Issue 012: The Talent Pipeline That Doesn't Exist Yet

The US manufacturing skills gap could leave 2.1 million jobs unfilled by 2030, according to Deloitte and the Manufacturing Institute. That's not a story about too few workers. It's a story about roles being invented faster than any education or credentialing system can keep up with, in semiconductors, nuclear, offshore wind, batteries, and now the AI data centres pulling electricians away from every other trade at once. Microsoft's own president has called the electrician shortage the single biggest obstacle to America's AI build-out, and the standard electrical apprenticeship still takes four to five years to produce one fully qualified person, whatever anyone does about it this year.

This week is about five companies who stopped waiting for a labour market to show up and built the pipeline themselves: one that turned a four-year credential into five weeks with a guaranteed interview at the end, one running its apprenticeship as the actual delivery mechanism for a national infrastructure project, one that's run the same programme for nine years and still calls it its biggest intake yet, one that made the training pipeline a condition of the investment before the company would even commit to the site, and one that funded the whole industry's pipeline instead of building a proprietary one it couldn't staff alone. Three Micro-Playbooks on telling a credential problem from a licensure problem, the wage spike that hits every niche talent pool the moment demand outruns supply, and the strategic choice most companies never actually make on purpose: whether you're training your own workforce or growing the whole market's.

Helen Dighton and Rob Baker are living a smaller version of the same problem right now. They're HR Director and Talent Development Lead at Paragraf, a Cambridgeshire and San Diego based company that makes graphene electronic devices using standard semiconductor manufacturing processes, a category that barely had a commercial supply chain a decade ago. In August 2025 Paragraf closed a $55 million Series C, led by Mubadala, to fund the jump from lab-scale production into mass manufacturing. Helen and Rob are the two people building the people function for that jump, in an industry with almost no existing talent market to hire from. They join The Work Life Reporter Live this Tuesday, the first time we've ever run two guests in one seat.

Welcome to the twelfth issue of The Work Life Reporter. This week you get:

Let’s get into it.

TSMC Arizona needed to fill more than 100 equipment technician roles by the end of 2026, at a pace no existing degree pipeline could match. In May 2026, TSMC and Arizona State University launched the Foundations for Equipment Technician Program, offered in three formats: a five-week Monday to Friday accelerator, a sixteen-week evening intensive, and an eighteen-week Saturday-only option. Training happens in labs built to mirror TSMC's actual fabrication floor, the course is free, and admission runs on aptitude and motivation, not a related degree or prior credential.

The insight isn't the speed for its own sake. It's that TSMC didn't wait for a two-year associate's degree pipeline to mature and graduate its first cohort, it compressed the credential itself down to match its actual hiring timeline, and removed the resume gap that normally keeps career changers out of a fab. Graduates get a guaranteed TSMC interview plus an industry-recognised credential that's portable across the wider semiconductor sector, so it holds its value even for the people TSMC itself doesn't end up hiring. Rose Castanares at TSMC Arizona put the stakes plainly: "Technicians play a vital role in fab operations and are essential to the precision and reliability required." Adam Eklund at ASU described what made the partnership different from an ordinary degree pathway: "This one is unique in that it does both", meaning it delivers the credential and the direct route into a job, not one or the other.

Rolls-Royce SMR was named preferred bidder to build Britain's first three small modular reactors, each designed to power roughly a million homes for 60 years or more. The UK hasn't built a new domestic reactor design in decades, so there is no existing pool of engineers anywhere who've done this specific job before, at any employer. In 2026 the company opened recruitment for its second apprentice cohort across seven departments, including manufacturing, supply chain and project management, based in Derby, Warrington and Manchester.

The insight is what the apprenticeship actually is inside this business. It isn't a graduate scheme running alongside the real workforce, it's the primary route into a workforce that doesn't exist anywhere else to hire from, run in serial cohorts tied directly to the build's construction phases. Clare Sandman, Rolls-Royce SMR's People Director, calls it "a pathway into an exciting, dynamic organisation at an incredibly important time for our business and for the wider nuclear sector." That's a fairly literal description of the bet underneath it: a piece of national infrastructure resting on people who are still partway through learning the job, because nobody with ten years' experience building this exact reactor design exists yet, anywhere in the world.

The UK government wants 43 to 50 gigawatts of offshore wind capacity by 2030, which means roughly tripling the sector's workforce to more than 100,000 skilled technicians. Ørsted's answer has been a four-year Wind Turbine Technician Apprenticeship, run every year for nine years running with North Lindsey College: one year of classroom training toward MOET and BTEC Level 3 qualifications, then three years of practical experience across its wind farms, including sites at Barrow-in-Furness, Birkenhead and Grimsby.

