Artificial intelligence (AI) is not only driving up technology prices, as evidenced by a recent development concerning Maryland’s electricity grid. Residents of this state will now face an “additional charge” on their electricity bills due to the growing popularity of AI data centers. The local government has informed its citizens that their electricity bills will increase. The reason for this is to cover a portion of the cost for the grid expansion necessary to accommodate the growth of data centers, particularly those linked to AI. It’s worth noting that a significant portion of this demand is reportedly generated outside the state itself.
This situation mirrors what is happening in Europe, as discussed this week. The European Union intends to compel all operators to remove infrastructure from Chinese companies Huawei and ZTE. Reports estimate the cost of replacing equipment from Chinese manufacturers to be between 17 and 21.5 billion euros. Similar to Maryland, European users could see price increases on their bills, presented as a contribution to improving this infrastructure for the common good. Consequently, whether driven by AI or political decisions, the end consumer is the one footing the bill.
Maryland Residents Must Contribute $2 Billion for New AI Electricity Grid Infrastructure
Obviously, this doesn’t mean each of Maryland’s 6.2 million residents will pay a $2 billion bill. This figure represents a surcharge on electricity rates that they will bear for decades. According to the Maryland Office of the People’s Counsel (OPC), which reported the issue to the Federal Energy Regulatory Commission, the total cost will be approximately $22 billion. Of this amount, $2 billion will be covered by the surcharge on electricity rates. The OPC estimates that after 10 years of this additional tariff, $1.6 billion would be collected.
The distribution projected by the OPC is quite specific: $823 million will fall on residential customers, averaging about $345 per customer; $146 million on commercial customers, averaging about $673 per customer; and $629 million on industrial customers, with an average impact of $15,074. The OPC contends that these costs are “not reasonably related to the actual growth in electricity demand within Maryland.”
This case directly addresses one of the major challenges of AI expansion: who pays for the new energy infrastructure it demands. The demand from data centers has become a primary source of pressure on PJM Interconnection, the major regional grid operator covering 13 states and Washington D.C. Reuters has already reported that PJM anticipates a demand increase of 32 GW by 2030, with 30 GW originating from data centers. Capacity prices in the PJM market have soared in recent years.
This State is Just the Tip of the Iceberg for Higher Price Hikes in More Severely Affected States
Maryland indicates that this is a minor inconvenience compared to other markets. Virginia, Ohio, Pennsylvania, and Illinois were cited as states with much greater data center growth. Virginia is highlighted as a critical point due to the concentration of facilities in the northern part of the state. The complaint notes that data centers in Virginia represented 3.6 GW of demand in December 2024, and their MWh consumption had increased by 660% since 2013. The math is clear: if a less significant state like Maryland must pay around $2 billion through its bills, imagine states where the infrastructure needs to scale more broadly and rapidly.
On March 4, 2026, the White House pledged that major AI and data center operators would not shift their electricity grid costs to households. This highlights the political contradiction that emerges alongside the complaint. If companies like Google, Microsoft, Meta, Oracle, xAI, OpenAI, and Amazon committed to covering the energy and infrastructure costs of their data centers, Maryland finds it illogical that a portion of these costs ends up on ordinary consumers’ bills. These companies are even building their own power plants to reduce costs.
The OPC’s proposed solution is for FERC to compel PJM to change its rules so that transmission costs incurred by data centers are allocated to the zones where those centers are located, or directly to the large data center customers. It also warns of an additional risk: if infrastructure is built based on demand forecasts that do not materialize, utilities can still recover these investments, leaving the risk with current consumers.
