Manufacturing's Demand Charge Problem Is Getting Worse, Not Better.

The electricity cost trajectory for manufacturing is worsening. Goldman Sachs Research estimates that AI data center load growth will add 40-50 gigawatts of demand to the U.S. grid by 2030 — a demand surge that is already driving wholesale capacity prices sharply higher in multiple regional transmission organization markets.

Capacity charges are a component of most industrial electricity bills — often embedded as a demand-related rider rather than a clearly labeled line item. As wholesale capacity prices rise, those charges increase automatically, often without any corresponding notice to the commercial customer.

The audit implication is twofold. First, as capacity charges grow, billing errors in those charges become more expensive. An error that inflated capacity charges by 10% when capacity was $5/kW-month costs twice as much when capacity is $10/kW-month. Second, the tariff provisions governing capacity charge calculation are complex and frequently applied incorrectly.

Industrial manufacturers operating in PJM, MISO, and NYISO territories have the highest exposure to capacity charge errors, as these markets have the most complex capacity pricing structures and the most significant recent capacity price increases. The demand curve behind those capacity prices has only gotten steeper. Goldman Sachs Research raised its 2030 forecast again in 2026, now projecting a 220% increase in U.S. data center electricity consumption from 2023 levels, with domestic data center capacity climbing 197% between 2025 and 2030 to roughly 95 gigawatts. Every additional gigawatt of that load flows into the wholesale capacity prices embedded in industrial customers' bills.