What's notable isn't the format, it's the posture. Most companies treat an apprenticeship intake as an annual announcement. Ørsted's 2025 round was described as its "biggest intake" yet, in year nine of the same programme, which tells you it was never a side initiative getting scaled up for good publicity, it's the company's actual primary hiring channel for this role, and it's been treated that way since well before the current hiring crunch existed. More than 95% of apprentices convert into full-time roles at the end, a rate that only makes sense if the final year is designed as a genuine extended interview and onboarding, not a training exercise followed by a separate hiring decision. Coleen Forde, Ørsted's Head of Competency Management, put the aim plainly: "This isn't just about learning a trade, it's about building a meaningful career in a fast-growing industry."

When Kansas was competing to land Panasonic Energy's new lithium-ion battery plant, in what the state calls its largest economic development project in history, its pitch to Panasonic included Johnson County Community College as the lead workforce training provider before the company had committed to build there at all. The 300-acre De Soto site now runs on a training pipeline adapted from JCCC's existing Automation Engineer Technology programme: eight weeks of full-time classroom instruction followed by on-site factory training, paid throughout, with more than 200 people trained through it every year, plus a two-year associate degree route for anyone the plant doesn't end up hiring directly.

The sequencing is the whole story. Most companies build the factory, hire, discover the local talent pool doesn't match the job, and bolt on a training partnership afterward as damage control. Kansas inverted that order: the training capacity was part of the investment case itself, a precondition of Panasonic choosing the site at all, not a fix applied once the plant was already running short-staffed. It also means the pipeline survives independently of any single hiring decision Panasonic makes. JCCC's associate degree keeps producing qualified people for the wider EV battery and advanced manufacturing sector even in years Panasonic itself isn't hiring at pace.

Journeyman electricians building AI data centres in Northern Virginia now earn upward of $120,000 a year, closer to $200,000 with overtime, and it still isn't enough people. The US needs more than 300,000 additional electricians this decade just to keep pace, against roughly 20,000 retiring every year, and a standard electrical apprenticeship still takes four to five years to produce one fully qualified person, according to reporting compiled by Fortune in March 2026. Microsoft's own president has called the shortage the single biggest obstacle to the country's data centre build-out. Google's response, a $50 million commitment aiming to train 300,000 workers across more than 20 states through its Skilled Trades and Readiness programme, didn't go toward a Google-branded academy. It went to existing union and trade infrastructure: TradesFutures, the Electrical Training Alliance, the plumbers' and pipefitters' International Training Fund, and the sheet metal workers' training institute.

The obvious competitive move, when your whole industry is short on a skill everyone needs, is to build a proprietary pipeline that locks graduates into working for you. Google did the opposite, funding shared infrastructure that also produces electricians for every other company bidding for the same tiny labour pool, competitors included. Maggie Johnson, Google.org's Global Head, framed the logic as a scale problem rather than a company one: "Building and maintaining America's future infrastructure requires a large, coordinated workforce of skilled tradespeople." That's a bet that growing the whole market's supply is worth more to Google than owning a smaller, exclusive slice of it.

Not every talent shortage responds to the same fix. TSMC closed its technician gap in five weeks because the bottleneck was curriculum, a course that could be redesigned and compressed once someone decided to do it. The US electrician shortage doesn't work the same way. Even if every company started training tomorrow, the standard apprenticeship still runs four to five years, meaning today's new starters won't reach journeyman level until 2030 or 2031, whatever urgency gets applied in the meantime. Apprenticeship applications are already up more than 70% nationally since 2022, and it barely moves the timeline, because the constraint isn't interest, it's a fixed number of years the law and the trade require someone to spend learning it.

Before you promise your board a fix by a certain date, work out which problem you actually have. A curriculum problem is genuinely fixable in months, the way TSMC and ASU proved. A licensure or years-in-trade problem has a floor nobody can shortcut, no matter how much budget gets thrown at it. Deloitte and the Manufacturing Institute put the scale of this at 2.1 million unfilled US manufacturing jobs by 2030 if nothing changes, and a good chunk of that gap is the second kind of problem wearing the first kind's urgency.

Conflating the two gets you a plan that promises a 2027 fix for what's actually a 2031 problem, and a board that stops trusting your timelines the next time you bring one.

When demand outruns a labour pipeline with a fixed, non-negotiable length, price becomes the only lever still moving. That's what happened to electricians once the AI data centre boom hit: journeyman pay in Northern Virginia is now running $120,000 to $200,000 with overtime, roles that paid meaningfully less only a few years ago, and Microsoft's own president has called the resulting shortage the single biggest obstacle to America's data centre build-out. None of that was in anyone's workforce plan three years ago, because nobody was modelling wage volatility into a role that used to be considered stable and unglamorous.

