Only 9% of executives say they have an excellent understanding of their AI dependencies as organizations face growing vendor lock-in and operational resilience challenges.
Key Takeaways
- Only 9% of executives say they have an excellent understanding of their AI dependencies.
- 71% say switching their primary AI vendor or model would be difficult today.
- 72% would accept a 20% cost increase to maintain multiple AI vendors for strategic flexibility.
- 81% say a seven-day outage at a primary AI vendor would have a severe or critical impact on operations.
- Organizations with stronger control across data, models, and infrastructure protect 55% more operating profit from AI-driven disruption.
The IBM Institute for Business Value released a new report examining how organizations are managing AI sovereignty as AI becomes increasingly embedded in business operations.
Based on a survey of 1,000 senior executives, the report found that many organizations are expanding AI adoption while struggling to maintain visibility into the vendors, models, infrastructure, and data dependencies that support those systems.
The findings suggest AI sovereignty is emerging as a governance issue as organizations seek greater control over the technologies driving operational decisions.
Organizations lack visibility into AI dependencies
Only 9% of executives said they have an excellent understanding of their dependencies on AI vendors, models, and infrastructure.
At the same time, 71% said switching their primary AI vendor or model would be difficult if required today. IBM said the findings highlight the growing challenge of maintaining flexibility as organizations become increasingly reliant on AI services and providers.
The report defines AI sovereignty as an organization’s ability to move data, swap models, and shift workloads across environments when technical, commercial, or regulatory conditions change.
These dependencies have become more complex over the past two years, with executives reporting unexpected price increases, usage restrictions, service deprecations, changes to data-handling practices, and geographic access limitations.
Vendor concentration creates resilience concerns
The study found that 81% of executives believe a seven-day outage affecting a primary AI vendor would have a severe impact on business operations.
In addition, 75% of executives who switched or attempted to switch AI vendors over the past two years said the process was difficult due to data portability issues, model revalidation requirements, compliance obligations, and being tied to a specific vendor’s technology.
Despite those challenges, many organizations are willing to invest in flexibility. The report found that 72% of executives would pay 20% more to maintain multiple AI vendors if doing so improved strategic freedom and reduced dependence on a single provider.
IBM noted that 28% of organizations already use four or more AI vendors as part of a strategy to maximize flexibility and avoid lock-in.
Governance and control linked to business outcomes
The report found that organizations with the strongest control across data, models, infrastructure, and applications protect 55% more operating profit from AI-driven disruption than organizations with less control.
The report also found that 57% of executives said replacing a core AI model would require significant decoupling or even a complete system rebuild. In comparison, 56% said moving core AI systems and applications to a different vendor would take at least six months.
As AI systems become more deeply embedded in business operations, IBM said organizations are increasingly evaluating how much control they maintain over the data, models, and infrastructure that support those capabilities.
