Three separate pieces of workforce research landed in September 2026, and read together they describe an industry that has correctly identified its problem and is spending money almost everywhere except on the fix.
The short version is that the shortage is real, it is getting worse, and the dominant response has been to bid harder for the people who already have the skills rather than to make more people who have them. The newest proposed solution, which is to recruit technicians out of adjacent industries, is a good idea that quietly inherits the exact problem the pay raises were avoiding.
The numbers are not ambiguous
The 2026 AGC and NCCER workforce survey, built on 1,830 responses collected in July and August, puts electricians at the top of the hardest to fill list at 81 percent, with mechanics at 79 percent and HVAC technicians at 77 percent right behind them. 87 percent of firms have open hourly craft positions right now.
The number that matters more than the ranking is this one: 50 percent of firms say the candidates who do apply are not qualified, because they lack the skills, certificates or licenses the work requires. That is not a shortage of bodies. That is a shortage of prepared bodies, and those are different problems with different solutions.
The response has been price, not capacity
Here is where it gets uncomfortable. Against that 50 percent finding, 55 percent of firms raised hourly craft pay over the prior year, while only 35 percent increased spending on training and development and only 25 percent added in person onsite craft training.
Read those three numbers next to each other and you have an industry that diagnosed a preparation problem and responded with a compensation solution. Raising pay is rational for any single firm, because it moves qualified people from a competitor’s payroll onto yours. It does nothing at all to increase the number of qualified people in the region, which means every firm doing it is competing for the same pool at a higher price and calling the result a labor shortage.
This is a mismatch, not a boom
The detail that settles the argument showed up in Construction Dive’s September reporting, and it is easy to skip past.
58 percent of contractors say data center construction has increased competition for skilled workers, and 49 percent report direct wage pressure from those projects. 88 percent of firms with craft openings say filling them is as hard as or harder than it was a year ago. All of that sounds like a demand boom, and it would be, except for one finding, which AGC chief economist Ken Simonson put plainly on a September webinar: this is the first year since the pandemic that more firms reduced headcount than increased it.
Demand is not up. Headcount is down and the openings are still not getting filled. That combination rules out the explanation everybody reaches for first, and it leaves the one nobody wants, which is that the work has changed faster than the workforce has been trained to follow it.
What a multiskilled worker is actually worth
The same reporting carried the most useful ratio I have seen on this question, out of work by the Simplar Foundation with Brian Lines at the University of Kansas. For every 10 multiskilled workers a contractor employs, they would need 14 to 17 single trade workers to cover equivalent work.
Sit with that for a minute, because it prices cross training directly. A worker who can take a job from the conduit through the terminations through the drive parameters and the startup does not just save you a handoff. They save you between 4 and 7 additional people across a crew of 10, and every one of those people would have needed hiring, onboarding, supervision and a truck.
That ratio is also the argument against the way most plants buy training, which is one narrow topic at a time for one narrow group at a time.
The adjacent pool, and the word doing the work
Deloitte and the Manufacturing Institute extended the same picture into manufacturing in a report released in September, projecting roughly 2.3 million openings across manufacturing and adjacent industry technician occupations between 2025 and 2030, with technician employment growing about six times faster than production occupations. Maintenance and repair technicians sit among the three steepest demand curves in the whole projection.
Their genuinely new contribution is an adjacent pool. They identify nearly 2 million technicians already working in construction, retail and wholesale trade, government, professional, scientific and technical services, transportation and warehousing, and information, whose skills they describe as substantially transferable.
That is a real finding and it should change how plants think about recruiting. It should not change how they think about training, and the word transferable is where those two get confused.
What transfers, and what has to be taught
I want to be fair to the adjacent technician here, because the argument is not that they arrive unprepared.
A good HVAC technician or a good automotive technician shows up with things that take years to build and that you cannot interview for. They have meter discipline, meaning they know what their instrument is actually measuring and they do not trust a reading they cannot explain. They have troubleshooting logic, meaning they narrow a field instead of swapping parts. They have lockout habits that are automatic rather than remembered. They will read a print before they touch anything. Every one of those transfers intact, and any plant that can hire that person should hire that person.
