Lesson 6 examines the rapidly expanding electricity demands associated with artificial intelligence and hyperscale data centers. The central case is a 2026 energy-supply study for a large data center campus in Virginia. The report estimates facility demand, compares grid electricity, solar photovoltaic systems, virtual power purchase agreements, renewable energy certificates, natural-gas generation, and battery storage, and recommends several combined energy strategies.

The report is also an example of how technical analysis embeds ethical choices. The authors must define the problem, select a study boundary, estimate an operating load, choose technologies for comparison, forecast costs and emissions, determine what counts as reliable power, and decide how much weight to give deployment speed, environmental impacts, public costs, corporate commitments, and community concerns. Each choice affects which strategy appears feasible and defensible.

The report was prepared as an independent academic project using public information and modeled assumptions. The project team did not represent QTS Realty Trust, and measured operating data from the selected facility were unavailable. Those limits are part of the case. Students should evaluate the strength of the evidence, the role of uncertainty, and the difference between observed data and modeled claims.

Central Question: How do assumptions, system boundaries, accounting methods, and comparison criteria shape which data-center energy strategy appears feasible and ethically defensible?

Learning Objectives

• Distinguish measured, estimated, assumed, and projected claims in a technical report.
• Explain how facility capacity, load factor, reliability requirements, and study horizon shape an energy-supply analysis.
• Distinguish physical electricity supply from contractual renewable-energy claims.
• Evaluate whether cost, emissions, and risk comparisons use compatible boundaries and methods.
• Identify stakeholders who receive detailed treatment and stakeholders who remain weakly represented or outside the analysis.
• Apply Ethics Matrix C to consequential methodological choices in the report.
• Assess whether the report’s recommendations follow from the evidence, assumptions, and weighting of decision criteria.

Core Case Report

Baldasare, Kevin; John Gossman; Neelesh Ramseebaluck; and Christopher Simon. 2026. Power and Energy Supply Strategy for Hyper-Scale Data Centers in Virginia. Final Project Report, EME 589, The Pennsylvania State University.

The report models the QTS Richmond 1 campus as a 238 MW facility and uses a 95 percent base-case load factor to estimate annual electricity consumption of approximately 1.98 TWh. Students should treat the load estimate, hourly demand curve, future costs, emissions trajectories, and implementation timelines as analytical claims that require close reading.

Two-Week Lesson Structure

Week 1: Constructing the Data-Center Energy Problem

Required Reading:

• Executive Summary, pp. 2–4
• Overview of the Problem and Study Scope, pp. 7–9
• Stakeholders for Data Center and Energy Projects, pp. 9–11
• Data Center Load Profile, pp. 17–21
• Final Comparison Table, p. 80, read as a preview of the report’s conclusions

Companion reading: International Energy Agency, Key Questions on Energy and AI, Executive Summary

Reading Focus: Week 1 asks how the report establishes the scale and character of the energy problem. Students should identify the evidence supporting the 238 MW facility capacity, the assumed load factor, the annual energy estimate, the modeled hourly load profile, the reliability requirements, the 20-year study period, and the selection or exclusion of energy options. Mark important claims using four categories: measured, estimated, assumed, and projected. The categories help separate direct evidence from calculations, modeling choices, and future scenarios.

Week 1 Work: Complete the Assumption and Scope Audit in Canvas. Identify three consequential assumptions, two major exclusions, one term that requires clearer definition, one stakeholder who is absent or weakly represented, and one reasonable alternative assumption that could change the analysis.

Week 2: Comparing Strategies and Producing a Recommendation

Required Reading:

• On-site solar scale and output, pp. 31–34
• Off-site solar contribution to demand, pp. 45–46
• VPPAs and Renewable Energy Certificates, pp. 54–60
• On-site Natural Gas Power Plants, pp. 60–69
• Energy Storage, pp. 69–74
• Comparison Analysis, Recommendations, and Conclusions, pp. 74–85

Companion reading: Greenhouse Gas Protocol, Scope 2 Guidance Executive Summary

Reading Focus: Week 2 asks whether the report compares the energy pathways through compatible methods and whether the recommendations follow from the evidence. Pay particular attention to differences among physical electricity supply, annual renewable-energy matching, contractual renewable attributes, direct emissions, life-cycle emissions, generation, storage, facility reliability, and grid resilience.

