The data-centre build-out has a hiring problem money can’t solve

The figures attached to Canada’s data-centre build-out are the kind that stop a conversation. Alberta alone is carrying more than 12,600 megawatts of data-centre capacity in its planning and development pipeline — roughly the size of the province’s entire electrical grid — and the province has set out to attract a hundred billion dollars of investment over five years.¹ Meta broke ground north of Edmonton in July on a campus it has put at thirteen billion dollars and a gigawatt of capacity.² And Meta is only the most visible of them. Amazon, Microsoft and Google are all building or expanding Canadian capacity, Bell and TELUS are standing up AI campuses of their own, and a wave of proposed, developer-led Alberta mega-projects — Wonder Valley and Beacon among them — would each run to several gigawatts if they are built. The capital is real, it is committed, and it is arriving.

What is far less certain is whether the people exist to build these facilities and to lead the projects that deliver them. On the evidence now in front of us, they don’t — at least not in the numbers the announcements assume. That gap is the story, and it is a hiring story before it is anything else.

The capital is the easy part

Money has been the limiting factor on big infrastructure for most of the last two decades. It isn’t here. The cheques are being written at a speed the sector has rarely seen, and the constraint has moved downstream — to power, to permitting, and, increasingly, to the workforce that turns a signed term sheet into a running facility.

Power is the constraint everyone names first, and for good reason. But power problems get solved with engineering and capital, both of which are in the room. The quieter problem is that the engineering itself has to be done by people, and the people who can do it at this scale are not sitting idle waiting for the work.

The constraint has moved to people

The Information and Communications Technology Council modelled what happens when Alberta’s pipeline meets Alberta’s labour market. If half of the proposed capacity were built over five years, the province could be short more than eight thousand construction and skilled workers at the point of peak demand. If eighty per cent of it moved forward in the same window, the gap could approach seventeen thousand.¹ The Council is careful to call these scenarios rather than forecasts, and that caution is fair. The direction, though, is not in dispute: the work being announced outruns the workforce available to do it.

A single facility shows why. Industry figures cited in the same report put a large data-centre build at well over a thousand electricians at peak, alongside hundreds of mechanical trades.¹ Multiply that across overlapping projects competing for the same regional pool, on compressed schedules, and the arithmetic stops working long before the money runs out. South of the border the same squeeze is further along: the U.S. data-centre construction sector has been described as short somewhere between 349,000 and 499,000 workers against current demand.³

The work being announced outruns the workforce available to do it. That is a recruitment problem wearing a construction-budget disguise.

These are not entry-level seats

It would be convenient if this were only a volume problem — more apprentices, more intake, a few years of training. Part of it is. But the roles that decide whether a mission-critical project lands on time are senior ones, and they can’t be trained up inside a build cycle.

The Council’s analysis of more than 4,800 Canadian data-centre job postings found that among those that specified experience, 81 per cent asked for at least two years and 39 per cent asked for six or more.¹ The acute shortages it names are not in the entry tiers at all — they are in experienced facility operators, specialist project engineers, service technicians, and commissioning specialists, the people who actually hold a mission-critical build together.¹ These are people with a track record, which means they are employed, busy, and already known to everyone else hiring in the sector.

In other words, the person you most need to lead one of these builds is almost never on the market. They are mid-project somewhere else, they have been approached before, and a posting will not reach them. This is the same reason senior hiring has always been won by reaching people rather than finding them — a point we’ve made at length elsewhere — except that the build-out has taken a chronic condition and made it acute.

A premium is necessary, not sufficient

The first instinct, reasonably, is to pay more. And the premium is real. One construction-hiring platform put average pay on U.S. data-centre projects at about 32 per cent above comparable non-data-centre builds.³ Demand for the adjacent trades has moved hard in the same direction — one analysis of job postings had roles such as HVAC engineering up well over half, with electrical and welding trades up by double digits, over three years.³ We’ve put our own number on the premium it takes to move a strong engineer, and it is not small.

Here is the part that catches companies out. When the shortage is this broad, everybody pays the premium. It stops being an advantage and becomes the cost of entry — the thing that gets you into the conversation, not the thing that wins it. If you are bidding for the same two or three dozen qualified leaders as Meta, the hyperscalers, and every other Alberta project at once, and all of them are offering strong money, then money is no longer the variable that decides who accepts.

And for the people who lead these projects, it rarely was. A senior construction director or commissioning lead who is already well paid does not move for another increment. They move for the scope of the build, the chance to run something career-defining, the team they’d inherit, and a clear read on whether the project is actually funded and real. Money that only leads the role is necessary. It is not what closes the hire.

What actually moves the person who runs the job

Two things move someone who isn’t looking, and neither can be bought at the last minute.

The first is reach that lands — a warm line to the handful of people who could genuinely run the project, and the standing to get twenty minutes of their time on a call they did not ask for. In a market where the whole sector is chasing the same names, speed decides most of it. And speed is a function of relationships that already exist, not effort applied once the search begins. We get the call answered because we have spent twenty-six years earning the right to make it, and the firm that reaches the right person first, and is trusted enough to be heard, is usually the one that places them.

