At the World Economic Forum in Davos, Nvidia chief executive Jensen Huang sought to reframe one of the most heated debates around artificial intelligence: the risk that machines will sap human jobs. Huang argued that rather than producing long‑term mass unemployment, the current phase of AI development is spawning a vast infrastructure boom — data centres, cabling, cooling and on‑site maintenance — that will sharply increase demand for skilled trades and raise wages for electricians, plumbers and construction crews to “six‑figure” levels.
Huang described the build‑out supporting large AI models as “one of the largest infrastructure projects in human history,” saying new investment could reach into the trillions. He told the audience that pay in the sector has “almost doubled” and stressed that high incomes in the AI supply chain do not require a PhD in computer science; vocationally trained workers can benefit directly from the boom.
Other industry figures echoed the point. Palantir chief executive Alex Karp praised vocational training and suggested AI will create more local, on‑the‑ground employment, reducing reliance on migrant labour. Michael Intrator, CEO of data‑centre operator CoreWeave, noted the industry’s “physical” side: the rising need for plumbers, electricians and carpenters to build and maintain the facilities that power modern AI.
Nvidia stands to be a major beneficiary of this wave. As the primary supplier of the specialised chips that train and run large language models and other advanced AI systems, it is at the centre of data‑centre demand. Analysts cited in Davos expect Nvidia’s data‑centre chip sales to approach $200 billion by 2025, while large cloud customers such as Microsoft, Meta, Amazon and Alphabet remain the company’s biggest buyers. The article also notes that major tech firms and others have already committed roughly $500 billion to data‑centre leasing over the coming years.
Yet the optimism about blue‑collar gains sits uneasily beside warnings from other AI leaders about white‑collar disruption. Anthropic CEO Dario Amodei cautioned that automation could eliminate a large share of entry‑level white‑collar roles — he suggested up to 50% of such positions might disappear — and that AI systems already have the capacity to perform many tasks traditionally assigned to junior software engineers. Amodei described a transition in which AI handles a growing portion of routine engineering work, raising the prospect of significant dislocation for many workers.
The juxtaposition of booming demand for skilled trades and the hollowing out of certain office roles highlights a likely bifurcation of labour markets. If Huang’s forecast materialises, regions that can supply trained electricians, plumbers and construction workers will capture outsized gains from data‑centre projects. But the gains will not automatically offset losses among junior professionals and clerical staff whose tasks are more easily automated.
Policymakers and firms face a narrow path. Meeting demand for physical infrastructure requires rapid expansion of vocational training, streamlined permitting and a workforce able to be mobilised where data centres are built. At the same time governments must confront the social and economic consequences of displaced white‑collar workers: retraining budgets, income support, and adjustments to immigration and labour policies will influence whether the AI transition broadens opportunity or deepens inequality.
For investors and corporate strategists, the message is twofold. Hardware and facilities players — chipmakers, cloud providers, construction firms and specialised contractors — are likely to see strong near‑term growth as AI scales. But businesses that rely on easily automated routine labour need to rethink staffing models and human capital investments to avoid painful churn. The coming years will test whether society can translate an AI‑driven capital boom into broadly shared prosperity, or whether the gains concentrate in a narrower slice of the economy.