The GHG Protocol reading distinguishes location-based emissions, which reflect the grid serving the facility, from market-based emissions associated with qualifying contractual instruments. Use that distinction to evaluate the report’s treatment of VPPAs, RECs, grid electricity, and the proposed natural-gas-plus-VPPA strategy.

Week 2 Work: Complete Ethics Matrix C and the Recommendation Audit in Canvas. Select three to five consequential choices in the report. For each choice, identify the technical or methodological decision, the rationale, the value embedded in the decision, the affected stakeholders, a reasonable alternative, and the likely effect of that alternative on the final recommendation.

How to Use the Lesson Pages

The Lesson 6 webpages provide analytical tools for reading the report. The webpages do not reproduce the report’s technology descriptions, cost tables, or calculations. Use the report for the case evidence and the webpages for distinctions, questions, and methods of close reading.

Assignment 6 Summary

Assignment 6 asks you to apply Ethics Matrix C to three to five consequential methodological choices in the report. The analysis should explain how assumptions, boundaries, evidence, accounting methods, and decision criteria influence the preferred energy-supply strategy. The complete instructions, word count, submission requirements, and evaluation criteria are provided in Canvas.

Main Point

Technical reports do not produce recommendations from data alone. Recommendations emerge from choices about what to measure, what to estimate, what to assume, what to exclude, how to compare alternatives, and which values receive priority. Lesson 6 uses the data-center energy case to make those choices visible and available for ethical analysis.

Source Notes

• Core report: Power and Energy Supply Strategy for Hyper-Scale Data Centers in Virginia (2026), selected sections listed above.
• International Energy Agency: Key Questions on Energy and AI (2026), Executive Summary.
• Greenhouse Gas Protocol: Scope 2 Guidance, Executive Summary.

Reading the Report Before Judging the Recommendation

The central report for Lesson 6 evaluates possible energy-supply strategies for the QTS Richmond 1 data center campus in Virginia. Before evaluating any technology or recommendation, students should examine how the report defines the problem that the energy strategy is intended to solve.

The authors must select a facility, establish an expected load, define reliability, choose a time horizon, determine which technologies belong in the comparison, identify a baseline, and decide which costs, emissions, risks, and stakeholder concerns will receive attention. Those choices construct the decision space.

Central Question: What must be assumed before the report can define the scale and character of the data-center energy problem?

The Report as a Constructed Analysis

The report selects the QTS Richmond 1 campus as its case and describes an estimated maximum electrical capacity of approximately 238 MW. Detailed operating data from the facility were not publicly available, so the authors developed a representative demand model based on published characteristics of hyperscale AI data centers.

The base case assumes a 95 percent load factor. Multiplying 238 MW by 8,760 hours per year and by the assumed load factor produces estimated annual electricity demand of approximately 1.98 TWh. The report also constructs an hourly demand profile with a relatively flat load and a modest afternoon increase associated with cooling. The calculation is transparent, but transparency does not convert an assumption into a measurement.

238 MW × 8,760 hours × 0.95 = approximately 1.98 TWh per year

The estimate becomes the basis for every later comparison. A different load factor changes required generation, the percentage of demand met by solar, grid purchases, fuel use, emissions, storage requirements, and twenty-year costs.

Four Kinds of Claims

Close reading begins by separating direct evidence from calculations, modeling choices, and future scenarios.

A projected twenty-year cost may combine measured historical prices, estimated current costs, assumed escalation rates, and projected future demand.

Capacity, Load, and Energy

Power and capacity are measured in watts, kilowatts, or megawatts. Energy is power used or produced over time and is measured in kilowatt-hours, megawatt-hours, or terawatt-hours.

