Businesses are racing to adopt artificial intelligence, but a new study from IBM suggests many companies may not be as in control as they think.
The report, based on a survey of 1,000 senior executives worldwide, paints a picture of organizations that have become heavily dependent on AI vendors while lacking a full understanding of the risks that come with that dependence. As AI becomes more deeply woven into daily operations, that could become a costly problem.
SEE ALSO: Sovereign AI sounds independent until you notice everything still runs on NVIDIA
According to IBM, 71 percent of surveyed executives say it would be difficult to switch away from their primary AI vendor or model. That’s a troubling number when you consider how quickly the AI landscape changes. Models get retired, pricing shifts, usage limits appear, and companies alter their strategies all the time.
Perhaps even more concerning is that 91 percent of respondents admit they do not fully understand their organization’s dependencies across AI vendors, models, and infrastructure. In other words, many businesses may not know exactly what would break if a key AI provider suddenly disappeared tomorrow.
The study also highlights ongoing concerns around data sovereignty. Sixty-eight percent of surveyed executives say meeting data residency and sovereignty requirements across multiple regions is challenging. As governments continue introducing new regulations around AI and data handling, that challenge is unlikely to get easier.
The risks aren’t merely theoretical. Survey participants reported experiencing an average of six AI-related disruptions over the past two years. Most of those disruptions were linked to vendor services, yet 81 percent of respondents said a seven-day outage involving a key vendor would cause severe or critical disruption to their business.
IBM argues that organizations need more flexibility and control over their AI environments. The company found that businesses with the most advanced AI control capabilities experienced less downtime and were better able to protect operating profits when disruptions occurred. However, only 7 percent of surveyed organizations currently operate at that level.
What’s particularly interesting is that cost does not appear to be the biggest concern. Nearly three-quarters of executives surveyed said they would be willing to pay 20 percent more to keep their AI options open if it meant gaining greater strategic flexibility.
The report also suggests that many multi-vendor AI environments are not the result of careful planning. While 73 percent of organizations describe themselves as intentionally multi-vendor, respondents pointed to independent business unit decisions, geographic requirements, and legacy technology choices as the primary reasons multiple AI providers ended up in the mix.
The AI industry loves to talk about new models, bigger context windows, and increasingly capable agents. But IBM’s research focuses on a less glamorous issue: control. If companies don’t fully understand the AI systems they rely on today, they may find themselves scrambling when the next outage, pricing change, or regulatory requirement arrives.
For organizations betting heavily on AI, that may be a bigger threat than the technology itself.
Support independent tech journalism
NERDS.xyz is independently owned and operated. If you enjoy my coverage of Linux, AI, hardware, cybersecurity, and tech culture, consider supporting the site on Ko-fi.
Support NERDS.xyz