Nutanix bets on 'agentic' AI

Nutanix went to investors this week with a pitch that sounds bigger than a new feature: companies will soon need a traffic-control tower for artificial intelligence workers that can make decisions on their own. Nutanix says it wants to be that control layer, deciding where those systems run and what company data they are allowed to touch. (finance.yahoo.com) The timing was not accidental. Nutanix held its investor day on April 7, 2026 alongside its.NEXT conference in Chicago, where chief executive Rajiv Ramaswami argued that “agentic” artificial intelligence will behave like a hybrid enterprise application spread across private data centers, public clouds, and company software systems. (diginomica.com) (stocktitan.net) An artificial intelligence agent is basically software that does more than answer a question once. It can take a goal like “resolve this support ticket” or “prepare this financial report,” then call tools, pull data, and execute several steps with less human hand-holding than a normal chatbot. (siliconangle.com) (diginomica.com) That creates a mess large companies already recognize. If one agent reads payroll records in a private data center, another calls a language model in a public cloud, and a third writes back into a customer database, someone has to enforce the rules in every step. (finance.yahoo.com) (siliconangle.com) Nutanix’s answer is a stack, not a single model. Its April 7 announcement described new Nutanix Agentic AI capabilities, including a multitenant management portal meant to give artificial intelligence engineers and developers self-service access while still keeping usage secure and governed. (nutanix.com) The company also framed cost control as part of the product. Nutanix executive Thomas Cornely said the software is designed to reduce complexity and make token spending more predictable, which is another way of saying companies want a meter on every artificial intelligence action before bills spiral. (computerweekly.com) This is why storage companies and infrastructure vendors keep showing up in the artificial intelligence story. Nutanix used the same event to talk up support for NetApp storage and a Lenovo partnership, because an agent that moves across systems is only useful if the underlying data, compute, and access rules move with it. (finance.yahoo.com 1) (finance.yahoo.com 2) Nutanix is not selling this only to classic corporate information technology teams. Its April 7 release said the next wave of features is also aimed at “neoclouds,” a new crop of artificial intelligence-focused cloud providers that want to offer secure, scalable services to customers building agents. (nutanix.com) The same week, Infosys and Harness announced a partnership built around agentic artificial intelligence for software delivery. Their plan combines Infosys consulting and cloud services with Harness tools to speed releases, improve reliability, and automate more of the work that sits between writing code and keeping systems running. (prnewswire.com) (infosys.com) Put those two announcements together and the market starts to look different from the 2023 artificial intelligence boom. Vendors are shifting from “here is a model” to “here is the operating system for fleets of models, tools, permissions, logs, and costs,” because that is where large companies actually get stuck. (siliconangle.com 1) (siliconangle.com 2) Nutanix still has to prove that customers want it as the referee for all of that. But by using investor day to tie artificial intelligence agents to hybrid cloud, storage, governance, and spending controls, it told Wall Street that the next enterprise artificial intelligence sale may look less like buying a brain and more like wiring a whole building. (finance.yahoo.com) (diginomica.com)