The report treats the selected data center as a near-baseload consumer with high and relatively stable demand. That representation favors resources that can provide continuous or dispatchable power. Students should ask whether workload management, staged buildout, or demand flexibility could alter the comparison.

Scope, Boundaries, and Exclusions

The report evaluates electricity-supply strategies over twenty years and uses continued purchases from Dominion Energy Virginia and the PJM grid as the main baseline. The analysis excludes the cost of constructing and operating the data center and excludes energy-efficiency and water-efficiency measures.

Conventional nuclear power, small modular reactors, enhanced geothermal systems, and offshore wind are also excluded because the authors judge them to have low feasibility, long development times, high costs, or insufficient maturity. Each exclusion may be defensible within a limited study. Each exclusion also narrows the possible conclusions.

Questions Raised by the Study Boundary

• Should reducing, shifting, or staging electricity demand be compared with adding generation?
• Does a twenty-year period compare technologies with different construction times and operating lives fairly?
• Which costs are assigned to the owner, utility, ratepayers, or public?
• Which environmental effects are quantified, and which remain qualitative?

Stakeholder Visibility

The report identifies owners, tenants, utilities, developers, investors, regulators, residents, suppliers, researchers, advocacy groups, policymakers, and the general public. Presence in a stakeholder table does not guarantee equal analytical treatment. The owner receives detailed cost and reliability analysis, while community concerns often enter as land-use, permitting, opposition, or project risk.

Consider whether the analysis gives sufficient visibility to:

• residential and low-income electricity customers;
• communities near generation, pipeline, substation, and transmission infrastructure;
• water users and communities facing competing water demands;
• workers in fuel, mineral, equipment, and construction supply chains;
• future ratepayers and residents who may inherit costs or emissions.

How to Read the Assigned Sections

Annotate the report rather than summarizing it. For each important claim:

1. Identify the claim.
2. Classify it as measured, estimated, assumed, projected, or a combination.
3. Locate the source, calculation, or stated rationale.
4. Identify the stakeholders affected by accepting the claim.
5. Consider a reasonable alternative and explain how it could change the analysis.

Questions for Close Reading

1. Which claims about the selected campus are based on measured operating data?
2. Which assumption most strongly determines the size of the energy problem?
3. What is treated as fixed that could be treated as variable?
4. How does the modeled load profile affect the apparent value of solar, gas, grid power, and storage?
5. Which options are excluded before the comparison begins?
6. Which stakeholders receive quantitative attention, and which appear mainly as sources of risk or opposition?
7. What additional data would be needed before the demand model could support an actual infrastructure decision?
8.  

Week 1 Work: Assumption and Scope Audit

Complete the Assumption and Scope Audit in Canvas. Identify three consequential assumptions, two important exclusions, one term that requires clearer definition, one stakeholder who is absent or weakly represented, and one reasonable alternative assumption that could change the results. Support each point with a specific reference to the assigned report pages.

Required Reading for Part 1

• Executive Summary, pp. 2–4.
• Overview of the Problem and Study Scope, pp. 7–9.
• Stakeholders for Data Center and Energy Projects, pp. 9–11.
• Data Center Load Profile, pp. 17–21.
• Final Comparison Table, p. 80, read as a preview of the report’s conclusions.

Companion reading: International Energy Agency, Key Questions on Energy and AI, Executive Summary

Main Point

The report’s energy problem is produced through choices about facility size, expected utilization, hourly demand, reliability, study period, available technologies, and stakeholder relevance. Close ethical analysis begins by making those choices visible. Students should understand how the problem was constructed before judging which solution appears best.

Comparing Pathways That Provide Different Services

The central report compares grid electricity, on-site solar, off-site solar, virtual power purchase agreements, renewable energy certificates, natural-gas generation, and battery storage. The options do not perform the same function. Some supply physical electricity. Some provide dispatchable capacity. Some provide contractual renewable-energy attributes. Battery storage shifts and stabilizes electricity but does not create primary energy.

A comparison becomes ethically and technically defensible only when the analyst states what service is being compared, which boundaries are used, and which costs, emissions, risks, and uncertainties are included. A single summary table can make unlike options appear directly comparable even when the underlying methods differ.