The lesson isn't really about electricians. It's about any role your business depends on that doesn't have an existing, mature labour market yet, a graphene device technician, a quantum systems engineer, a small modular reactor specialist. Right now those roles are cheap precisely because almost nobody else is bidding for them. That changes the moment the industry around you matures and demand catches up before supply does, exactly what happened to electricians the moment every hyperscaler needed thousands of them at once.

Build wage volatility into your workforce plan for any role with no mature market yet, as a real planning variable, not a surprise you discover during next year's budget round. The companies caught flat-footed by the electrician spike didn't lack the money. They simply never asked the question early enough to plan for the answer.

Look across this issue's five Plays and there are really only three underlying strategies, not five. TSMC and ASU built a credential that's portable across the whole semiconductor industry, useful even to people TSMC itself doesn't hire, which grows the market but creates no lock-in at all. Rolls-Royce SMR and Ørsted train almost entirely for roles inside their own build, with conversion rates north of 90%, a bet that pays off precisely because there's nowhere else in the industry yet for a graduate to go. Google funded shared union infrastructure with zero obligation on any trainee to ever work for Google, the least self-interested option on paper, and arguably the most honest about the size of the problem it's actually trying to solve.

None of these is the correct answer in the abstract. They're three different bets about whether you'd rather own a smaller, guaranteed slice of a talent pipeline or help grow a bigger one you don't fully control. A proprietary pipeline gives you retention at the cost of the whole industry's goodwill and a permanent obligation to keep it running alone. A shared one gives you access to a bigger pool at the cost of watching some of the people you helped train walk straight to a competitor.

Most companies never actually choose between these on purpose. They default into whichever one their first training initiative happened to look like, usually the proprietary kind, because it feels like it protects the investment. It's worth deciding this one deliberately, in the open, before you build anything, because retrofitting a shared model onto a proprietary one later is a much harder conversation than having it now.

Helen Dighton is HR Director at Paragraf, having spent eight years as Head of Sales and Marketing at Encocam before moving into HR. Rob Baker is Paragraf's Talent Development Lead, having spent most of his career recruiting from the outside before shifting to building talent from within as Paragraf scales.

Since Paragraf closed its $55 million Series C in August 2025, led by Mubadala, to fund the jump from lab-scale graphene electronics into mass production, Helen and Rob have taken on building a people function that has to scale at the same pace as a graphene chip factory moving from lab-grade output to commercial manufacturing. Helen's brought a genuinely commercial read into an HR seat, her sense of what a scale-up like this actually needs comes from years running sales and marketing, not a traditional HR career. Rob's shifted the talent function's whole orientation, from finding people outside the company to growing the people already inside it, because the industry has almost no external pipeline to hire from yet. Together they run one people function across two continents, sites in Cambridgeshire, UK, and San Diego, US.

In this week's Leader Spotlight, we look ahead to sitting down with both of them to unpack what changes in a people function when a deep tech company jumps from lab to mass production, where you actually find people for roles that barely existed five years ago, and what it takes to run one people strategy across two continents. A note on what follows: Tuesday's conversation hasn't happened yet as this issue goes out, and this is the first time we've run two guests in one Spotlight, so instead of five direct quotes, here are five threads from their own guest brief worth bookmarking ahead of it.

Paragraf's $55 million raise is funding exactly that jump, and the people strategy has to scale at the same speed as the manufacturing line. Worth watching for what Helen and Rob say actually breaks first when a scale-up moves this fast.

Graphene device manufacturing is still a young enough field that there's no obvious place to go looking for people who've done this exact job before. Worth watching for where they've actually found people instead.

This is Rob's whole remit right now, growing the people already inside Paragraf rather than hiring them in from an industry that doesn't yet have a mature labour market. Worth watching for what that actually looks like week to week.

Helen's read on this comes from eight years in sales and marketing, not a traditional HR path. Worth watching for how that shows up in the calls she makes that a more conventional HR background might not.

Cambridgeshire and San Diego are very different sites with very different local labour markets. Worth watching for how much of the people strategy stays consistent across both versus flexes locally.

Thank you in advance, Helen and Rob, see you Tuesday!

And that's a wrap for Issue 012. Something land? Hit reply and tell us which section you're taking back to your team.

Follow The Work Life Reporter on LinkedIn for what we share between issues, and join us live on Tuesday for our first double-guest episode. Visit theworklifereporter.com for more.

PS. Know someone who leads people and would find this useful? Forward it on. Ten seconds to share, a lot longer to write.

PPS. Got a story, a leader, or a topic you'd like us to dig into? Reply with your ideas.