Maryland Challenges $2 Billion PJM Grid Upgrade Bill Tied to Out‑of‑State AI Data Centers

Announcement

The Maryland Office of People’s Counsel (OPC) lodged a formal complaint with the Federal Energy Regulatory Commission (FERC) on April 30, 2026, demanding a revision of PJM Interconnection’s $22 billion transmission‑upgrade plan. OPC says Maryland should not be on the hook for $2 billion of those costs, which it estimates will translate into $1.6 billion of extra charges for the state’s residential, commercial and industrial customers over the next decade.

“Without FERC action, Maryland customers face paying billions for transmission infrastructure that PJM is advancing to benefit data centers,” said Maryland People’s Counsel David S. Lapp.

---

Technical specs and cost allocation methodology

PJM’s upgrade program

• Total spend: $22 billion (2024‑2034) for new high‑voltage lines, sub‑stations and control systems across 13 states and the District of Columbia.
• Primary driver: Projected load growth from AI‑intensive data centers, especially in Virginia, Ohio, Pennsylvania and Illinois.
• Capacity added: ~12 GW of new transmission capacity, enough to power roughly 1.5 million additional homes.

How PJM allocates costs today

1. Load‑forecast weighting – each state’s share is proportional to the projected increase in megawatt‑hours (MWh) demand over the planning horizon.
2. Geographic proximity factor – projects that cross state borders receive a small “regional benefit” multiplier.
3. Historical usage adjustment – states that have historically drawn more power from PJM’s network receive a modest discount.

Using this formula, Maryland’s forecasted load growth of 3.5 GW (≈ 0.3 % of PJM’s total) still yields a $2 billion allocation because the model spreads a portion of every interstate line cost to all participating states.

OPC’s counter‑analysis

• Growth mismatch: Maryland’s AI data‑center pipeline is projected at 0.8 GW of new load, compared with 4.2 GW in Virginia and 5.1 GW in Illinois. The OPC argues that the model inflates Maryland’s share by a factor of 2‑3×.
• Benefit mis‑alignment: Most of the new lines slated for Maryland’s region terminate in Virginia and Pennsylvania, where the bulk of the AI compute clusters will sit. OPC cites PJM’s own internal studies showing < 15 % of the new capacity will be utilized by Maryland customers.
• Ratepayer protection pledge: The OPC references voluntary commitments from major hyperscalers to fund grid upgrades that directly serve their facilities. PJM’s current allocation ignores these pledges, shifting the burden to existing utility customers.

Cost impact breakdown (2026‑2035)

Customer class	Total extra cost	Avg. per‑customer cost
Residential	$823 million	$345 per household
Commercial	$146 million	$673 per business
Industrial	$629 million	$15,074 per plant

These figures assume PJM’s original allocation holds; OPC warns the numbers could rise if additional data‑center projects materialize.

---

Market implications and broader context

Immediate financial pressure on Maryland ratepayers

• The $345 residential surcharge is comparable to a modest monthly utility bill increase, but compounded over ten years it represents a 12 % rise in average household electricity spend.
• Industrial customers, many of whom already operate on thin margins, face a $15 k surcharge that could influence site‑selection decisions for future manufacturing or logistics facilities.

Potential shift in FERC’s cost‑recovery framework

• If FERC sides with OPC, it may require project‑specific cost allocation, where only the states hosting the physical infrastructure bear the expense. This would set a precedent for other RTOs/ISOs facing similar AI‑driven load spikes.
• A ruling could also compel PJM to track “beneficiary” loads more granularly, possibly integrating real‑time metering data to validate who actually uses new transmission assets.

Ripple effects for AI hyperscalers

• Companies such as Microsoft, Google, Amazon have pledged to cover grid upgrades that directly serve their data centers under the “ratepayer protection pledge.” A FERC decision that forces utilities to absorb those costs could pressure hyperscalers to renegotiate contracts or accelerate the build‑out of on‑site renewable generation and energy‑storage to reduce reliance on external transmission.
• Some hyperscalers may explore micro‑grid solutions in states with more favorable cost‑allocation rules, potentially reshaping the geographic distribution of AI compute capacity.

Political and community backlash

• The complaint arrives amid a wave of local moratoriums on data‑center construction; 69 jurisdictions have enacted at least a temporary ban on new high‑power facilities.
• Public sentiment surveys show 48 % of Americans oppose a data center in their neighborhood, citing concerns over noise, visual impact, and electricity demand. The Maryland case underscores how grid‑cost disputes can become a flashpoint for broader community resistance.

---

Outlook

The OPC’s filing forces PJM to defend a cost‑allocation model that was designed for a more predictable, utility‑driven load profile. As AI workloads continue to grow at double‑digit percentages year over year, the tension between interstate transmission economics and local ratepayer protection will intensify.

Stakeholders to watch:

• FERC – expected to issue a preliminary decision by Q4 2026.
• PJM – may propose a revised allocation methodology that incorporates beneficiary‑based tagging.
• Hyperscalers – could accelerate direct‑billing agreements or invest in grid‑adjacent renewable projects to mitigate future disputes.

The resolution will likely set the tone for how the United States funds the massive grid upgrades required to keep AI‑driven data centers online without saddling existing consumers with unforeseen costs.

---

For further reading on PJM’s transmission planning, see the official PJM 2024‑2034 Grid Development Plan.