IBM Sovereign Core aims to put AI control back in the hands of the enterprise

Companies worldwide are demanding more than regulatory compliance from their technology stacks — they want digital sovereignty with verifiable operational control over AI workloads and the data infrastructure underpinning them.

More than 75% of all enterprises outside the U.S. will have a digital sovereignty strategy by 2030 — a signal that the market has moved well past data residency concerns into something far more consequential. As agentic AI accelerates automated decision-making inside enterprises, that governance gap has become a strategic liability organizations can no longer defer. IBM Corp. made a direct move to close it at Think 2026, announcing the general availability of its Sovereign Core platform — a software-only stack designed to give enterprises a fully in-boundary control plane. To understand why that matters, it helps to trace how digital sovereignty itself has evolved, according to Sripriya Srinivasan (pictured), general manager for Core and ALM software products at IBM.

“It started off as data sovereignty … but it very quickly shifted from that to operational sovereignty. Who runs my platform? Where is my control plane? Where are the keys and secrets?” Srinivasan said. “Much more recently in the last couple of years, it’s been all about AI sovereignty. Where do my models run? Is my inferencing governed? Who has access at all times?”

Srinivasan spoke with Dave Vellante at the Think 2026 event, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed the four pillars of digital sovereignty, the architecture of Sovereign Core and IBM’s open-ecosystem go-to-market strategy. (* Disclosure below.)

IBM Sovereign Core reframes digital sovereignty as a runtime requirement

Digital sovereignty means different things to different organizations, but most enterprises have made the same mistake — reducing it to a regulatory compliance exercise. The emergence of AI agents has made that reductionism dangerous, Srinivasan explained. The spread of agents across different departments is now creating internal governance chaos that no policy document alone can solve.

“The proliferation of agents is real in enterprises. Everybody is starting to build agents,” she said. “How do you as a CIO, how do you as a CTO, make sure that there’s a level of standardization, there is a level of consistency, there is a level of governance, there’s a level of orchestration? All of these are internal fears — not necessarily coming just based on regulations.”

To respond to these fears, IBM built Sovereign Core as a software platform that embeds the control plane, secrets management, identity, base infrastructure services and an extensible catalog of IBM, third-party and open-source technologies entirely within the client’s sovereign boundary — a design the company calls Sovereign 2.0. Built on Red Hat OpenShift and Red Hat AI, it works with ecosystem partners including Dell Technologies Inc., Advanced Micro Devices Inc. and Intel Corp., giving customers hardware-agnostic flexibility, according to Srinivasan. IBM’s first design principle was no vendor lock-in — a principle that required the company to resist favoring its own hardware.

“Simply put, it would be an anti-sovereignty pattern if we said, ‘Thou shall only run on our hardware, our software,'” she said. “Our principle was give the control and put the control back in the client’s hands.”

Sovereign Core’s go-to-market reflects the same philosophy, with IBM routing the platform through local managed service providers, ecosystem hardware partners and its existing enterprise customer base rather than pushing a single turnkey stack. The goal is to have customers in AI production in days rather than the 12 to 18 months most enterprise deployments currently take, Srinivasan noted. Ultimately, success will be measured not by what IBM claims but by what customers are able to say for themselves, she added.

“I would actually want my customers to say more than what I say,” she said. “[For them to say,] ‘Hey, we finally have control of our solution. We have a great vendor that we can trust that we partner, but at the end of the day we have the control, we have the independence, and guess what? I can be in production [with] AI in matter of days. It’s no longer a 12 to 18 month project.'”

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

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

Photo: SiliconANGLE