U.S. policymakers have issued a 60-day ultimatum demanding urgent action on grid capacity constraints, citing AI data center energy consumption as a primary driver of strain on electricity infrastructure and rising residential costs.
America's power grid just received new marching orders. The Federal Energy Regulatory Commission has directed six regional grid operators to defend or rewrite how they connect data centers, crypto mines, factories, and other massive electricity users to transmission lines—a move designed to accelerate connections without quietly shifting costs to households.
This is no mere Washington paperwork fight. Data centers have become the power-hungry engine driving AI, cloud computing, and modern internet infrastructure. Their explosive growth now collides with electricity bills, summer reliability, and a fundamental question: who pays when the grid must expand?
FERC's orders give regional grid operators and transmission owners 60 days to justify current tariffs or file changes. They must also spend the next 30 days explaining how they will ensure adequate generation capacity for existing and new large loads. The directive covers PJM, MISO, SPP, CAISO, ISO New England, and NYISO—notably excluding Texas's ERCOT, a reminder that America's grid remains a patchwork of regional systems rather than a unified network.
The agency frames this as enabling faster connections for large users while protecting reliability and affordability. In practical terms, it aims to prevent data centers from leaving families and small businesses holding inflated electric bills.
The numbers drive this urgency. A Lawrence Berkeley National Laboratory report, backed by the Department of Energy, found that U.S. data centers consumed roughly 4.4% of national electricity in 2023 and could reach 12% by 2028. This growth concentrates unevenly. In some regions, a single proposed data center campus resembles a small city suddenly demanding to plug in.
FERC defines "large loads" as demand generally exceeding 20 megawatts. Many local disputes involve projects far larger, pushing the issue from utility planning rooms into public meetings, rate cases, and household conversations about monthly bills.
Data centers create grid stress beyond simple consumption. They can destabilize it when many disconnect nearly simultaneously. Last summer in Northern Virginia, Reuters reported that 60 data centers abruptly dropped offline and switched to on-site generators following a voltage fluctuation. The sudden imbalance forced PJM and Dominion Energy to scale back power plant output to protect equipment and prevent cascade failure—a paradox of plenty: the grid had excess power in the wrong place at the wrong time.
Cost-shifting stands as a central concern. If a utility builds new transmission lines or other infrastructure for a proposed data center that never materializes, who absorbs the cost of stranded assets? Consumer advocates warn the answer cannot be households, particularly for natural gas plants and pipelines designed to last decades while data center contracts may shift within years.
FERC is therefore demanding grid operators address cost transparency and rules preventing large customers from pushing infrastructure expenses onto others. Nick Guidi of the Southern Environmental Law Center called the action less ambitious than Energy Secretary Chris Wright requested, though consistent with FERC's traditional approach.
The orders also examine "co-location" and "behind-the-meter" arrangements—where data centers site near power plants or build their own generation while relying on the grid for backup. This sounds straightforward: let data centers provide their own power and leave the grid alone. But the grid is not a spare tire available cost-free when needed. If private generation fails during peak summer demand, the regional system may still intervene, obligating grid operators to determine how much these customers should pay for backup access, flexibility studies, and system stability.
FERC clarified it is not commandeering every facet of electricity business. States retain authority over power plant siting, permitting, and retail electricity rates—critical because households do not pay FERC directly; they pay local utilities shaped by state regulation.
The Sierra Club and consumer advocates have pressed for protections preventing large-load costs from spreading across the broader customer base. Their concern is straightforward: while AI infrastructure offers clear benefits, ordinary customers should not become the default backup for risky forecasts.
The next 60 days will reveal whether regional grid operators produce meaningful reforms or defend the status quo. FERC permits regional solutions, though this flexibility means final rules may differ substantially from California to New England.
This is not fundamentally a question of supporting or opposing data centers. It concerns whether the grid can absorb this new class of massive electricity users without triggering higher bills, excess gas capacity, or reliability crises on power-hungry summer days when air conditioning and internet traffic peak simultaneously. The pressure is mounting faster than politics can move.