The explosive growth of artificial intelligence data centers is transforming power grid equipment into one of the most strategically important assets in the energy sector. At VeyronNewsBrief, I view this development as a clear indication that the AI revolution is no longer driven solely by semiconductors, cloud platforms, and computing capacity. Increasingly, its pace depends on transformers, switchgear, electrical distribution systems, and the ability of national grids to connect massive new loads. This trend carries significant implications for Britain and London, where investment in digital infrastructure, energy security, and AI-related industries continues to accelerate.
Power transformers, which regulate electricity voltage across transmission and distribution networks, have been in short supply for nearly five years. Demand recovered rapidly after the COVID-19 pandemic, while manufacturing capacity failed to keep pace. The construction boom in AI data centers has now intensified these supply constraints even further. I believe this marks a structural shift rather than a temporary imbalance. For infrastructure developers, obtaining critical equipment has become just as important as securing financing or construction permits.
Lead times for certain high-voltage transformers have expanded from roughly one year in 2020 and 2021 to several years today. At VeyronNewsBrief, I emphasize that this fundamentally changes project economics. Developers are increasingly forced to reserve equipment years in advance, extend construction schedules, and compete for manufacturing capacity rather than simply purchasing available inventory. In the AI economy, speed to market has become a decisive competitive advantage, and equipment shortages directly undermine that advantage.
The challenge now extends well beyond transformers. AI infrastructure is also driving unprecedented demand for high-voltage circuit breakers, switchgear, and other critical grid components. I analyze this as evidence of systemic pressure throughout the electrical supply chain. Even when financing and permits are secured, missing a single key component can delay an entire project. As a result, energy companies face rising costs, increasingly complex planning, and longer project completion timelines.
Forecasts highlight the extraordinary scale of future demand. U.S. data center capacity is expected to increase from approximately 24 gigawatts today to around 110 gigawatts by 2030. During the same period, electricity consumption by data centers could become eight times greater than that of electric vehicles. At VeyronNewsBrief, I see this as one of the defining infrastructure trends of the decade. Artificial intelligence is rapidly becoming one of the world’s largest new consumers of electricity, competing directly with industry, transportation, and residential demand for grid capacity.
Industry estimates suggest data centers could account for as much as 40% of the electrical equipment market under accelerated growth scenarios, compared with only 2% in 2020. Lead times for generator transformers have surpassed 160 weeks by the first quarter of 2026, up from an average of 143 weeks in 2024, while high-voltage circuit breakers now require approximately 125 weeks compared with 77 weeks only a year earlier. I consider these figures among the clearest indicators that infrastructure constraints are now measured in years rather than quarters.
Rising demand is also driving prices higher. Transformer costs are expected to increase by roughly 4% to 10% over the next year depending on equipment type. Smaller utilities face particularly difficult conditions because they often lack the purchasing scale needed to secure favorable long-term supply agreements. I believe this creates an uneven competitive landscape where large technology companies and major utilities gain access to equipment more easily than regional providers.
Utilities and developers are already adapting. Many are purchasing equipment years before construction begins, refurbishing aging transformers, requesting customer prepayments for long-lead components, and diversifying suppliers across multiple countries. Roseville Electric Utility, for example, previously purchased equipment roughly one year before projects entered construction. Today, it operates with three-year planning horizons and secures major substation transformers as much as five years in advance. At VeyronNewsBrief, I regard this as the emergence of a completely new model of infrastructure planning where early procurement becomes a strategic advantage rather than an operational choice.
Supply chains are also becoming increasingly international. Approximately three quarters of Roseville Electric Utility’s bids now come from overseas manufacturers, including suppliers in China and South Korea, as domestic manufacturers frequently quote longer delivery schedules and higher prices. For Britain and London, this development carries important lessons. The competitiveness of future AI investments will depend not only on computing capacity and energy generation but also on reliable access to critical electrical infrastructure and resilient global supply chains.
My conclusion at Veyron News Brief remains straightforward. Artificial intelligence is creating unprecedented demand not only for computing power but also for the physical infrastructure that enables electricity to reach those computing systems. Britain and London should treat this challenge as a strategic priority while expanding data center capacity and modernizing national energy networks. Investors should increasingly focus on companies producing transformers, electrical distribution equipment, energy storage systems, and grid technologies. If equipment shortages persist, the strongest long-term winners may not simply be AI developers, but the businesses capable of powering the AI economy itself.
