Loading...
Loading...
Virginia's electric grid is under more simultaneous pressure than any other state in PJM's footprint, and possibly in the country. Dominion Energy Virginia, the investor-owned utility serving 2.7 million customers across eastern and central Virginia, is managing three structural load changes at once: a data center explosion in Northern Virginia's Data Center Alley (Ashburn, Loudoun County) that has added roughly 1,000-2,000 MW of new load per year since 2022, the integration of the Coastal Virginia Offshore Wind (CVOW) project — the largest offshore wind project in the Americas at 2.6 GW — and the simultaneous operation of two nuclear units at North Anna and two at Surry, providing roughly 3,400 MW of baseload capacity under Nuclear Regulatory Commission oversight. Appalachian Power, the AEP Virginia subsidiary, serves western Virginia from the Roanoke Valley to the coalfield counties, a geographically and economically distinct service territory that faces different load characteristics and different generation mix. The State Corporation Commission (SCC) regulates Virginia's investor-owned utilities under a framework that has been substantially modified by the Virginia Clean Economy Act (VCEA) of 2020, which mandates 100% carbon-free electricity for Dominion by 2045. Virginia is in PJM, the largest wholesale electricity market in the world, and PJM capacity market rules heavily influence investment economics for Virginia generation assets. LocalAISource connects Virginia utility operators, data center developers, offshore wind contractors, and SCC stakeholders with AI specialists who understand PJM market dynamics, VCEA compliance requirements, and the specific grid physics of integrating 2.6 GW of offshore wind with a data center load that can surge by hundreds of megawatts within months of a new hyperscaler lease signing.
Loudoun County, Virginia hosts more data center square footage than any jurisdiction in the world — an estimated 35% of global internet traffic routes through facilities in Ashburn and the surrounding Data Center Alley corridor. Amazon Web Services, Microsoft Azure, Google Cloud, and Meta all operate major campuses here, with dozens of smaller operators filling in between. The load growth rate from 2022 to 2025 has been without historical precedent in Dominion's planning models: new data center connections adding 1,000+ MW annually to a grid that historically grew at 1-2% per year. Dominion's load forecasting team has had to rebuild its planning models from scratch — traditional degree-day regression models trained on residential and commercial load growth are useless for predicting when a hyperscaler will announce a 400 MW campus expansion based on AI infrastructure demand signals that don't appear in any utility dataset. The AI irony here is real: the AI infrastructure boom is what's causing the load surge that requires better AI forecasting to manage. Dominion has deployed ML-based load forecasting that incorporates data center construction permit filings, power purchase agreement announcements, and real estate data on leased critical-power capacity — signals that anticipate load additions 18-36 months before the meters turn on. The SCC approved Dominion's 2024 Integrated Resource Plan update with explicit provisions for data-center-driven load growth, and Dominion has committed to accelerating transmission build-out in Northern Virginia to serve the load. The practical constraint operators report: even with better forecasting, transmission permitting timelines (5-7 years for new 500 kV lines in Northern Virginia) lag load-growth timelines by years, creating a structural grid constraint that AI forecasting can anticipate but not solve.
The Coastal Virginia Offshore Wind project, under construction by Dominion Energy with Stonepeak as an infrastructure partner, will deliver 2.6 GW from 176 offshore turbines roughly 27 miles off Virginia Beach. The first phase was commissioned in 2020 (two-turbine pilot), with commercial-scale construction underway through the mid-2020s. Integrating 2.6 GW of variable offshore wind into a grid that also serves data centers, nuclear baseload, and a population of 2.7 million requires AI grid operations capabilities that Dominion is building in parallel with the physical construction. Wind power forecasting for offshore turbines is materially different from onshore: ocean surface wind patterns, sea surface temperature effects, and wake interactions between turbines in large arrays produce variability that standard NOAA weather products don't capture at the spatial resolution needed for real-time dispatch decisions. Dominion has contracted with DTN and other specialized marine weather AI providers for CVOW-specific wind forecasting, and its Energy Management Center in Richmond is building the SCADA integration layer that will allow real-time offshore wind output to be coordinated with nuclear baseload dispatch and gas peaking. The North Anna Power Station in Louisa County (two Dominion-operated PWR units) and the Surry Power Station in Surry County (two older units) provide the baseload floor that makes CVOW integration manageable — without nuclear, the intermittency of 2.6 GW of offshore wind would require much larger storage or gas backup. AI predictive maintenance on North Anna's non-safety-critical balance-of-plant systems has been active since 2022 and has contributed to improved capacity factors on both units.