Central Question: Are the energy strategies compared through compatible services, boundaries, metrics, and assumptions?

What Each Pathway Provides

The first step is to identify the primary service provided by each pathway. The table below summarizes the role each option plays in the report.

Distinctions Needed for Comparison

Power, Capacity, and Energy

Power and capacity describe the rate at which electricity is demanded or supplied. Energy describes electricity used or produced over time. A solar project can generate a meaningful amount of annual energy while providing limited output during many hours. A data center can require both annual energy and dependable power at specific times.

Annual Matching and Hourly Matching

Annual matching compares total renewable generation with total annual consumption. Hourly matching asks whether generation and consumption occur during the same hours. Annual equality does not establish continuous physical supply, local delivery, or resource adequacy.

Physical Supply and Contractual Claims

A physical supply arrangement delivers electricity through a grid connection or on-site generator. A VPPA is a financial contract associated with a separate generating project. A REC represents the renewable attribute of one megawatt-hour of generation. VPPAs and RECs can support market-based renewable claims without changing the physical electricity delivered to the data center.

Generation and Storage

Generation converts an energy resource into electricity. Storage receives electricity, retains part of it, and delivers electricity later. Storage can provide rapid response, power quality, backup, peak management, and renewable integration. Storage performance depends on duration, power capacity, losses, degradation, operating strategy, and the source of charging electricity.

Facility Reliability and Grid Resilience

Facility reliability concerns the data center's ability to maintain operations. Grid resilience concerns the larger system's ability to withstand and recover from disruptions. A strategy can improve facility independence while reducing investment in shared infrastructure or shifting risks to fuel, local emissions, or other customers.

Location-Based and Market-Based Electricity Accounting

The Greenhouse Gas Protocol distinguishes two methods for reporting emissions associated with purchased electricity.

The two methods answer different questions. A data center can report market-based renewable procurement while continuing to receive electricity from a fossil-intensive regional grid. A transparent analysis should distinguish the facility's physical electricity use, contractual attributes, and any claim that the procurement caused or enabled additional renewable generation.

Additionality and Renewable-Energy Claims

Additionality concerns whether a purchase or contract causes new renewable generation or another change beyond what would otherwise occur. A long-term VPPA associated with a new project may provide stronger evidence of additionality than an unbundled REC from an existing facility. Additionality is still an empirical and contractual claim. The analysis should state the evidence supporting it.

A Compatibility Test for Energy Comparisons

Before comparing two pathways, ask whether the analysis uses compatible definitions and boundaries.

• Functional service: Are the options providing energy, firm capacity, reliability, attributes, or a combination?
• Demand level: Are all pathways evaluated against the same modeled load?
• Time horizon: Are capital-intensive and long-lived assets evaluated over a fair period?
• System boundary: Are upstream fuel, manufacturing, transmission, and end-of-life effects treated consistently?
• Emissions boundary: Are direct, life-cycle, location-based, and market-based emissions clearly separated?
• Financial treatment: Are financing, taxes, fuel, replacement, infrastructure, and residual value included consistently?
• Risk treatment: Are permitting, construction, market, fuel, interconnection, and operational risks assessed on similar terms?
• Uncertainty: Are uncertain assumptions tested with comparable sensitivity ranges?

Where the Report's Comparisons Require Close Reading

The report's final comparison tables bring costs, emissions, timelines, demand coverage, and risk into a common format. The underlying calculations use different methods, which makes the tables useful for analysis and also requires caution.

• Grid emissions are projected to decline through an assumed annual reduction in emissions intensity.
• Solar pathways use life-cycle greenhouse-gas factors rather than direct operational emissions.
• Natural-gas pathways emphasize direct combustion emissions; upstream methane and full fuel-cycle effects require separate attention.
• The VPPA pathway is assigned renewable life-cycle emissions even though the data center continues to rely physically on the grid.
• Battery emissions focus on manufacturing; charging losses, charging source, degradation, augmentation, and replacement may change the result.
• Costs depend on different assumptions about future grid prices, fuel prices, financing, tax treatment, project life, and market contracts.

