AI May Replace White-Collar Jobs, But Who Will Build Its Physical World?

One of the major concerns surrounding artificial intelligence, or AI, has been repeated around the world for years: Will AI take away everyone’s jobs? While discussions continue over the future of software developers, content writers and data analysts, the world’s leading technology and financial giants are quietly facing a completely opposite and unprecedented reality.

Companies such as Meta, Google and BlackRock are laying off software engineers and white-collar employees on one hand, while desperately searching for thousands of electricians, welders, plumbers and construction workers on the other.

At first glance, this may appear contradictory. But when economic data and the global trajectory of the technology industry are examined, the underlying reason becomes clear. The enormous computing power required to operate AI systems cannot be achieved simply by purchasing several hundred thousand processors or GPUs. The real bottleneck in the AI revolution is no longer silicon chips or AI algorithms. It has become the availability of uninterrupted electricity and skilled human hands capable of building the physical infrastructure required to support it.

1. Tech Giants’ Technical Training Initiatives and Hundreds of Millions of Dollars in Funding

To make AI technology functional in the real world, thousands of hyperscale data centres need to be built and connected to regional power grids. However, there are not enough skilled workers in the global market to build this infrastructure. To address the situation, world-renowned Silicon Valley companies are now investing hundreds of millions of dollars directly in technical education and skilled-trade training.

Meta: Technology giant Meta, together with its own and primary construction partners, has established a massive fund of approximately $112 million to $115 million. The money is being used through programmes including its “America’s Workforce Academy” and other partnerships. These initiatives provide completely free training, airfare for transportation, free accommodation and stipends. After an intensive five-week boot camp, participants are given opportunities to work directly at data-centre construction sites. The first cohort is expected to include around 5,000 people, and applicants do not need prior experience in construction or skilled trades to participate.

Google: Technology giant Google has announced a special $50 million grant to accelerate growth in the skilled trades. The project is being operated through the joint network of the International Brotherhood of Electrical Workers (IBEW) and the National Electrical Contractors Association. Google’s primary goal is to increase the number of technical apprentices accepted annually in the United States from 19,500 to 30,000, ensuring an adequate supply of licensed electricians for rapidly expanding data-centre hubs.

BlackRock: The foundation arm of BlackRock, the world’s largest asset-management company, has allocated $100 million over the next five years to develop skilled workers and build a frontline workforce. The fund aims to train approximately 50,000 young people for jobs in advanced manufacturing, substation development and high-voltage grid connections.

2. Why the Acute Shortage? OpenAI and the White House Economic Review

A clear picture of the unprecedented workforce required to build AI infrastructure has emerged from an economic review submitted by OpenAI to US policymakers and the White House.

The OpenAI report clearly states that the company’s proposed mega data-centre projects, along with the new power plants and power-grid connections required to operate them, would require nearly 20 percent of all experienced and skilled electricians and technicians in the United States. In other words, one out of every five experienced electricians in America would need to be engaged solely in building infrastructure planned by one AI company.

There are several deeper reasons behind this acute shortage:

First, the average age of experienced electricians and skilled-trade workers currently employed in the United States is above 45 to 46. Every year, a significant number of experienced tradespeople retire, while trade schools are not receiving enough new students because younger generations are less interested in physical labour.

Second, a modern hyperscale AI data centre can have a power demand ranging from 100 megawatts to 1 gigawatt. One gigawatt of electricity can supply approximately 700,000 to 800,000 ordinary homes. Bringing this enormous amount of electricity to data centres requires massive high-voltage substations, transformers and thousands of miles of heavy cabling. There is an acute shortage of highly qualified, licensed and certified electricians capable of carrying out this specialised work.

3. The Equation of Salaries, Premiums and the Reality on the Ground

Because of this extreme imbalance between demand and supply, the salaries of licensed electricians and technicians have risen to record levels. Popular American television host Mike Rowe visited a high-tech data-centre construction site in Texas and highlighted the nature of the work and the reality of the wages there. He interviewed several young licensed electricians, each under the age of 30 and without the burden of student loans. Yet, because of the urgent deadlines of critical projects, they are earning as much as approximately $260,000 a year through overtime, double-time bonuses and project-completion allowances.

