The manufacturing skills gap, by the numbers.

Every manufacturing leader I talk to is fighting the same fire: they can’t find enough skilled people, and the ones they have are stretched too thin to do the work that actually matters.

The data backs up what we’re all feeling on the floor. U.S. manufacturing needs 3.8 million new employees by 2033, and Deloitte and The Manufacturing Institute project up to 1.9 million of those roles could go unfilled if the skills and applicant gaps aren’t closed. As of December 2025, there were already 433,000 open manufacturing positions in the U.S.

Now zoom into the trades specifically, because that’s where the pain is sharpest for shops like ours.

The American Welding Society projects the industry needs more than 330,000 new welding professionals by 2028 — about 82,500 a year — while the average welder is 55 years old and roughly 30% of that workforce is retirement-eligible. Machining is no better off: CNC machinists average 50 to 55 years old, and for every five master machinists who retire, only one or two apprentices are coming in behind them. That’s not a slow leak. That’s a workforce cliff.

And here’s the part that should keep every ops leader up at night: apprenticeship programs take three to four years to produce a fully qualified machinist, and even an accelerated CNC program requires up to 4,000 hours of training over two years. You cannot hire your way out of a multi-year training pipeline problem in a hot labor market. The math doesn’t work.

This is a Theory of Constraints problem, not a recruiting problem

If you’ve sat through my Lego TOC simulation, you already know where I’m going with this. When you have a constraint in your system, you don’t throw more undifferentiated effort at it and hope. You identify it, you exploit it, and you elevate it — deliberately, in that order.

Right now, skilled trade labor is the constraint for most manufacturers I know. Nearly half of the new roles manufacturing needs by 2033 are at risk of going unfilled if we don’t close the skills gap, and more than half of the advanced manufacturing workforce will need upskilling to meet changing demands by 2030. That’s not a “post more job ads” problem. That’s a systems problem, and it needs a systems answer.

Most companies respond to a labor constraint in one of two broken ways:

1. They chase bodies. Raise wages, widen the net, hope quality holds. It doesn’t scale, and it doesn’t fix the pipeline.
2. They chase automation as headcount reduction. Buy the robot, cut the seats, call it a productivity win. This is where I think most manufacturers get it backwards.

Automation’s real job is to exploit your constraint — not eliminate the people in it

Here’s the reframe I use with every client and every subsidiary leader: automation isn’t there to replace your people. It’s there to free your most valuable people from the work that isn’t actually using their skill.

Every hour a journeyman welder or a master machinist spends on a task a machine could do — material handling, repetitive deburring, basic programming a well-trained operator could run — is an hou

r they’re not spending on the work only they can do: process troubleshooting, tolerance-critical work, training the next generation, solving the problems that actually require 20 years of pattern recognition.

That’s what “highest and best use” means in a TOC framework. You don’t get more throughput out of a constrained resource by working it harder on low-value tasks. You get more throughput by stripping everything off its plate except the work that exploits its actual capability. Automation, done right, is exactly that stripping mechanism.

The research backs this up. The World Economic Forum’s Global Lighthouse Network — tracking Fourth Industrial Revolution frontrunners across 220+ sites and 35 industries since 2018 — has found that wearables supporting training and upskilling improve overall equipment effectiveness, not just capacity. Automation and workforce development aren’t competing investments. They compound each other.

But there’s a trust gap we have to close first. Only 17% of manufacturing job seekers view employers’ use of AI positively, and two-thirds believe AI will eliminate more jobs than it creates. If we roll out automation without an explicit, communicated plan for how it elevates our people rather than replaces them, we’re going to fight our own workforce on the way to solving the workforce problem. That’s on us as leaders to get right — not just in strategy, but in how we talk about it on the floor.

What this looks like in practice — the IDEA approach

This is exactly the framework I use for every operational bottleneck and the skills gap is no different:

• Identify — Map your actual constraint. Is it headcount, or is it that your skilled people are burning hours on work that doesn’t require their skill? Most shops assume it’s the former when it’s really the latter.
• Develop — Build the internal training pipeline you can’t outsource. The data on credentials is telling us where the industry is already voting with its feet — certificate programs, built for foundational and occupational-specific training, have seen real growth because they get people productive faster than a four-year degree track ever will. Don’t wait on the labor market to solve this for you. Build the apprentice-to-journeyman pathway inside your own walls.
• Execute — Deploy automation deliberately at the points in the process where it exploits your constraint — where it takes low-skill-required tasks off your best people’s plates — not just wherever capital happens to be available.
• Assess — Track whether your skilled people’s time is actually shifting toward higher-value work, not just whether machine utilization went up. If your master machinist is still doing debur work eighteen months after the new equipment showed up, you didn’t solve anything. You just spent money.

The bottom line

The skills gap isn’t going to close because we found more welders. With CNC machining alone projected to grow at a 7.2% CAGR on the back of aerospace, medical device, and telecom demand, the gap widens even as the retirement wave accelerates. The companies that come out ahead won’t be the ones that hired the most people. They’ll be the ones that built internal development systems and used automation to put their existing people at their highest and best use.

That’s not a talent strategy. That’s a constraint strategy. And it’s the only one that actually scales.

---