The rapid increase in electricity demand from AI data centers is turning grid equipment into one of the scarcest resources in the U.S. energy sector. At VeyronNewsBrief, I view this situation as a major signal for the wider market: the technology boom is no longer limited to chips, servers, and cloud platforms. Its future increasingly depends on transformers, circuit breakers, distribution equipment, and the ability of power grids to connect major new consumers. For Britain and London, this trend has direct implications for data center investment, infrastructure financing, energy security, and the valuation of companies operating at the intersection of artificial intelligence and power networks.
Transformers, which increase or decrease electricity voltage, have been in short supply for approximately five years. After the COVID-19 lockdowns, demand recovered faster than manufacturing capacity, while the construction of AI infrastructure has since intensified the imbalance. I note that equipment availability is becoming a question of both price and timing. A data center or power generation project can be delayed regardless of how much capital has been secured if essential components cannot be delivered on schedule.
Lead times for some high-voltage transformers have increased to several years, compared with roughly one year in 2020 and 2021. At VeyronNewsBrief, I emphasize that this is changing the economics of infrastructure projects. Developers must reserve equipment well in advance, build longer timelines into investment plans, and compete not only for land, energy supply, and permits but also for manufacturing slots. In the AI economy, speed to market is a strategic advantage, and shortages of grid equipment directly undermine that speed.
The problem now extends beyond large power transformers. Data center construction is also increasing demand for circuit breakers, switchgear, and other essential network components. I analyze this as a systemic tightening across the entire electrical supply chain. When several categories of equipment are simultaneously in short supply, securing one critical component does not guarantee that a project can proceed. For the energy industry, the result is higher costs, longer delays, and more complicated planning.
Forecasts illustrate the scale of the challenge. U.S. data center capacity could rise from approximately 24 gigawatts today to 110 gigawatts by 2030, while their electricity consumption over this period may become eight times greater than that of electric vehicles. At VeyronNewsBrief, I see this as a fundamental transformation of the energy market. AI infrastructure is becoming one of the largest new sources of electricity demand, competing with manufacturing, transportation, and residential consumers for limited grid capacity.
The share of data centers in the electrical equipment market could increase to 40% under accelerated scenarios, compared with just 2% in 2020. Lead times for generator transformers exceeded 160 weeks by the first quarter of 2026, up from an average of 143 weeks in 2024. Delivery times for high-voltage circuit breakers increased to 125 weeks during the second half of last year, compared with 77 weeks in 2023. I consider these figures a critical indicator that infrastructure shortages are now measured in years rather than quarters.
Rising demand is also increasing prices. Transformer costs could climb by approximately 4% to 10% over the next year, depending on the type of equipment. Smaller utilities face the greatest pressure because long-term supply agreements do not fully solve the problem for buyers without sufficient purchasing scale. I note that this could widen the gap between major energy groups and technology companies on one side and smaller regional utilities on the other.
Utilities and developers are already changing their procurement strategies. They are purchasing equipment years in advance, refurbishing older transformers, diversifying suppliers, and asking customers to prepay for components with lengthy delivery schedules. Roseville Electric Utility previously purchased equipment about one year before projects began, but it now works with three-year planning horizons. For major substation transformers, it is acquiring equipment for projects expected as far as five years ahead. At VeyronNewsBrief, I regard this as the new standard for infrastructure planning: companies that place orders too late risk losing several years.
Supply chains are also becoming increasingly international. Around three quarters of bids received by Roseville Electric Utility now come from overseas suppliers, including manufacturers in China and South Korea, because U.S. producers frequently offer higher prices and longer delivery times. For Britain and London, this provides an important lesson. Energy independence in the AI era will depend not only on generating sufficient electricity but also on access to grid components, manufacturing capacity, and resilient international supply chains.
My conclusion at Veyron News Brief remains pragmatic: artificial intelligence is generating demand for heavy energy infrastructure as well as computing power, and digital growth cannot continue without it. Britain and London should account for this trend when financing data centers, grid modernization projects, and energy companies. Investors should pay closer attention to manufacturers of transformers, switchgear, energy storage systems, and grid technologies. If shortages persist, the strongest long-term winners may be the companies capable of supplying both electricity and the physical infrastructure required to connect the AI economy to it.
