
Every new campus needs a leader who has opened one before. For a competitor building nearby, that person often works at the site down the road. I think this is the one operators underrate most of all the retention risks.
Losing a chief engineer can cost a site years of knowledge that never made it into a procedure. No job posting brings that back.
In Uptime Institute's 2026 Global Data Center Survey of more than 800 owners and operators, 53% struggled to find qualified candidates, up from 46% in 2025.
Uptime calls turnover a persistent problem and says staff are often lured away by other data center companies.
The reasons frontline data center staff leave, as well as the fixes that keep them, apply to leaders too. My question: if your site director, chief engineer or VP of Data Center Operations quits tomorrow, who runs the facility?
Why Do Competitors Go After Operations Leaders?
There's more to build than there are people to run it. JLL's midyear 2026 report counts more than 66 GW of data center capacity under construction in North America, with vacancy around 1% for a third straight year.
Every one of those campuses has an opening date and none opens without someone who's commissioned a site before.
Hiring one away from a rival is the quicker route for any company willing to pay.
Iron Mountain, which runs roughly 30 data centers globally according to Fierce Network, knows both sides of this. Mark Kidd, EVP and GM of its data center and asset lifecycle management businesses, told the publication, "We end up hiring from each other and driving up wages."
He also said the customer ends up paying for it. And he's right.
Fierce Network reports that experienced electricians managing multi-site crews in Northern Virginia now command more than $200,000.
Anyone recruiting a site director there is shopping in that same expensive market.
This Year’s Leadership Moves
- May: JLL split its strategy, hiring Brandon Keesee externally as managing director of hyperscale and colocation while promoting Michael Martin from inside the ranks.
- April: Vantage Data Centers hired former Iron Mountain CHRO Emma Jeffries as global chief people officer, putting talent acquisition and leadership development squarely on her plate.
- June: CyrusOne brought in Bob Hennegan, formerly Meta's VP of global data center facility operations and a retired U.S. Navy rear admiral, to run its U.S. operations.
- September: Matt Landek left JLL, where he oversaw 2,500 employees across 340 data centers, to take the president and CEO role at L5 Mission Critical, a new services venture backed by Walbridge. No counteroffer competes with a CEO seat.
What Does a Site Lose When a Chief Engineer Leaves?
Uptime found average modal rack density passed 11 kW for the first time in 2026, and 24% of respondents ran at least some racks at 30 kW or higher, up from 19% in 2025. The job has gotten harder, which is part of why data centers now need AI-native leadership.
Before anyone hires a leader for an AI facility, I'd want to know exactly which power and cooling systems they've run and whether that matches what's going into the building.
In Uptime's skills data, operations management was a hard-to-fill gap for 35% of operators, close behind electrical and junior operations roles at 38% each.
Those are skills gaps across the whole workforce, and Uptime doesn't track turnover for chief engineers or site directors.
My own read goes past what Uptime measured. I see the junior operations shortage as a warning about the leadership pipeline, because today's junior operators are tomorrow's chief engineers.
Why Doesn't Matching a Competitor's Offer Work?
The usual fix is a counteroffer, and I don't think it holds. A competitor building a new campus has room to offer your site director more money and a bigger job. Matching the salary buys time. Then the next offer shows up.
I'd give leaders a next move inside the company instead. A chief engineer might take on commissioning the next facility, and a site director might move up to running several sites.
Where equity is part of pay, vesting tied to a construction cycle gives people a reason to stay until the project's done.
Planning matters for the same reason. In my view, a chief engineer who helped design a new hall's electrical systems will want to see it run, and someone handed a finished building doesn't feel that pull.
How Do Operators Protect Their Leadership Bench?
Digital Realty is growing more of its own talent. Audrey Escalante, its head of global operations learning and development, told Fierce Network that a master electrician or former HVAC technician sometimes enters at a mid-level engineer role after some bridge training.
The company also pairs veterans with newer hires, so knowledge gets passed on before someone leaves. The same approach works for data center leadership, where every chief engineer needs a deputy who knows the site's equipment history and has made real maintenance calls under supervision.
