Across the enterprise, AI agents have outpaced the infrastructure meant to support them

As open-source platforms scramble to close the infrastructure gap, enterprises are discovering that autonomous operations expose a deeper problem than performance: control.

The conversation around enterprise resilience has shifted dramatically. A few years ago, resilience meant uptime. Today, it means the ability to absorb shocks from geopolitical, technological and financial angles simultaneously — all while AI agents begin driving autonomous operations across core business workflows. That pressure is landing directly on IT leaders who must build infrastructure that is flexible enough to support an agentic future without surrendering control, according to Abhinav Puri (pictured), general manager of portfolio and community at SUSE S.A.

“When your language models are more embedded in your workflows — when your AI agents are driving your autonomous operations — you’re effectively handing over control of your IP, of your data, of your business processes, all of which combined contribute to your organization’s competitive moat,” Puri told theCUBE. “That is where choice and the ability to have control as an IT leader becomes absolutely critical.”

Puri spoke with Paul Nashawaty at SUSECON 2026, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed how enterprises can bridge the production gap between legacy infrastructure and autonomous operations as well as where SUSE is positioning its platform as the open-infrastructure backbone for agentic AI. (* Disclosure below.)

Closing the autonomous operations production gap

The central challenge enterprises face is not building or experimenting with AI — it’s extracting real return on investment from production deployments. Legacy infrastructure, built to host applications rather than to reason and act autonomously, creates a structural mismatch with what agentic workloads demand, Puri noted.

“The infrastructure that most organizations have today, that’s based on a legacy foundation, was designed to host applications and host workloads. That was designed to be managed. It wasn’t designed to think,” he said. “SUSE, for the last 30-plus years, we’ve mastered the art of software-defined. Now, we are elevating the infrastructure for IT leaders from simply a platform or a foundation to a digital coworker.”

In practice, that means infrastructure that actively orchestrates autonomous operations, managing AI agents rather than just the workloads they run on. To help organizations move without getting locked in, SUSE is shipping Model Context Protocol servers — effectively wrapping the intelligence of its products into a plug-compatible layer that works across any existing AI agentic platform. The approach is not unlike a USB adapter: connectivity without forcing a stack overhaul, according to Puri. This at a time when many enterprises are confronting that the cost of lock-in goes far beyond just the invoice.

“As an IT leader, I don’t want to get locked into a proprietary stack,” Puri said. “I’m paying a tax, so to speak, and it’s not a dollars-per-token kind of a tax. The tax that I’m paying as an IT leader is [one] of control. I’m losing control over my proprietary business processes, my data, my intelligence. That is a challenge which organizations strategically need to step up and think about.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of SUSECON 2026:

(* Disclosure: TheCUBE is a paid media partner for SUSECON 2026. Sponsors of theCUBE’s event coverage do not have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE