize performance. With the advanced Nutanix Flow Virtual Networking solution, the network data plane can be offloaded to NVIDIA BlueField. This means high-performance networking and reduced utilization of CPU and memory resources.

Basic services for AI data

Agentic AI applications require basic services for AI data. The Nutanix Unified Storage solution follows the NVIDIA AI Data Platform reference design and delivers linearly scalable read-write performance for thousands of GPU clients. Nutanix provides a high-capacity tier for KV cache offloading and S3-over-RDMA and NFS-over-RDMA support. This creates a scalable, low-latency data fabric with the highest possible GPU efficiency across enterprise AI workloads.

Nutanix Agentic AI includes products that are generally available or in early access and are expected to be available soon. Nutanix is trusted by more than 30,000 customers worldwide to digitally transform and operate their hybrid multicloud environments consistently, easily and cost-effectively.

Further information on AI and Nutanix can be found on the Nutanix website available.