Appalachian Power's Virginia service territory — from Roanoke west through the New River Valley and into the coalfield counties of southwest Virginia — is structurally different from Dominion's eastern footprint. Load is declining in the legacy coal counties as Appalachian Power retires coal generation under Virginia's VCEA mandates, and the customer base includes a mix of industrial customers (Virginia Tech's Blacksburg campus, Celanese's Narrows facility, paper and chemical operations) and rural residential accounts with high heating loads. AI applications in Appalachian Power's Virginia territory are shaped by these contrasts: demand-response programs targeted at industrial flexible loads (Celanese and similar large customers can participate in PJM demand response auctions), predictive maintenance on an aging distribution system serving geography that includes river valleys and ridgelines that stress line hardware, and customer automation for a service territory where average household income is lower than Dominion's Northern Virginia footprint. The SCC's cost-recovery mechanisms for distribution modernization apply to Appalachian Power as they do to Dominion, and AEP's national grid modernization program provides technology access that individual state subsidiaries couldn't fund independently. Virginia's VCEA creates a CCOS (cost of compliance filing) mechanism that allows Appalachian Power to recover the cost of renewable transition investments, and AI tools that support VCEA compliance tracking and renewable integration planning are increasingly part of the utility's technology roadmap.
Connecting AI systems to existing business infrastructure and workflows
Workflow automation using AI, including Make.com-style automation and RPA
Predictive models, data analysis, and ML pipeline development
Image recognition, object detection, video analysis, and visual inspection systems
Dominion is managing the data center load surge through a combination of transmission expansion, advanced load forecasting, and demand-response programs for large customers. The transmission build-out includes new 500 kV lines in Northern Virginia that require multi-year permitting. On the AI side, Dominion's load forecasting has been rebuilt around data-center-specific leading indicators — construction permits, PPA announcements, and critical power lease data — that provide 18-36 month advance notice of load additions. The SCC approved Dominion's 2024 IRP, which projects continued data center growth through 2030 and includes provisions for accelerated transmission investment. Data center operators in Data Center Alley who want to participate in demand-response programs can access PJM's Emergency Load Response Program for large commercial customers.
Offshore wind forecasting at the CVOW scale requires specialized marine weather AI — specifically, high-resolution ocean-surface wind models that account for sea surface temperature effects and turbine wake interactions that standard NOAA NWS products don't model at the necessary spatial resolution. Dominion uses contracted marine weather analytics providers for day-ahead and intra-hour wind forecasting at the CVOW array. The Energy Management Center in Richmond integrates CVOW output forecasts with real-time nuclear dispatch, gas peaking availability, and PJM market signals to optimize the generation portfolio. The core challenge is managing ramp events — periods when CVOW output drops by hundreds of megawatts within 30-60 minutes due to weather front passage — without triggering frequency deviations that require PJM emergency actions.
The Virginia State Corporation Commission allows Dominion to recover qualifying transmission and distribution investments through rate adjustment clauses (RACs) and base rate cases. The VCEA created a specific regulatory mechanism for clean energy transition investments that includes associated grid modernization. AI-enabled distribution automation, advanced metering infrastructure, and demand-response programs that support VCEA compliance are generally recoverable through these mechanisms. The SCC scrutinizes Dominion rate cases closely — Virginia has had contentious rate proceedings in recent years — but reliability-improving technology investments with documented customer benefit typically receive favorable regulatory treatment. Appalachian Power follows a similar cost-recovery framework under AEP's Virginia regulatory relationships.
Large data center operators in Data Center Alley can participate in PJM's demand response programs through Dominion's large-commercial DR program, which pays participants for curtailing load during PJM emergency events. AI-managed cooling and UPS systems in hyperscale data centers can defer non-critical cooling load for 30-60 minute windows without affecting compute operations, creating a demand-response resource. AWS, Microsoft, and Google all have sustainability commitments that align with demand response participation. The practical challenge is that most hyperscale data centers in Ashburn operate at near-100% utilization, which limits curtailment flexibility — the better demand-response candidates are colocation providers with mixed-workload tenants who have load-shedding agreements. Several energy management firms in the Ashburn and Reston corridor specialize in data center demand response program design.
Virginia industrial customers pursuing PJM demand response through an aggregator typically pay $0 upfront — aggregators take a revenue share from PJM capacity and energy payments. The AI building management systems that enable demand response cost $100,000-$400,000 for a large industrial facility, recoverable through energy savings and DR revenue over 3-5 years. Independent generators in PJM's Virginia zone typically invest $200,000-$800,000 in predictive maintenance AI for major rotating equipment, with ROI driven by PJM capacity market revenue protection — an unplanned outage during a PJM performance assessment interval can result in capacity market penalties that dwarf the cost of the monitoring system.
Reach Virginia businesses looking for your expertise.