Ask a C-suite executive whether their company controls its data, and the answer is likely to be a confident yes. Ask how quickly they could move their most critical workloads away from their biggest infrastructure provider, and the confidence evaporates.
That disconnect sits at the heart of the Leaseweb Digital Independence Survey 2026, which polled 800 C-suite leaders across the UK, US, Germany and the Netherlands. Some 76% of respondents said they hold complete or high control over their ability to move data to another provider or environment. Yet only 11% said they could migrate business-critical workloads away from their largest cloud or infrastructure provider within three weeks. For the remaining 89%, the exit door is a lot narrower than the floor plan suggests.
More striking still, 55% said such a move would take longer than three months or is not currently possible at all.
Portability on paper, lock-in in practice
Lex Boost, co-CEO at Leaseweb, put the problem plainly. “Organisations may believe they control their data, but once applications have been built around a provider’s platform, moving them is rarely quick or simple,” he said. The distinction matters. Owning your data is not the same as being able to relocate the applications, pipelines and integrations that give that data its value. Every proprietary service adopted and every architecture decision tuned to one platform’s quirks adds friction to a future departure, and the cost of that friction rarely appears on any balance sheet until the day someone tries to leave.
The research shows how unevenly that friction is distributed. The Netherlands recorded the highest share of businesses expecting migration to take more than three months, at 65%, and Dutch respondents gave the longest average estimate of 4.7 months. The UK followed at 59%, then the US at 55%. Germany was the outlier at 40%, though even there, four in ten leaders expect a move to stretch beyond a quarter of a year.
AI is tightening the grip
If the current picture is uncomfortable, the trajectory is worse. More than three-quarters of leaders (79%) expect AI to increase their reliance on a small number of providers over the next three years. The expectation is strongest in the US at 89%, followed by the UK at 83%, Germany at 76% and the Netherlands at 67%.
The logic is not hard to follow. AI workloads demand scarce GPU capacity, specialized tooling and vast volumes of data, and the organisations best positioned to supply all three are a handful of hyperscalers. As companies build models, training pipelines and inference services on top of those platforms, they deepen the same dependencies that already make exits slow and costly. Respondents seem to see it coming. AI-related technology and provider dependence was the most frequently cited barrier to digital independence overall, selected by 28%, although German leaders ranked legacy-system complexity first, at 30%.
Geopolitics is forcing the issue
Dependence is no longer just a technical or commercial concern. Some 78% of respondents said they have already acted to reduce reliance on a particular technology or infrastructure provider in response to geopolitical challenges. That figure reaches 91% in the US and 87% in the UK, and 45% of respondents overall plan to go further. The era of treating infrastructure choice as a purely operational decision appears to be over, replaced by a recognition that where workloads live, and who ultimately controls them, is a matter of strategic risk.
Yet there is an uncomfortable tension between those actions and the stated reality of migration timelines. Leaders are taking steps to diversify while simultaneously acknowledging that most could not execute a meaningful exit in under a quarter. Intent is running well ahead of capability.
Independence has a price, and leaders will pay it
Perhaps the most telling finding concerns money. Almost all respondents said they would pay more to gain greater control, reduce provider dependence or improve their freedom to move. Nearly three-quarters (73%) would pay at least 11% more, and the average premium respondents named was roughly 16%. Willingness was greatest in the Netherlands, where 81% would pay at least 11% more and 26% would pay at least 21% more. Germany followed, with 79% prepared to pay at least 11% more.
For CIOs, that figure reframes a familiar conversation. Optionality is usually discussed as a cost to be minimized, a hedge that eats into efficiency. These results suggest that executives increasingly see it as an asset with a measurable value, one worth a double-digit premium.
A surprising ranking of priorities
The survey also surfaces some uncomfortable data about what leaders treat as non-negotiable. Fewer than three in ten regard regulatory and legal compliance (28%) or security and data protection (28%) as absolute requirements when choosing a cloud provider, and two-thirds (67%) would accept at least a minor trade-off on compliance. Priorities vary by market: availability and uptime lead in the US at 35%, provider reputation and trust in the UK at 34%, and security and data protection in Germany at 30%. In the Netherlands, regulatory compliance and exposure to geopolitical or trade-related risk tied at 27%.
Taken together, the findings suggest that many organisations are making trade-offs on the very dimensions that shape their long-term exposure, and doing so at the moment when AI is set to make those choices harder to reverse.
The message from Boost is one of urgency. Leaders, he argued, need to understand and act on their dependencies now, “not when a pricing change or geopolitical shift forces their hand.” That is sound advice, and it begins with an honest audit. The relevant question is not whether you own your data, but how long it would actually take to run your critical workloads somewhere else, and what would break along the way. Few organisations can answer that with precision today. The 11% who can are the ones with genuine leverage when contracts renew, prices change or the geopolitical weather turns.
In an AI-driven decade, digital independence will not be won by declaration. It will be built, deliberately and early, in the architecture decisions made long before anyone needs to leave