80% of technology leaders expect to retain their top people this year. 44% of technology workers say they’re planning to leave.
Both numbers come from the same study, Nash Squared’s 2025 Digital Leadership Report, surveying over 2,000 technology leaders across 62 countries, nearly a quarter of them in the UK. Both can’t be true at once, and having sat in enough retention conversations from the leadership side, I know which one I’d bet on.
Why the gap exists
Confidence about retention tends to be built on relationship, not evidence. Leaders point to how engaged someone seemed in their last one-to-one, how long they’ve been with the organisation, how much they seem to enjoy the team. All of that is real, and none of it is a substitute for actually asking people what they’re weighing up.
Part of what’s driving the gap right now is specific and measurable. 51% of technology leaders report an AI skills shortage this year, almost double the 28% who said the same the year before. That’s one of the sharpest year-on-year jumps in this survey’s history, and it means every organisation with genuine AI capability is suddenly a live option for anyone who has it, whether or not that person was actively job hunting a year ago.
The part that should worry you more than the shortage itself
Here’s what makes the retention risk sharper rather than milder. Nash Squared’s data shows 67% of technology leaders report no measurable return on investment from their AI pilots so far. Only among the very largest technology budgets, above $500 million, does a majority start seeing genuine returns.
At the same time, 52% of companies still aren’t upskilling their existing workforce in generative AI at all. And 65% of digital leaders now say they prioritise AI capability over years of experience when hiring.
Put those three numbers together and a fairly uncomfortable picture emerges. Many organisations are chasing AI capability externally, through hiring, while doing relatively little to build it in the people already on the payroll, even as most of those same AI investments aren’t yet proving their worth. If you’re a capable engineer with five years of tenure and no AI training offered to you this year, watching your employer prioritise a stranger’s AI skills over your accumulated institutional knowledge is not a subtle signal.
What retention risk actually looks like from the inside
I’ve sat across the table from enough departing senior engineers to recognise the pattern by now. The real story usually starts months before any counteroffer negotiation, with a quiet sense that the organisation’s stated priorities and its actual investment don’t match.
Someone hears leadership talk about AI transformation in every town hall, then watches their own request for a training budget get deprioritised for another quarter. They see a job advert for an “AI-native” hire at a higher band than their own. They notice a peer leave for a competitor doing exactly the kind of AI work their own organisation talks about but hasn’t yet funded properly.
None of that shows up in an engagement survey score until it’s already decided. By the time someone hands in their notice, the decision was usually made weeks or months earlier.
What I’d actually ask a board to look at
Retention risk in this environment has less to do with pay, although pay still matters, than with whether your workforce can see a credible path to the AI capability the organisation says it wants, without leaving to get it elsewhere.
I’d start with a simple audit. Of your current technical staff, how many have had any formal AI upskilling in the last twelve months, properly resourced, not a link to a self-serve course nobody has time for. Compare that number honestly with how much you’ve spent trying to hire AI capability externally in the same period. If the external number dwarfs the internal one, you have your answer about where your best people think they stand.
I’d also ask, specifically and by name, whether your three or four most technically capable people currently have a visible route to the kind of work they’d need to stay engaged. Not a vague promise about “exciting projects coming,” a specific answer about what they’re doing differently in six months’ time.
The AI skills shortage is a real, well-documented pressure, and it isn’t going away this year. But the response to it is a choice. Organisations spending on external hiring while their own people watch from the sidelines are, whether they intend it or not, actively training their best staff to start looking elsewhere. The 44% who say they’re planning to leave didn’t arrive at that answer by accident.
Source: Nash Squared / Harvey Nash, “Digital Leadership Report 2025” (26th edition, published 19 May 2025), based on a survey of 2,015 technology leaders across 62 countries, 924 of them in the UK.