Enterprises embedding AI into core operations face a growing control gap, according to a new IBM study of 1,000 senior executives released June 17, 2026. The research finds that 71% of respondents say switching their primary AI vendor or model would be difficult, while 91% admit they do not fully understand their AI dependencies across vendors, models, and infrastructure. Organizations with the most advanced AI control capabilities protect 55% more operating profit from AI-driven disruptions, yet only 7% of those surveyed operate at that level.
The study, conducted by the IBM Institute for Business Value with Oxford Economics, surveyed executives across 16 countries and 17 industries. It reveals that 68% of respondents find meeting data residency and sovereignty requirements across geographies challenging, complicating efforts to move AI systems or data between environments. Over the past two years, surveyed leaders reported an average of six AI-related disruptions, largely from vendor services. An 81% majority said a seven-day vendor outage would still cause severe or critical disruption, effectively halting operations.
"AI has introduced new forms of dependency that evolve faster than traditional governance, procurement, or technology cycles were designed to handle," said Ana Paula Assis, IBM Senior Vice President and Chair, EMEA and APAC, in the study foreword. "That is why AI sovereignty has become one of the most defining leadership issues of this moment. The stakes are no longer technical; they are economic. Any loss of control can translate directly into margin pressure, compliance exposure, or outright business disruption."
The high cost of vendor lock-in
Unexpected changes across the AI ecosystem-price increases, usage restrictions, model deprecations, and performance degradation-compound the problem. Despite these risks, 72% of surveyed executives said they would accept a 20% cost increase to maintain AI vendors if it improved strategic flexibility. The findings suggest that many organizations are willing to pay a premium for control they currently lack.
Multi-vendor environments driven by necessity, not strategy
Most respondents (73%) described their AI environments as intentionally multi-vendor. However, the data points to less deliberate drivers. Independent business unit decisions (69%) and geographic necessity (69%) were the leading factors, while legacy complexity from mergers, acquisitions, and historical choices was cited by 57%. This pattern indicates that vendor diversity often arises from operational realities rather than a coherent architecture strategy.
For AI for IT & Development teams, these findings highlight the operational risk of unmanaged dependencies across the AI stack. Without clear visibility, infrastructure decisions made in one part of the business can create constraints that ripple across the organization.
Control profiles separate the resilient from the exposed
The study segmented organizations based on how they structure control across data, models, infrastructure, and applications. Only a small fraction-7%-qualified as having the most advanced control capabilities. These organizations experienced less AI downtime and shielded a larger share of operating profit from disruptions. The gap between these leaders and the rest signals a widening divide in AI resilience.
For AI for Executives & Strategy, the message is clear: building adaptable AI systems that can switch components as conditions change is no longer a technical concern but a boardroom priority. The study offers a roadmap for how to design such systems, emphasizing sovereignty as a foundation for business continuity.
Why this matters for general, IT and development, and management roles
For executives, the study quantifies the financial risk of AI dependency: a loss of control translates into margin pressure and operational paralysis. For IT and development professionals, the data underscores the need to map dependencies across every layer-data, models, infrastructure, and applications-to prevent vendor outages from becoming business outages. The full study is available at ibm.biz/ai-sovereignty.
Your membership also unlocks:
