He has tracked IT infrastructure shifts for 30 years, seeing new innovations disrupt the enterprise tech stack, layer by layer. But artificial intelligence blew them all up at once.
These days, CEO and principal analyst at HyperFRAME Research Steven Dickens bounces between industry trade shows and media interviews, bearing witness to enterprises scrambling to embrace new AI capabilities. He said the IT industry is moving at a pace like never before.
"It's making a fundamental change all the way up and down the stack, whether that's compute infrastructure, networking, operations, the way applications are deployed and managed, Kubernetes, virtual machines, literally all the way up the stack," Dickens told The Forecast in Chicago during the 2026 .NEXT event.
"The landscape I've known for far too many years is all being thrown up in the air. We're almost relearning the entire stack."
The Enterprise AI Stack Has a Long Way to Go
The scale of that relearning is visible in the data. The HyperFRAME Research Lens: The State of the AI Stack reported on the first half of 2026 survey findings from 544 qualified enterprises. Only 14% classify their core data architecture as fully modernized for AI workloads today, while 23% are still running a legacy on-premises data warehouse.
Meanwhile, 79% of enterprises anticipate deploying multiple foundation models concurrently, signaling multi-model architecture as the emerging enterprise standard.
Dickens emphasizes that speed is what sets this shift apart from every prior one.
How the Model Context Protocol Became the Fastest-Adopted Standard in AI History
Consider the Model Context Protocol, or MCP, a technical standard originally developed by Anthropic that defines how AI agents communicate with external tools, databases and services. It's what lets AI systems go from answering questions to actually doing things. People often describe it as "USB-C for AI" because it standardizes how AI agents connect to tools and services across vendors. In most sectors, a standard like that takes years to coalesce. Just 17 months after its 2024 release, nearly 80% of enterprise AI teams had MCP-backed AI agents in production, according to Digital Applied.
"That standard has gone from being a sparkle in some developers' eyes to being an agentic foundation that the Linux Foundation is collaborating on, and mass adoption," Dickens said.
"Every vendor has got MCP baked into it. Getting a telco standard involved would take 10 years to get approved. We just speedrun it in a six-month period, then moved to adoption."
The protocol has drawn backing from Amazon Web Services (AWS), Block, Bloomberg, Cloudflare, Google, Microsoft and OpenAI.
Why Kubernetes Is the Default Platform for Deploying Enterprise AI Apps
Then there's Kubernetes. The open-source platform for deploying and managing containerized applications has been around since 2014. For most of that time, it was a developer curiosity rather than an enterprise staple. That conversation has changed. According to the 2026 Enterprise Cloud Index, 85% of IT executives believe AI is accelerating their organization's adoption of the containers Kubernetes enables in meaningful ways, including 29% who say it's accelerating adoption to a great extent.
"Platform engineering, for me, is looking at non-functional requirements: security, availability, performance, scalability," Dickens said.
"Things that a tinkerer doesn't think about. We've moved beyond that. Now it's not only 'can we deliver fast?' but 'is it production ready? Can we secure it? Can we manage it? What about backup? What about availability? What happens when something goes 'bang' in the middle of the night?'"
Those are the questions of a mature platform. And the timing of that maturity matters.
Kubernetes spent a decade becoming enterprise-grade just as enterprises needed an enterprise-grade way to deploy AI. As companies build the next wave of AI applications, Dickens estimates that a business with 1,000 apps today will add 300 new ones. All agentic AI applications are built for containers.
"I don't see those going onto bare metal or onto virtual machines," Dickens said. "I see them going onto containers."
One arc enabled the other.
"If AI had come along and Kubernetes was two or three years old, we wouldn't have seen the same explosion," he said. "But we've got a really robust and mature way to deploy all these applications at scale, both cloud and on-prem, so the timing is perfect."
Dickens describes Kubernetes today the way you'd describe someone in their mid-20s: career established, edges rounded off, ready for serious work, but still being refined.
"Do we ever fix it all?" he said. "We're still making development on the mainframe space. We're still making developments on virtual machines." Like painting a bridge, he says. A continual process.
The AI Ecosystem Is Sorting Itself Out
The surge in partnerships visible at .NEXT 2026 reflects the next phase of that process. Nutanix announcing deals with NetApp, Everpure and Lenovo in the storage space would have been unusual for a hyperconverged infrastructure company four years ago. The ecosystem is sorting itself out.
"The ecosystem is having to coalesce and reestablish itself around a new norm, which is this AI factory and how we deploy," Dickens said. "I think we're about a year into that ecosystem, with the strata starting to be established. There's still about another year of that to go. Then it's going to be execution."
As for how far along AI itself actually is, Dickens is cautious about anyone who thinks the hard part is behind them. The cloud took 20 years to mature. The iPhone launched in 2007 and continues to shape the industry. AI is still in its early stages of evolution.
"It's the first innings of a seven-game World Series for me," he said.
Related articles:
How Kubernetes Catalyzes Enterprise IT
From Cloud Native to AI-Native: The Evolution of Enterprise Infrastructure
Dual Native Runs Containers in Virtual Machines or Bare Metal
ECI 2026 Report Shows Strain Between AI Innovation and IT Governance
AI's Defining Decade: Nutanix Leaders on Inference, Agents, and the Road Ahead
David Rand is a business and technology reporter whose work has appeared in major publications around the world. He specializes in spotting and digging into what’s coming next–and helping executives in organizations of all sizes know what to do about it.
Ken Kaplan contributed to this article. He is Editor in Chief for The Forecast by Nutanix. Find him on X @kenekaplan and LinkedIn.
© 2026 Nutanix, Inc. All rights reserved. For additional information and important legal disclaimers, please go here.
Related Articles
