Jensen Huang, chief executive of NVIDIA, forecast that the rush to build AI infrastructure will push demand for skilled trades—electricians, plumbers and similar workers—so high that they could command six‑figure annual salaries. The remark captures a routinely overlooked dimension of the current AI expansion: it is not only software engineers and data‑scientists who benefit, but also the workforce that physically constructs and maintains the data centres, power systems and cooling networks that underpin large models.
NVIDIA sits at the centre of that physical ecosystem. The company’s AI accelerators have become a prerequisite for the largest generative‑AI deployments, driving orders for racks, power distribution, specialised cooling and on‑site engineering work. As hyperscalers, cloud providers and sovereign projects rush to increase compute capacity, the result is a simultaneous spike in demand for on‑the‑ground installers, electricians who handle high‑voltage distribution, and plumbers and HVAC technicians who install and maintain liquid cooling and chilled water systems.
This is a labour story as much as a technology story. Many advanced economies already complain of chronic shortages in skilled trades; the entry of data‑centre construction into that labour market intensifies competition for talent, squeezes supply chains and pushes wages upward. For developing economies, the wave presents an opportunity to convert infrastructural investment into higher‑paying jobs, but only if training and certification systems scale fast enough.
The implications are broad. Higher wages for manual trades could help rebalance the narrative that AI only rewards elite tech workers, while raising broader costs for AI deployment as labour becomes a larger component of capital projects. It could also accelerate credentialing and apprenticeship programmes, change migration patterns for skilled workers, and feed into political debates about industrial policy, immigration and vocational education.
There are also operational and strategic risks. Rapidly rising labour costs and shortages may encourage further automation of installation and maintenance tasks, increased vertical integration by cloud providers, or shifts in data‑centre siting toward regions with available skilled labour or cheaper energy. In parallel, companies such as NVIDIA will need to work with partners on standards and training to ensure the specialist trades meet the technical requirements of high‑density, high‑power AI systems.
For investors and policymakers the takeaway is clear: AI’s value chain extends far beyond silicon and models. The physical realities of power, cooling and cabling are now bottlenecks and competitive advantages. How governments, education systems and companies respond to the ensuing labour crunch will shape the pace, geography and social distribution of the AI boom.