
IBM has introduced Sovereign Core, an open source–based software stack designed to help enterprises and governments deploy AI workloads with full regulatory control, compliance, and operational autonomy.
IBM has announced IBM Sovereign Core, describing it as the industry’s first AI-ready, sovereign-enabled software built on open-source foundations to help enterprises, governments, and service providers build, deploy, and manage sovereign AI environments.
The launch comes amid growing pressure on organisations to retain control over their technology infrastructure as regulatory requirements tighten and AI workloads intensify sovereignty concerns. Digital sovereignty now extends beyond data residency to include who operates the systems, how data is governed, where workloads run, and under whose jurisdiction AI models operate. Many organisations currently lack a modern platform to re-host applications, modernise workloads, deploy AI, and maintain continuous compliance reporting.
According to Gartner, “more than 75% of all enterprises will have a digital sovereignty strategy by 2030, often sovereign cloud strategies.”
Built on Red Hat’s open-source foundation, IBM Sovereign Core embeds sovereignty directly into the software rather than layering controls afterward. It enables customer-operated control planes, in-boundary identity and encryption key management, continuous compliance evidence generation, and governed AI inference using local GPU clusters without exporting data externally. Deployment can be completed within days, with flexibility across on-premises infrastructure, in-region clouds, and IT service providers.
“As AI adoption accelerates in India, businesses will need to innovate while meeting tightening regulatory requirements,” said Sandip Patel, Managing Director, IBM India & South Asia, adding that Sovereign Core delivers control, compliance, and operational autonomy while avoiding infrastructure lock-in through open standards.
IBM will roll out the software initially with European partners Cegeka and Computacenter. A tech preview is scheduled for February 2026, with general availability planned for mid-2026.












































































