The artificial intelligence boom has created an unexpected supply chain crisis in an unlikely place: the industrial power equipment market. As compute-intensive AI workloads demand ever-increasing electricity, major data center operators face a critical bottleneck—new gas turbines now require six years or longer to manufacture and deliver. This timeline mismatch has forced companies to explore unconventional solutions, including repurposed jet engines and refurbished turbines originally designed for other applications. What began as a temporary workaround has become a structural feature of how the industry secures its energy infrastructure.
The economics driving this shift reveal deeper tensions in AI's path to mainstream deployment. New turbine procurement from manufacturers like GE and Siemens involves extended lead times, substantial capital expenditures, and regulatory approval processes that can stretch into years. Rather than wait, data center operators have begun acquiring and retrofitting used jet engines—hardware originally built for commercial aviation—to generate the gigawatts their operations require. This creative repurposing reflects both entrepreneurial problem-solving and a quiet acknowledgment that traditional energy infrastructure cannot scale as rapidly as AI demand accelerates. Some operators are also turning to used industrial turbines from decommissioned power plants, essentially creating a secondary market for aging generation equipment.
However, this strategy carries significant complications. Environmental regulators view retrofitted jet engines and aging turbines with skepticism, citing emissions concerns and air quality impacts. Permitting remains arduous even for reconditioned equipment, as municipalities grapple with hosting massive data centers that consume electricity equivalent to small cities. Supply constraints on new turbines simultaneously drive up prices for used alternatives, compressing margins and making power procurement a genuine business risk for AI infrastructure investors. The calculus is brutal: deploy older, dirtier equipment quickly, or wait years for compliant new turbines that may never materialize in sufficient quantities.
This turbine shortage illuminates a fundamental constraint on AI's expansion that venture capital narratives often gloss over—the physical world cannot always adapt as quickly as software demands. Data centers require reliable, abundant baseload power, and that requirement cannot be solved through clever engineering alone. Whether this pressure ultimately accelerates nuclear energy adoption, forces renewable infrastructure investment, or simply perpetuates reliance on improvised fossil fuel solutions remains an open question with profound implications for both AI's sustainability trajectory and energy markets broadly.