Rebecca PY MALAYSIA, 21 JULY 2026 – Nutanix, a leader in hybrid multicloud computing, will showcase its vision for enterprise Agentic AI at AMD Advancing AI 2026, demonstrating how organisations can build scalable AI infrastructure while improving cost control, governance and data sovereignty.
As enterprise AI moves from isolated experimentation to production-scale deployment, autonomous AI agents are increasingly retrieving context, interacting with enterprise tools and continuously executing complex workflows. This shift is driving new challenges around infrastructure scalability, governance and the economics of AI. As agents consume and generate millions of tokens daily to perform routing, validation and execution tasks, the financial reality of rented infrastructure is becoming a significant bottleneck to scale.
Nutanix offers a pragmatic path to a two-tier intelligence strategy, allowing organisations to reserve frontier AI models for complex reasoning while routing high-volume workloads to optimised models running on private infrastructure. Using an agent gateway, organisations can intelligently route requests, enforce policy, manage access and gain greater control over AI costs.
At AMD Advancing AI 2026, Nutanix will demonstrate how to make this two-tier Agentic AI model sustainable. With the Nutanix Cloud Platform and Nutanix Enterprise AI, combined with AMD EPYC processors and AMD Instinct MI355X accelerators, enterprises will be able to gain greater control over their AI workloads.
Key capabilities include running long-lived AI agents while helping manage token costs, supporting data sovereignty by keeping AI workloads and proprietary data within governed hybrid multicloud environments, and maximising infrastructure utilisation through shared inference infrastructure powered by AMD EPYC CPUs and AMD Instinct MI355X GPUs.
Debo Dutta, Chief AI Officer at Nutanix, said owning your intelligence does not mean completely abandoning frontier models, but rather taking control of routing, volume and costs. He explained that this approach allows enterprises to reserve expensive frontier models for a small fraction of tasks requiring complex, edge-case reasoning, while routing the vast majority of high-volume tasks to highly optimised models running securely on private infrastructure.