CHICAGO — Nutanix Inc. is positioning itself as the architectural backbone for a rising category of specialized cloud companies as the industry shifts from training massive artificial intelligence (AI) models to the complex task of running them.
On Tuesday, the enterprise cloud computing company announced a major expansion of its Nutanix Agentic AI solution at its .NEXT conference here. Scheduled for release in the second half of 2026, the new capabilities are specifically engineered for neoclouds — a new generation of agile cloud providers offering on-demand GPU access — seeking to move beyond simple infrastructure and into high-value AI services.
Until recently, the AI boom was defined by a gold rush for raw computing power to train Large Language Models (LLMs). However, Nutanix executives say the market is entering a second phase: the era of agentic AI.
This shift requires more than just raw GPUs; it demands security, sovereign data control, and cost-efficient inference, the process of running AI applications in production.
“Demand for sovereign and specialized AI clouds is accelerating as organizations look for ways to access AI while maintaining control over their data,” said Thomas Cornely, executive vice president of product management at Nutanix. He added the new platform is designed to help regional providers offer “powerful AI capabilities” that compete with global hyperscalers while maintaining local data integrity.
The upcoming updates focus on a multitenant, multiservice portal. This allows neocloud providers to offer a diverse menu of as-a-service products, including Kubernetes, virtual machines, vector databases, and models.
The framework introduces enhanced tenant isolation, which lets multiple companies share the same physical GPU hardware without risking data leaks or performance “jitter” from neighboring users. For the neocloud provider, the update to Nutanix Cloud Manager adds usage-based metering, allowing them to bill customers based on precise GPU usage or API calls.
Industry analysts suggest Nutanix’s infrastructure pivot is necessary for the next wave of corporate adoption. Scott Sinclair, practice director at research firm Omdia, noted that legacy infrastructure cannot handle the governance and security risks inherent in autonomous agents.
“Nutanix’s focus on strong governance and tenant isolation provides a welcome option for CIOs seeking an enterprise-grade foundation for their AI agent strategy,” Sinclair said.
Enterprises’ steady adoption of AI agents has grown beyond a drumbeat and into a crescendo, rallying many businesses to use it for coding and multiple workloads, Nutanix Chief AI Officer Debo Dutta said in an interview. “AI agents are here and real, people are modernizing their apps, and organizations are revving up their architectures,” he said.
Indeed, Nutanix added 1,000 customers last quarter and averaging 500 to 1,000 a quarter, CEO Rajiv Ramaswami said in a press conference on Wednesday.
In a secondary announcement, Nutanix unveiled NKP Metal, an extension of its Kubernetes platform that allows the software to run directly on bare-metal hardware. This move is designed to squeeze maximum performance out of AI workloads by removing the virtualization tax, further lowering the cost-per-token for developers.
By providing the full stack for AI delivery, Nutanix aims to turn regional neoclouds into formidable competitors in the global AI race, ensuring that the next generation of digital agents has a secure place to reside.