These differences do not make comparison impossible. They require the analyst to state what each number represents and avoid presenting results from different boundaries as though they measure the same phenomenon.

Reading the Strategy Sections

Questions for Close Reading

• Do all pathways provide the same combination of energy, capacity, reliability, and environmental attributes?
• Which costs are included for one pathway and omitted for another?
• Are direct emissions, life-cycle emissions, and contractual emissions claims clearly distinguished?
• What does a VPPA change physically, financially, and contractually?
• How does the selected treatment of battery storage affect the comparison?
• Which pathway is most sensitive to changes in the modeled data-center load?
• Which result would benefit most from sensitivity analysis?
• What comparison appears precise in the final table but depends most strongly on uncertain assumptions?

Week 2 Work: Pathway Comparison Audit

Complete the Pathway Comparison Audit in Canvas. Select one pathway and identify its primary service, major data sources, three consequential assumptions, cost and emissions boundaries, treatment of reliability, affected stakeholders, major uncertainty, and one methodological change that could alter the result. Compare the pathway with one other option and explain whether the report evaluates them through compatible services and boundaries.

Required Reading for Part 2

Report sections: on-site solar, pp. 31-34; off-site solar, pp. 45-46; VPPAs and RECs, pp. 54-60; natural gas, pp. 60-69; battery storage, pp. 69-74; and comparison analysis, pp. 74-80.

Companion reading: Greenhouse Gas Protocol, Scope 2 Guidance Executive Summary

Main Point

Energy pathways cannot be compared responsibly through adjacent cost, emissions, and timeline values alone. Analysts must identify the service each pathway provides, use compatible boundaries, separate physical electricity from contractual claims, and disclose the assumptions producing each result.

From Technical Results to a Preferred Strategy

The final sections of the data-center report move from calculations and comparison tables to a set of recommendations. That transition requires judgment. Cost, emissions, reliability, deployment time, interconnection risk, land use, corporate renewable-energy commitments, community response, and public infrastructure do not combine themselves into a single preferred answer. The analysts must decide which criteria carry the greatest weight and which tradeoffs are acceptable.

Embedded ethics appears in those decisions. The report's preferred strategy is shaped by the way the problem was defined, the options included, the assumptions selected, the boundaries used for cost and emissions, and the priority assigned to speed, control, and continuous power. Part 3 asks whether the recommendation follows convincingly from the evidence and how a different set of defensible priorities might alter the result.

Central Question: Which technical and ethical commitments make the recommended strategy appear preferable?

What the Report Recommends

The report recommends different strategies for different development conditions. The recommendations combine several options because no single pathway supplies every service the authors consider necessary.

The report does not recommend simple-cycle natural gas because of its lower efficiency, higher fuel use, and higher emissions. It also does not recommend the small on-site solar project as a central strategy because the project supplies only a very small share of modeled annual demand.

A Recommendation Depends on Weighting

The report gives significant weight to several criteria: continuous power, time to deployment, control over energy supply, modeled twenty-year cost, avoidance of interconnection delays, and the ability to support a corporate renewable-energy claim. The preferred gas-based strategy becomes plausible because combined-cycle generation performs strongly under those criteria.

Other criteria are present but receive different forms of treatment. Carbon emissions are quantified, while community effects, ratepayer burdens, public infrastructure, policy alignment, water use, and long-term fossil fuel dependence are often discussed qualitatively or through project risk. Differences in measurement can affect perceived importance. A precise cost estimate can appear more decision-relevant than a less quantified public burden even when the public burden is ethically significant.

Questions About Weighting

• Why should deployment speed receive the weight assigned to it?
• Whose definition of reliability governs the recommendation?
• How much environmental impact is treated as acceptable in exchange for private control and faster energization?
• Should corporate renewable-energy claims count as a benefit when the facility physically burns natural gas or consumes grid electricity?
• How should costs borne by ratepayers, communities, or future infrastructure users enter the comparison?