What does not transfer is everything specific to the machine they have not worked on. Reflected wave behavior on a long motor lead does not exist in automotive work, so there is nothing there to transfer from. The tradeoff between carrier frequency, audible motor noise and the thermal headroom you have left in a hot cabinet is not a concept an HVAC tech has encountered. Neither is a bearing that fails by electrical discharge rather than by load, which is the failure that gets misdiagnosed as a bad bearing three times before anybody asks why the new ones keep failing the same way.
None of that is a criticism of the adjacent technician. It is a description of what a training plan has to contain if you hire one. Transferable is a recruiting claim. It is not a training claim, and treating it as one is how a plant ends up with a capable person who cannot yet be left alone with the machine.
Where to start, and it is not with drives
This is the part I would argue hardest, and it runs against what most people expect me to say.
Electricians came in at 81 percent, the hardest craft in the survey to fill. Not drives specialists. Electricians. The binding constraint in most plants is not a shortage of people who understand vector control, it is a shortage of people who can safely and competently work the electrical system that a drive is bolted into.
Drives content taught to somebody who is shaky on grounding, on bonding, on how a three phase system actually behaves under fault, or on what a megger reading means, produces a technician who can change parameters and cannot diagnose anything. I have watched that go wrong more than once, and the failure mode is always the same, which is a person who knows the menu and not the machine.
So sequence it. Electrical fundamentals first, then the motor, then the drive, then the application. That order costs more time up front and it is the only order that produces somebody who can be handed a fault code and trusted to find the cause rather than the code’s definition.
Where this fits
If you are building a training plan rather than a job posting, the training programs are built in that sequence, and the assessment page covers how the pre-test and post-test measure what a technician can actually do at each stage. The VFD Installation Guide and the VFD Troubleshooting Guide are the other half, since together they describe the reasoning a trained technician applies when a machine goes down.
Before the First Fault is the field record the whole approach came out of.
Workforce figures on this page come from the 2026 AGC and NCCER Workforce Survey, September 2026 reporting by Construction Dive, and the September 2026 report from Deloitte and the Manufacturing Institute. The multiskilled crew ratio is from Simplar Foundation research presented by Brian Lines of the University of Kansas.
Frequently Asked Questions
Is the technician shortage a hiring problem or a training problem?
The 2026 data points hard at training. Half of firms say applicants lack the required skills, certificates or licenses, and headcount actually fell over the past year, which rules out a simple demand boom and leaves a preparation gap.
Can we hire experienced technicians out of other industries?
Yes, and the research says nearly 2 million of them exist in adjacent fields. Just budget for training when they arrive, because their general technical discipline transfers and the equipment specific knowledge does not.
What skills actually transfer from HVAC or automotive work?
Meter discipline, troubleshooting logic, lockout habits and print literacy all transfer intact and are worth a great deal. Reflected wave behavior, carrier frequency tradeoffs and electrical bearing damage do not, because those conditions do not exist in those trades.
Why has raising pay not solved the shortage?
Because raising pay moves qualified people between employers without creating any new ones. 55 percent of firms raised craft pay while only 35 percent increased training spend, which is a lot of competition for a pool that is not growing.
What is a multiskilled worker worth compared to a single trade worker?
Roughly 1.4 to 1.7 of them. Research reported in September 2026 found that 10 multiskilled workers cover the work that would otherwise require 14 to 17 single trade workers.
Should drives training come before or after electrical fundamentals?
After. Electricians are the hardest craft to fill for a reason, and drives content sitting on a weak electrical foundation produces someone who can navigate a keypad without being able to diagnose the system the drive is connected to.
The annual report: The State of the VFD 2026
Nineteen drive brands assessed, thirteen scored on what they will actually put on the public record. Free, in a Plant Edition for specification work and a Channel Edition for quoting.