Maryland Citizens Slapped with $2 Billion Power Grid Upgrade Bill for Out-of-State AI Data Centers — State Complains to Federal Energy Regulators, Says Additional Cost Breaks ‘Ratepayer Protection Pledge’ Promises

Why It Matters

Shifting the $2 billion burden could protect Maryland households from inflated utility bills and set a precedent for how grid costs are allocated to AI‑driven data centers nationwide.

Key Takeaways

• Maryland consumers face $1.6 B extra grid costs over ten years.
• PJM allocates $2 B to Maryland despite low local AI demand.
• State urges FERC to shift costs to data centers or building states.
• 69 U.S. jurisdictions have moratoriums on new data‑center projects.

Pulse Analysis

The PJM Interconnection, which operates the largest high‑voltage transmission network in the United States, is embarking on a $22 billion modernization program to accommodate the surge in electricity consumption from AI‑driven data centers. While the overall spend is justified by rising demand in several member states, Maryland’s share of the load forecast is modest. Nevertheless, PJM’s current allocation formula assigns the state $2 billion of the cost, translating into an extra $1.6 billion for Maryland customers over ten years. Critics argue the methodology ignores geographic demand differentials and places an undue burden on consumers who do not host the new facilities.

Maryland’s Office of People’s Counsel has taken the dispute to the Federal Energy Regulatory Commission, invoking the ‘ratepayer protection pledge’ that tech firms allegedly made to bill themselves for grid upgrades. If FERC intervenes, the state could force PJM to re‑allocate the $2 billion to the jurisdictions where the data centers are built, potentially shielding residential users from a $345 per‑household surcharge. The outcome will also signal how regulators balance utility‑scale infrastructure financing with emerging AI workloads, a question that could reshape cost‑recovery rules across the PJM footprint.

The controversy arrives amid a growing backlash against AI data centers, with 69 U.S. jurisdictions imposing moratoriums and public surveys showing roughly half of Americans oppose nearby facilities. As hyperscalers chase cheap power for training large language models, they increasingly encounter community resistance and regulatory scrutiny. Maryland’s challenge may encourage other states to demand direct billing of data‑center operators, accelerating a shift toward usage‑based cost recovery. Ultimately, the case could influence both the pace of AI‑related grid investments and the political calculus of future data‑center siting.