Where Embedded Ethics Appears

Ethics Matrix C can be used to trace ethical choices through the report rather than treating ethics as a final comment added after the technical analysis.

Public Concerns: Ethical Claims or Project Risks?

The report recognizes public backlash, electricity-price concerns, environmental impacts, land use, local opposition, permitting, and possible conflict with Virginia's clean-energy direction. The framing of those concerns matters.

A community objection can be treated as evidence of potential harm, a claim about fairness, a request for participation, a permitting constraint, a schedule delay, or a reputational threat. Treating opposition mainly as project risk can convert an ethical concern into a management problem. A stronger analysis should ask what the concern is about, what evidence supports it, which groups are affected, and whether the proposed response addresses the underlying burden.

Uncertainty and the Burden of Proof

The recommendation depends on uncertain estimates of demand, grid prices, fuel prices, technology costs, project timelines, emissions, contractual prices, and future policy. A recommendation can remain useful under uncertainty, but the report should show whether the preferred strategy is robust across reasonable alternatives.

Burden of proof concerns who must demonstrate that a strategy is reliable, affordable, environmentally responsible, and fair. A developer may ask opponents to prove that a project will cause harm. Communities and regulators may instead ask the developer to demonstrate that private benefits will not create unreasonable public costs. Matrix C can make those competing expectations explicit.

Useful Sensitivity Questions

• Would the preferred strategy change under a lower data-center load factor or phased buildout?
• Would slower grid-price growth change the cost advantage assigned to natural gas?
• How would higher fuel prices or upstream methane emissions affect the gas strategy?
• Would a longer study period change the relative value of solar, transmission, or other long-lived infrastructure?
• Would the recommendation change if location-based and market-based emissions were reported separately?
• What happens if community, permitting, pipeline, or air-quality constraints delay the gas project?

Recommendation Audit Using Ethics Matrix C

1. Use the following process to connect close reading of the report with Ethics Matrix C.
2. Identify a consequential methodological choice in the report.
3. State the evidence, calculation, or rationale used to support the choice.
4. Explain the technical effect of the choice on the comparison.
5. Identify the value or priority embedded in the choice.
6. Identify the stakeholders who benefit, bear risk, or remain outside the analysis.
7. Propose a reasonable alternative assumption, boundary, metric, or weighting.
8. Explain whether the alternative could change the recommendation.

Questions for Close Reading

1. Does the preferred strategy follow from the evidence, or from the weighting of the evidence?
2. Which assumption creates the greatest risk that the recommendation will be wrong?
3. Which stakeholder interest receives the greatest priority?
4. Which burden is least visible in the final comparison?
5. Are community and ratepayer concerns treated as substantive ethical claims or mainly as project risks?
6. How does the treatment of VPPAs affect the apparent environmental credibility of the gas-based recommendation?
7. What additional data or sensitivity analysis would be needed before the recommendation could support an actual decision?
8. Would the recommendation remain defensible under a different definition of reliability, sustainability, or public benefit?

Assignment 6 Connection

Assignment 6 asks you to apply Ethics Matrix C to three to five consequential choices in the report. Your analysis should remain grounded in the assigned readings. For each choice, explain the methodological decision, the report's rationale, the values and stakeholders involved, a reasonable alternative, and the likely effect of that alternative on the preferred strategy.

The goal is not to decide whether data centers, natural gas, solar energy, or artificial intelligence are generally good or bad. The goal is to evaluate how a specific technical report constructs evidence, compares alternatives, handles uncertainty, and converts results into a recommendation.

Required Reading for Part 3

Report sections: Comparison Analysis, pp. 74-80; Recommendations, pp. 80-84; and Conclusions, pp. 84-85.

Main Point

Technical findings do not determine a recommendation without judgment. The preferred strategy reflects choices about which problem to solve, which evidence to trust, which risks to tolerate, which stakeholders to prioritize, and how to weigh cost, speed, reliability, emissions, and public responsibility. Embedded ethics becomes visible when those choices are identified, justified, and tested against reasonable alternatives.