Eman Sumair does not share the angst afflicting many of her contemporaries over artificial intelligence’s potential to devalue graduate skills.
The 19-year-old from Slough, a UK town housing Europe’s largest data centre cluster, is in the final part of a three-year apprenticeship at Equinix, which runs hundreds of the facilities worldwide. Her course of on-the-job training, plus one day a week at college, incurs no university debt and will give her a vocational qualification and experience as a critical facilities engineer — a key role in the infrastructure build-out underpinning the AI boom.
“It’s better than university,” she says. “I know I’m in this industry . . . Data centres are going nowhere. They’re your phone, banks, Fortnite.” Their significance, she adds, is under-appreciated, visible to consumers only when a digital service goes down.
Sumair’s career decision could be a smart choice. The global AI boom is generating a voracious appetite for infrastructure and power, and demand for workers trained to keep the vast, humming sheds of racks and cooling systems running. It is also reverberating across the economy. “Construction workers build data centres, the power industry bolts down turbines, telecom firms lay out networking equipment,” says Ajay Rajadhyaksha, global chair of research at Barclays, adding it may be the “biggest industrial investment cycle” in the US since the second world war.
Politicians are keen to present the data centre frenzy as one facet of a broader renaissance in well-paid, skilled work that does not require a university degree. They contextualise it in a broader revival of industrial policies promoting domestic manufacturing, public investments in civil infrastructure and — in Europe — rearmament.
In the run-up to his re-election, Donald Trump promised tax and tariff policies would kickstart a “new American industrialism”; last summer, the White House claimed it was overseeing a historic increase in blue-collar wages.
UK Prime Minister Keir Starmer has scrapped a long-term target for half of young people to go to university, and wants immigration curbs to encourage the training of homegrown apprentices. In Germany, Friedrich Merz is hoping fresh funding for roads, bridges, energy, communications and defence will sustain well-paid technical jobs even as automotive sector employment declines.
So far, however, there is little sign that any promised “blue-collar boom” will be powerful enough to reverse the long-running decline of traditional skilled manual jobs in the overall labour market.
Following a post-Covid rebound, US manufacturing has been steadily shedding jobs since 2023 as automation and, more recently, AI hammer the sector. It has lost 68,000 jobs since Trump’s return to the White House, according to the Bureau of Labor Statistics, continuing a long-running decline as a share of overall employment. Over the same period, US employment in mining and oil and gas declined by 13,000. Construction, even with the boost from the data-centre build-out, has added just 14,000.
In Germany, a recent hiring spree by defence contractors has so far failed to offset a drop of about 500,000 manufacturing jobs since the pandemic, according to government data, although wider fiscal stimulus will feed through this year.
In the UK, there are fewer jobs in manufacturing and construction than in 2019, although lower-paid sectors, such as transport and logistics, have grown. The increase in defence spending — once contracts are confirmed and work under way — could also give some towns with high levels of deprivation a new lease of life.
There are hotspots of stronger demand and skills shortages in sectors with an ageing workforce. Jim Farley, Ford’s chief executive, said in November the company had 5,000 open mechanic positions and bemoaned a lack of trade schools in the US. “We are in trouble in this country,” he said. “It’s a very serious thing.”
In the UK, where developers have lodged hundreds of planning applications for data centres, “we can all see the demand”, says Neil Crudden, managing director of CRG TEC, a critical infrastructure recruiter. His clients are fighting to hire from a small pool of technicians who can run uninterrupted power supplies — increasingly poaching military veterans and hospital engineers with transferable skills, or training their own apprentices.
Still, many economists are sceptical about a potential society-wide revival in the kind of employment politicians and unions hanker after: well-paid, stable factory jobs that anchor a community and allow people to support a family on a steady salary.
The lowest-paid workers in the UK and US saw stronger wage growth in the immediate aftermath of the pandemic than their white-collar peers, due to widespread labour shortages and, in many European countries, increases in minimum wages. But rising living costs have squeezed those on middle incomes, and in the US, college graduates are once again pulling ahead of the less qualified.
Manufacturers now rely on a slimmer, more skilled workforce and this is unlikely to change, says labour market economist Christopher Pissarides, because “you cannot pay blue-collar jobs the same as 10 years ago . . . the [product] would be so expensive there would be no demand for it”.
Non-graduates who would have seen manufacturing as a natural career path in the past are now taking more precarious service-sector work, as couriers, domestic cleaners or dog walkers, Pissarides adds. “That type of job will grow, but not traditional blue collar — these are ‘no-collar’ jobs, where you just wear a T-shirt, and there’s more risk involved.”
The precarious reality of non-graduate work is a counter to the popular view that hands-on skills will prove more resilient to AI than white-collar roles. Geoffrey Hinton, the so-called godfather of AI, recently advised listeners to “train to be a plumber . . . Someone like a legal assistant, a paralegal, they’re not going to be needed for very long.”
But an analysis of jobs that will be critical to the UK’s industrial strategy — conducted by higher education data provider QS, York university and consultancy Public First — suggests graduate skills are unlikely to be immediately superseded. It found 80 per cent of the roles key to expansion of the government’s eight chosen sectors — from advanced manufacturing to clean energy — require a minimum of degree-level skills.
“The financial returns for HE [higher education] are not what they were. But as the workplace becomes more highly skilled, there [are] large parts of the labour market you can’t get into without a degree,” says Charlie Ball, head of labour market intelligence at Jisc, a non-profit data agency focused on higher education.
Even in-demand blue-collar sectors such as construction suffer from unstable employment models and challenging conditions compared with graduate jobs. “They are generally the best paid, but also long hours, physically demanding, and very, very male,” Ball says. Unlike nine-to-five jobs with sick pay, maternity leave and holiday pay, “there’s a lot of self-employment, some are precarious. You are often at the whim of construction budgets.”
Debbie Phillips, campus director of data centre operations at Equinix, says on her company’s sites demand for contractors is indeed “very spiky — you might need 300 people for four months, then those numbers fall away”.
The hordes of telecoms engineers needed to wire data centres during construction are relatively easy to hire in the UK because of the cyclical nature of the industry, Crudden says. They are back on the market after a private equity-fuelled rush to roll out fast broadband fizzled out.
Finding critical infrastructure engineers to keep data centres running is tougher, Crudden says — partly because “babysitting” the equipment is seen as less exciting than problem-solving in the design and build phase. Unions also worry these long-term jobs will be small in number. It takes a small army of workers 12 to 18 months to build a data centre, but just a couple of hundred to run even the biggest.
“When it’s up and running, how many jobs does it sustain?” asks Paul Nowak, general secretary of the UK’s Trades Union Congress.
He believes to create good jobs ministers should prioritise clean energy and traditional industries threatened by US tariffs, rather than courting Big Tech. In his view, an expansion of companies such as TikTok in the UK would “employ a handful of people, pay little tax and add little value”.