The power industry finds itself in an uncomfortable bind. Demand for electricity is surging, driven by data center buildouts, broad electrification, and the retirement of aging coal fleets, but the equipment needed to meet that demand is stuck in a supply chain that is having trouble keeping up with the pace. Large power transformers, high-voltage switchgear, and transmission cables now carry multi-year lead times. Gas turbine rotor forgings and hot-section blades face severe production constraints. And the upstream inputs that feed it all, from copper and electrical steel to nickel superalloys and cobalt, are increasingly contested by competing industries. “The biggest bottlenecks are in capital-intensive, high-reliability equipment where capacity is limited and substitution is difficult,” said Rafed Hossain, senior technical director with Cognite. Those constraints, he noted, are compounded by utility-side process delays that can turn available supply into unusable inventory.

On the gas turbine side, the picture is especially stark. According to John Shingledecker, principal technical executive, and Bobby Noble, senior program manager for gas turbine research and development, both with the Electric Power Research Institute (EPRI), rotor forgings and hot-section blades are the current chokepoints, owing to a limited supplier base and highly technical manufacturing processes. In some cases, large frame turbines have been shipped without rotors or blades, with installation occurring later onsite just to keep construction schedules from slipping further. Materials are adding another layer of cost pressure. While raw material availability has not yet become a hard constraint, competition for nickel superalloys and cobalt has intensified as lithium-ion battery demand reshapes global commodity markets. Aerospace and energy were once the primary consumers of cobalt, but battery manufacturing now exerts significant pricing influence.

What makes this cycle different from the gas turbine boom of the late 2000s is the convergence of demand drivers. That earlier wave was fueled largely by low natural gas prices. Today’s demand reflects near-term data center needs, limited dispatchable generation options amid growing renewable penetration, and long-term load growth from electrification. New coal capacity is off the table for most utilities, and natural gas supplies, while more abundant than during the first boom, still must support a broader set of end uses. For utilities that need capacity by 2028 but face equipment backlogs stretching to 2030, the options are pragmatic: extending the life of existing thermal assets, repowering and uprating turbines through modifications such as inlet fogging, and pairing renewables with battery storage. EPRI, in collaboration with the Department of Energy’s National Energy Technology Laboratory, is exploring uprate scenarios that could yield double-digit percentage increases in output from some existing units. Data centers are also reshaping competitive dynamics within the turbine market. While they are not typically competing for large-frame machines, they are competing directly with utilities for small and mid-sized turbines in the 30 MW to 100 MW range—prime peaking units. This has opened the door for new entrants and pushed orders for sub-20-MW turbines to record levels in 2025, according to Shingledecker and Noble. The two experts said reciprocating internal combustion engines are gaining traction as well. Meanwhile, aeroderivative units offer rapid deployment potential, but they face their own backlog pressures due to competition with the airline industry.

Amid these hardware constraints, a growing chorus of supply chain experts argues that the path forward lies not just in building more capacity, but in fundamentally rethinking how utilities plan, procure, and execute. Hossain advocated for a shift from one-off project procurement to multi-year programmatic planning with reserved capacity and engineered recovery plans. “Treating supplier management as a core reliability discipline rather than a transactional function” is essential, he said. Better data management is central to that shift. Platforms that connect procurement status to specific assets, outage plans, and reliability risks can transform a late component from an ambiguous delay into a quantified threat to a corridor, unit, or customer obligation. Artificial intelligence (AI) adds another dimension, improving forecasts for critical spares, optimizing tradeoffs across service levels and capital, and surfacing early warning signals from unstructured supplier and field data. Experts at SSA & Co., a global management consultancy, see similar themes playing out across the manufacturing landscape. Jeff Krajacic, a managing director at the firm, described a shift from retrospective reporting toward AI-enabled flow control, where process intelligence and real-time monitoring surface emerging constraints before they impact output. But he cautioned that technology alone is not enough: “Technology surfaces the signal; the operating model determines whether action follows.” Matt Derganc, a senior director at SSA & Co., emphasized that working capital performance increasingly hinges on daily execution discipline rather than financial policy. Many organizations carry excess inventory not because their strategy is unclear, but because execution varies across plants, planners, and suppliers, he said. Meanwhile, SSA & Co. Principal Matt Wilson argued that the planning advantage in 2026 belongs to organizations that embed AI and analytics directly into decision workflows, translating signals into coordinated action across operations, procurement, and logistics. The broader geopolitical backdrop adds further urgency. A 2026 predictions report from supply chain risk intelligence firm Interos identified converging shocks—tariff policies, geopolitical tensions, AI infrastructure bottlenecks, and rare earth constraints—as forces reshaping the global supply chain risk landscape. Companies that handle disruption well, Hossain observed, treat the supply chain like an engineering system, with standardized parts, clear substitution rules, and feedback loops that turn each delay into quantified operational risk. Those that do not, remain perpetually reactive, and in a market where lead times are measured in years, reactive is a position the power industry can no longer afford. —Aaron Larson is POWER’s executive editor.