Dive Brief:
- Enterprises embedding AI into core operations and processes are getting locked into AI systems that can be difficult to change, IBM’s Institute for Business Value revealed in a study based on insights from 1,000 senior executives published earlier this month.
- More than 7 in 10 respondents said it would be challenging to switch from their primary AI provider. Another 68% of respondents said meeting data residency and sovereignty requirements make it difficult to move AI systems and data across business environments.
- As AI adoption grows, the need for organizations to strengthen control and oversight of systems is also rising, said Ana Paula Assis, IBM SVP and chair for IBM Europe, Middle East, Africa and Asia Pacific, in the study’s foreword. “AI has introduced new forms of dependency that evolve faster than traditional governance, procurement or technology cycles were designed to handle. That is why AI sovereignty has become one of the most defining leadership issues of this moment.”
Dive Insight:
IBM's Institute for Business Value study isn’t the only one pointing out enterprise concerns over vendor lock-in as more AI systems come online.
Research firm Forrester warned in a May blog post that changes to SAP’s API policies enacted earlier this month would restrict third-party AI agents, large-scale data extraction and other workarounds beyond SAP-endorsed pathways. Additionally, the ERP giant made its agentic Joule Studio 2.0 free through Dec. 31 as a way to bring on customers, “knowing that the customer will inevitably pass through a metered tollbooth down the road in 2027,” the Forrester blog said.
“This is bigger than an IT integration issue and demands board attention,” Forrester analysts wrote. “A vendor control event with budget and AI strategy implications with this level of strategic partner should be treated as such.”
Lock-in concerns are catching the attention of regulators.
The U.K. Competition and Markets Authority began investigating Microsoft’s business software ecosystem this year as concerns rise about AI adoption entrenching customers with the tech giant. The European Commission also on Wednesday said Microsoft and Amazon should be designated as gatekeepers under the Digital Markets Act for their cloud offerings.
“They both have vast and entrenched user bases and appear to benefit from lock-in effects and high switching costs, in addition to a large ecosystem,” the commission said.
Other vendors are changing their AI pricing strategies as enterprise adoption advances. GitHub began charging for token usage this month, while Workday and Zendesk changed their pricing structures based on AI features. AWS is also raising the price for its EC2 Capacity Blocks for ML, which allows companies to reserve accelerated compute instances for future start dates.
IBM’s Institute for Business Value study positions AI sovereignty as the key to enterprises maintaining control over their AI systems and lessening vendor dependency. Enterprises with advanced controls over their AI systems often see less downtime and “protect 55% more operating profit from AI-driven disruptions,” according to the study.
However, only 7% of organizations reported such levels of control, which IBM said signals a widening gap between companies with flexible AI systems and those “constrained by dependency.” Only 9% of businesses have an excellent understanding of AI vendor dependencies, according to the study.
While 73% of surveyed executives said their business environments intentionally have multiple AI vendors, IBM found that the practice is driven less by strategy and more by operational realities, such as legacy system complexity.
“Multivendor strategies only strengthen AI sovereignty when they are actively orchestrated,” the study said. “That means establishing common standards for data, models, and security; defining where variance is allowed and where consistency is required; and creating transparency across regions and business units. This lets leaders manage their AI vendors as a strategic portfolio, optimizing for adaptability, resilience, and long-term control.”