The second is positioning the opportunity honestly and well. A strong candidate in this market is weighing several approaches at once. What tips the decision is rarely the number on the offer. It is whether the opportunity has been framed around what actually matters to that person — the scale of the mandate, the quality of the people above and below them, the parts of the build that are genuinely hard. That framing is work, and it is most of what we do inside a search. It takes understanding the role well enough to tell a candidate something true that the job description left out, and reading the person well enough to know which true thing will matter to them. Do it well and a cautious “not right now” becomes “tell me more.” Do it from a template and the best people politely decline.

Speed and quality are usually set against each other in recruitment. In this market they are one capability seen from two sides: the reach that gets you to the right person quickly is the same relationship that earns you a conversation worth their time. You do not get one by trading away the other. You get both from a network built over years, or you get neither — and ours is the asset a competitor cannot subscribe to.

What the best candidates are really asking

There is a question the strongest candidates now ask early, and it has sharpened as the build-out has crowded: is the project actually going to happen? Not announced — funded to a final investment decision, with the power secured. They want to know whether the energy supply is contracted or still a hope, whether the grid connection is in hand, whether permitting is advanced or exposed. They have watched megaprojects launch with fanfare and then pause, and joining one that stalls is a risk they carry in their own career, not the company’s.

That instinct is well founded. The same analysis that sized the labour gap also found that Alberta’s grid cannot yet serve all the capacity in its pipeline, and that requests to energize new sites are set to outpace new generation into the early 2030s.¹ Candidates read that too, and they weigh it with every approach that reaches them.

It helps to be clear about who is actually doing the hiring. Behind the hyperscalers sits a layer of specialist developers and infrastructure capital — Brookfield, DigitalBridge and the like, and developer-operators such as Vantage, STACK and Beacon — who develop, and often own and run, the facilities the hyperscalers lease. The developer is frequently the employer, and its model sets the job: developing to own and operate is a different mandate from building to suit for a single hyperscale tenant, which is different again from developing to recapitalize. Each one changes which leaders a project needs, and how convincingly it can tell a candidate the build is funded, tenanted, powered, and going to finish.

So the projects that land these people are not always the ones paying the most. They are the ones that can show, concretely, that they are equipped to finish — the funding is committed, the power is contracted, the schedule is credible, and delivery will not stall six months in. A company that can demonstrate that holds an edge over a better-funded competitor that cannot, because to the candidate certainty is the whole difference between a career-defining build and a line on a résumé that led nowhere. Part of the work of a search is helping a client make that case to the market — and being straight with them when it is not yet there to make.

The pool is wider than your competitor set

When every developer fishes in the same small pond of people who have already built a data centre, the pond gets bid dry. The way through is to stop defining the candidate by the job title and start defining them by the skill that actually transfers.

The hardest parts of a mission-critical build — electrical and mechanical depth, commissioning rigour, cost and schedule control, managing the interface with a utility — are not unique to data centres. They are found in heavy industrial and power construction, in ICI (institutional, commercial and industrial) work, in large-scale industrial and warehousing builds, in the consulting and engineering firms that design this work, and in other industries that deliver capital projects at scale. Someone who has commissioned a cogeneration plant or delivered a large distribution facility often understands substations, switchgear, commissioning sequencing and utility interfaces better than a candidate whose only credential is a finished data hall.

Opening your brief to those adjacent pools does two useful things. It widens the addressable set well beyond the few dozen names every competitor is chasing, and because almost no one else is looking there, it reaches people who are not already weighing three other offers. The judgment it takes is knowing which experience genuinely carries across and which does not, then making that case to a hiring team anchored on a like-for-like résumé. That is a search decision as much as a sourcing one, and it is often the move that unlocks a mandate which has stood open for months.

What this means if you’re building

If you are standing up a data-centre or mission-critical project, the search method is not administrative overhead sitting beside the real work. In a market this tight, it is delivery risk. A leadership seat left open for an extra quarter doesn’t just cost the salary you’re not paying — it moves the schedule, and on a build measured in hundreds of megawatts the cost of that slippage dwarfs any fee.

The instinct an open seat provokes is to widen the advertising and enlarge the pile. The better moves run the other way. Name the small set of people who could actually run the job, and look for them in the adjacent industries most competitors ignore, not only among those who have built a data centre before. Reach them through someone they will take a call from. Give each a reason to move that is more than the cheque, and show them the project is funded, powered, and not about to stall. That is slower to start and far more likely to finish in a leader who stays through commissioning. It is the work we do in data-centre and mission-critical construction, in power and the grid, and across engineering and technology, every day.

The capital will keep arriving. Whether it turns into running facilities depends on a much smaller number of people than the headlines suggest — and on reaching them before everyone else does.