A name on a succession chart won't get a site through an outage. And the knowledge has to be written down while the leader's still there. Equipment that keeps faulting goes in that file, along with the decisions made after earlier incidents.
Vendor escalation contacts belong there as well. It works best when the deputy runs some maintenance reviews with the chief engineer in the room, explaining the reasoning.
Pay should get checked against the local market on a fixed schedule, since waiting for a resignation letter is too late.
I'd also track leader departures separately from technician turnover. If one site keeps losing experienced managers, that points to a career problem a raise is unlikely to fix.
What If the Deputy Gets Poached Too?
An operator that trains a strong number two has created another person a rival wants to hire, and that's a fair objection. I'd accept the risk anyway.
The damage is smaller, for one. When a chief engineer leaves a site with no deputy, most of what they knew goes too, and that's a much bigger hole to fill.
And I'd expect a deputy stuck behind a boss who isn't going anywhere to start looking around. A real path to chief engineer gives them a reason to stay, which is why I see the succession plan and the career path as the same plan.
When to Let a Leader Go
Some leaders will leave no matter what you offer, because the next job is bigger than anything their employer has open.
The goal then is a survivable exit, with a deputy who's ready and a search that starts before the notice period ends.
An internal successor is enough to keep an existing site running. But a new campus is a different problem, especially for a technology company building its first owned data center.
It needs someone who's already run critical facilities and understands the computing infrastructure going in. That person often has to come from outside.
How Should Technology Companies Choose a Data Center Recruiting Firm?
An AI company building its first campus needs a site leader who knows how to run critical facilities and also talk shop with the infrastructure engineers filling the halls. I'd expect that pool to be small. Anthropic has hired a number of former Googlers for its data center team, including Winnie Leung, its head of data center infrastructure.
So what would I ask a recruiting partner? Recent placements come up before anything else, in the exact roles you're filling. Christian & Timbers' guide to choosing data center recruiting for technology firms covers the full process.
A firm that placed a chief engineer or VP of Data Center Operations in the past year knows current pay and who's open to a move. If most of its infrastructure work is IT or cloud roles, it'll be learning on your search.
Off-limits agreements matter as much, because the people you want work at other operators and a firm that recruits for those operators is barred from approaching them.
I'd get that list in writing before signing. It's a boring question, but it decides whether the shortlist is real.
Commissioning and high-density experience need checking too. Someone who's only run low-density enterprise sites might not be ready for an AI hall, however good the résumé looks.
Who Runs the Site the Morning After?
I'd stop treating leadership retention as a pay problem. Pay matters, but the leaders I'd worry about are the ones who see a bigger job somewhere else and no next step at home.
The question I'd want every operator's CEO to answer is who runs each site the morning after its leader resigns. If the honest answer is a name on a chart, the retention plan isn't finished.
Frequently Asked Questions About Data Center Leadership Retention
What Is the Difference Between a Data Center Site Director and a Chief Engineer?
Think of the site director as the person who answers to corporate for the facility’s budget and its customers. Power and cooling, along with the maintenance program that keeps them running, sit with the chief engineer. Some smaller sites hand both jobs to one person. On a large campus they’re two hires, and each one needs a successor.
What Should a Retention Plan for Data Center Leaders Include?
A deputy for each senior leader comes first, ideally someone who already knows the site’s equipment history. Regular pay reviews help, though they won’t keep a leader who sees no next step at home. So the plan should also name an internal next role for every site director and chief engineer, and count their departures apart from technician turnover.
What Experience Should an AI Data Center Leader Have?
Hands-on time with the power and cooling systems going into your building, and at least one commissioning from start to finish. A résumé built on years at low-density enterprise sites can look strong and still miss what an AI hall demands.
Which Executive Search Firms Recruit Data Center Operations Leaders?
Senior operations searches usually go to retained firms that run dedicated data center practices. Christian & Timbers is a full executive search firm with roots in AI and technology search. Its data center work covers VP of Data Center Operations, Head of Data Center Infrastructure, VP of AI Infrastructure, CTO and board-level roles, and its global reach puts clients in front of passive candidates, such as chief engineers already running critical facilities who aren’t looking for a move.

