Artificial intelligence has commanded global attention in recent years, with trillions flowing into a new, turbocharged digital era. Tech giants sell a vision of limitless computation, yet that dream rests on a gritty industrial reality where the AI boom is tethered to electrical transformers from China.
Software does not power server halls. Racks of graphics processors require cooling pumps and backup batteries, and each component demands a chain of heavy electrical equipment. High voltage transformers must step down grid power, switchgear must direct the current, and generators must take over during outages. Without these tools, the most advanced AI cluster remains a silent stack of silicon.
The West faces a chronic shortage of these grid gatekeepers. Delivery times have stretched from eight months to four years, and many data centre projects slated for 2026 face delays or cancellation. The bottleneck is no longer microchips but heavy electrical transformers and switchgear.
Geopolitics now dictates industrial capacity. China has cornered nearly 60 per cent of global transformer production, and its supply chain operates at unmatched scale and cost efficiency. Rivals such as Hitachi Energy, Siemens, and General Electric still craft specialised high‑performance units, but they cannot match Chinese manufacturing speed. The United States sources roughly a third of its large power transformers from China while domestic factories can only satisfy a fifth of national demand.
The recent Trump-Xi summit in Beijing centred on semiconductors, fentanyl, rare earths, and tariff frameworks. Conspicuously absent from the agenda was the electrical transformer. Despite its critical role in powering data centres and stabilising grids, this unglamorous component never entered high-level trade discussions. That omission leaves infrastructure planners in the United States exposed to supply shocks that diplomacy has yet to address.
This is a crucial and myopic omission as American power networks are already buckling under unprecedented stress. Major outages across the United States rose 29 per cent between 2018 and 2024, while average disruption times for major events climbed to nearly nine hours. Extreme weather drives most of these failures, yet ageing infrastructure and delayed equipment deliveries leave utilities increasingly vulnerable. The rapid expansion of data centres will only intensify this strain, demanding millions of additional megawatts from grids that cannot currently replace their oldest components. Diplomacy that ignores transformer supply chains effectively gambles with national resilience during a period of compounding physical and digital demand.
The situation is hardly better in the United Kingdom. National Grid operators have repeatedly warned that transformer shortages are now the primary constraint on new grid connections. Many substations across England and Scotland rely on equipment installed in the 1970s and 1980s, with replacement lead times stretching beyond three years. Recent years have seen a sharp rise in substation failures and unplanned outages, driven by ageing insulation and extreme weather stress. The surge in data centre construction around London and the South East has overwhelmed local distribution networks, forcing developers into multi‑year queues for power connections. Whitehall has acknowledged the vulnerability, yet domestic manufacturing capacity remains negligible and import reliance mirrors the American dilemma. British grids are ageing at the exact moment digital demand accelerates, leaving the nation equally exposed to global supply chain bottlenecks.
This institutional inertia is not unique to energy infrastructure. Cost overruns remain an implacable feature of the Western industrial landscape, particularly the United States. Consider the California High‑Speed Rail project. Under its 2026 Draft Business Plan, the cost of Phase 1 from San Francisco to Los Angeles and Anaheim has risen to roughly $126.2 billion, a figure some analysts believe may balloon to $231.3 billion. Voters approved the project in 2008, construction began nearly a decade ago, yet service is not expected until 2040.
Compare this to the Beijing-Shanghai high-speed railway, which covers a comparable distance of 1,318 kilometres. Construction began in April 2008 and the line entered commercial service in June 2011, with total investment standing at approximately $34.7 billion in 2011 prices. Chinese high-speed rail infrastructure typically costs between $17 million and $21 million per kilometre, while European equivalents often run between $25 million and $39 million.
The disparity reflects more than geography or labour costs. It reveals divergent approaches to procurement, supply chain integration, and state capacity. When the West debates reshoring critical industries, it must confront not just the price of steel or silicon but the institutional ability to deliver complex infrastructure at scale and speed.
This structural inability to scale production leaves Western economies dangerously exposed to shifts in global trade policy. Any escalation would strip away immediate alternatives. Beijing could raise tariffs or restrict exports without warning. Rebuilding domestic capacity would demand billions in new factory investment and several years of construction. Crucially, manufacturers would also need reliable supplies of specialist electrical steel. China dominates that market as well, leaving no easy shortcut.
This procurement paralysis is not confined to the United States. Australia illustrates the transformer dilemma perfectly. Much of the national grid relies on equipment installed decades ago. Coal plants are retiring, renewables are expanding, and data centres are adding massive new loads to an ageing network that cannot cope. In October 2025, the $1 billion Waratah Super Battery north of Sydney suffered a failure of one of its three giant transformers just hours before final commissioning. The transformer was declared beyond repair, and a second unit was taken offline for safety inspection.
Despite the blowback, Canberra maintains anti-dumping duties to shield domestic manufacturers. Yet these firms only produce small distribution transformers, and lack the capacity to build large power units at scale. Developers therefore wait years for imported equipment while absorbing extra tariffs and regulatory delays. Here is where protectionist policy collides with engineering reality. Australia urgently needs foreign transformers but clings to an increasingly impractical model of industrial self-sufficiency.
The bottleneck in the developed world extends far beyond transformers. Facilities need reliable gas turbines for backup power, yet Western factories are fully booked until the end of the decade. Diesel generators face global equipment shortages, while industrial battery supplies rely heavily on Chinese refining of graphite and lithium. In a severe trade conflict, Beijing could restrict multiple components at once. The electrical ecosystem powering artificial intelligence would face simultaneous disruption, grinding Western ambitions for limitless computing to a costly and protracted halt.
Transformers rarely feature in glossy technology campaigns. Yet they govern every grid connection and every megawatt of computational power. Beijing holds the master key, forcing Western leaders to choose between steep industrial investment and prolonged reliance on Chinese exports. Rebuilding domestic capacity requires years of coordinated funding and strategic planning. The alternative is accepting hard limits on digital growth. Current lead times already outpace political ambition, and no algorithm can fix a missing transformer.
Dr Mathew Maavak is a retired transdisciplinary consultant who writes on technology, power, global risks and future systems. He is also the author of the dystopian techno-thriller ‘The Electric Reckoning’.