The demand for these skilled tradespeople is so high that, within an 18-month period, individual electricians have reportedly been “hijacked” or recruited away by other companies offering higher salaries as many as three times.

However, several important market and institutional differences need to be understood behind these salary figures:

Average or Median Salary vs Top-End Cases: According to data from the US Bureau of Labor Statistics (BLS), the annual average or median salary of a typical licensed electrician in the United States is around $61,000 to $66,000. The cases involving annual earnings of $260,000 are primarily specialised, project-based positions in data centres and critical infrastructure, where workers may have to perform additional physical labour for as much as 60 to 70 hours a week.

Pay Growth Premium: According to recent indicators from job-market data platform Indeed, salaries for electrical and skilled-trade jobs related to data centres and AI infrastructure are growing approximately 42 percent faster than those for ordinary commercial or residential trade jobs.

Short-Term Boot Camp vs Four-Year Institutional Training: Sean McGarvey, president of North America’s Building Trades Unions (NABTU), has sharply criticised the five-week boot-camp programmes offered by Meta and other technology companies, describing them essentially as a clever PR strategy. He said that five weeks of boot-camp training can teach only basic work such as that of a helper/labourer or cable pulling. However, becoming a qualified “licensed journeyman” or “master electrician” capable of working on dangerous high-voltage projects still requires four years of practical apprenticeship and institutional training hours.

CapEx vs Training Budget: Companies such as Meta and Google are spending more than $335 billion in capital expenditure (CapEx) this year alone on data centres and AI hardware. Compared with that enormous amount, their $265 million or $300 million investment in skilled-trade training is equivalent to only a few hours of their income or expenditure.

It is also important to remember that these job opportunities are intended entirely for the US domestic market. Without an appropriate US work visa and an electrician’s licence issued by the relevant state, there is no opportunity for people from outside the country to directly take up these jobs.

4. Our Social and Cultural Misconceptions and a Major Message for Bangladesh

In South Asian culture, particularly in Bangladesh, a familiar mindset has developed from childhood: “If you do not study hard, you will grow up and have to earn a living as a mechanic or carpenter.” Traditional manual labour and hands-on work have long been viewed as second-tier professions, while graduation and white-collar employment have been given greater priority. But the global economic situation of 2026 has completely overturned this traditional calculation.

Generative AI is rapidly automating coding, text writing, graphic design and conventional office work. However, AI and robotics have still not developed the ability to replace skilled human hands performing tasks in the physical world, such as connecting copper wires, repairing substations or building large-scale infrastructure.

There are three major lessons from this global transformation for Bangladesh’s economy and technical education system:

1. Degree vs Certified Hands-On Skills: The value of internationally recognised, hands-on practical skills is increasing rapidly in the global market compared with conventional general degrees.
2. Upgrading and Certifying Migrant Workers: Every year, hundreds of thousands of electricians, welders, plumbers and construction workers from Bangladesh travel to the Middle East and Western countries. However, because they often lack appropriate international licences or recognised certifications, such as NVQ or ISO-certified training, many are forced to work as unskilled or semi-skilled labourers for the lowest wages. If they can be trained and certified according to international standards before leaving the country, their annual income from the same type of work could potentially increase by up to five times.
3. Changing the Fundamental Perspective: Perhaps we are asking the completely wrong question about AI. The question is not: “Will AI take away our jobs?”

The real and correct question is: “Who will actually build and operate the infrastructure of the vast digital world that AI wants to create?”

In conclusion, no matter how far our digital world moves toward the cloud or the virtual realm, its very existence remains dependent on concrete walls, copper cables, substations and uninterrupted electricity. And when it comes to building that physical infrastructure, the global economy still has no alternative to skilled human hands.

Author: E-commerce analyst and founding member, e-CAB

Note: The entire opinion published in this section represents the personal views of the author. The Digital Bangla Media authorities have no association with these views. As a matter of policy, this article is published without editorial intervention as a reflection of diversity of opinion in the media. If anyone feels offended or provoked by the contents, that is entirely a personal matter.