Infrastructure Past, Present, and Future Casebook/Three Mile Island

Introduction

Three Mile Island, being renamed to the Crane Clean Energy Center with its reopening.

The Three Mile Island (TMI) nuclear power plant, located near Harrisburg, Pennsylvania, is one of the most well-known nuclear facilities in the United States due to the partial meltdown of its Unit 2 reactor in 1979. One of the most serious accidents in U.S. commercial nuclear power history, the meltdown led to major regulatory reforms and heightened public skepticism toward nuclear energy. Constellation Energy's 2024 plan to restart the Unit 1 reactor to provide power for Microsoft AI data centers offers an opportunity to examine the history and future of the plant.

Timeline of Events

The Unit 1 reactor entered operation in 1974, and Unit 2 entered operation in 1978. On March 28, 1979, a series of mechanical failures and human errors caused a partial meltdown in the Unit 2 reactor, prompting inquiry from the Nuclear Regulatory Commission and the temporary shutdown of the Unit 1 reactor. In 1985, cleanup efforts progressed to a point that removal of damaged fuel could begin and the Unit 1 reactor could resume operation. The Unit 1 reactor remained in operation until 2019.

Animated sequence of events associated with the 1979 Three Mile Island nuclear accident.

In September 2024, Constellation Energy announced its plan to restart operation of Unit 1 and sell power exclusively to Microsoft to meet AI energy demands. Licensing and approval for the project is, as of February 12, 2025, still ongoing.

Key Actors

   • Metropolitan Edison Company (Met Ed): The original operator of the Three Mile Island facility during the 1979 accident, later absorbed into larger utility companies.

   • Nuclear Regulatory Commission (NRC): The federal agency responsible for overseeing nuclear safety and operations. The NRC played a crucial role in investigating the 1979 accident and enforcing new safety measures. To deal with the novel occurrence of reopening an already decommissioned plant, the NRC has established an inspection program for reactor reopenings.

   • Constellation Energy: The current owner of TMI’s Unit 1 and the driving force behind its planned reopening. The company specializes in nuclear energy production and aims to leverage TMI’s existing infrastructure for clean power generation.

   • Microsoft: The tech giant has entered into a 20-year agreement with Constellation Energy to use electricity generated by TMI’s Unit 1 for its AI data centers. The project aligns with Microsoft’s commitment to achieving carbon-negative operations by 2030.

   • State and Local Governments: The Pennsylvania government and local municipalities are key stakeholders in the decision to reopen TMI. Economic and environmental concerns have been raised by different groups, with debates on job creation versus potential risks.

   • Environmental and Advocacy Groups: Organizations advocating for and against nuclear power, including groups concerned about safety risks, radioactive waste disposal, and long-term environmental impacts.

Public Response

Response to the 1979 Incident

Although the meltdown resulted in minimal radioactive releases with no detectable health effects on plant workers or the public, it had a profound impact on public sentiment. The immediate aftermath was characterized by confusion, further exacerbated by conflicting reports. Residents were uncertain about evacuation orders, leading to widespread anxiety, all the while media coverage intensified public concern about potential catastrophic outcomes.

The incident led to a significant shift in public opinion regarding nuclear energy. Surveys around near the time of the meltdown indicated a decline in support for nuclear power, with many Americans expressing heightened safety concerns. This shift influenced policymakers, resulting in stricter regulatory measures and a slowdown in the development of new nuclear facilities. The TMI accident brought about the need for improved safety protocols and transparent communication between nuclear operators, regulators, and the public.

Response to the Proposed Restart

In September 2024, Constellation Energy announced plans to restart TMI's Unit 1, which had been shut down in 2019 due to economic challenges. The restart of the plant is backed by a 20-year power purchase agreement with Microsoft, which hopes to supply energy to its data centers for AI. The proposal has elicited a variety of reactions. Supporters of the restart emphasize the benefits of adding approximately 835 megawatts of carbon-free energy to the grid. They highlight the potential economic advantages, including the creation of an estimated 3,400 direct and indirect jobs and the generation of over $3 billion in state and federal taxes. Opponents express concerns based on historical apprehensions and environmental considerations. Environmental groups raise issues about the management of radioactive waste and the long-term sustainability of nuclear energy. The proximity of Unit 1 to the decommissioned and damaged Unit 2 adds to the unease, with calls for comprehensive safety evaluations before any restart. Regulatory bodies, including the NRC, have initiated reviews to assess the feasibility and safety of the proposed restart. Public meetings have been held to gather input, during which residents have voiced both support and concerns. Some community members seek assurances about emergency preparedness and environmental protection, while others advocate for the economic and clean energy benefits the project promises. The collaboration with Microsoft underscores the tech industry's growing demand for reliable, carbon-free energy sources to power expansive data centers, especially with the rise of artificial intelligence applications. This partnership reflects a broader trend of corporations seeking sustainable energy solutions, potentially influencing public opinion on nuclear energy's role in addressing climate change.  

The public response to the TMI Unit 1 restart proposal is deeply intertwined with the legacy of the 1979 accident. While there is optimism about economic growth and clean energy production, lingering safety concerns and environmental considerations continue to shape the discourse. As regulatory evaluations proceed, the balance between these perspectives will play a crucial role in determining the project's future.

Evaluating the Disaster Response and Reopening

The planned reopening of Three Mile Island reflects broader shifts in the global energy landscape. As AI and cloud computing demand skyrockets, major tech companies like Microsoft are seeking stable and sustainable energy sources. Nuclear power, despite its controversial history, remains one of the most reliable carbon-free energy options available. The decision to restart TMI also underscores the financial and technological viability of repurposing existing nuclear plants rather than building new ones from scratch. However, concerns remain about regulatory hurdles, safety measures, and public perception given the facility’s history.

Ostrom et al lay out a framework for evaluating institutional performance in providing infrastructure, used here for evaluating both response to the 1979 incident and the reopening as projects based on the metrics of efficiency, equity, accountability, and adaptability.

Efficiency

The efficiency metric is a gauge of whether the benefits of a project outweigh operating costs. Following the 1979 incident, the NRC introduced stricter safety protocols, operator training, and emergency planning to prevent future accidents. Safety precautions often come with higher operating costs; if the cost of operating a reactor safely—including the cost of safe disposal of waste materials—is greater than the benefit of the energy it produces, there is little reason to operate the reactor at all.

The response to the 1979 incident highlighted inefficiencies in communication and crisis management, which worsened the situation. Ostrom emphasizes that effective governance requires systematic evaluation to ensure reforms genuinely improve performance rather than add bureaucratic burdens. Some scholars argue that post-TMI regulations led to excessive oversight, reducing the competitiveness of nuclear energy. While safety improved, the financial burden on energy providers suggests that a cost-benefit analysis could have optimized efficiency in policy implementation.

The actual efficiency of reopening the plant remains to be seen based on Constellation Energy's success at delivering the needed energy in a cost-effective manner.

Equity

Ostrom et al define equity, or fairness, with two different metrics: fiscal equivalence, the idea that "those who benefit from a service should bear the burden of financing that service," with those deriving greater benefit paying more, and redistribution, moving money from wealthier areas or populations to lower-income areas or population. These aims are, frequently, in tension with each other.

Skarbek argues that fiscal equivalence ensures accountability by linking service costs to the benefiting population. Three Mile Island adds a complication to this calculation with the existence of a disserviced population in the aftermath of the incident. While population health studies in the aftermath were initially funded, Wilson et al found that long-term health follow-ups for affected communities were never fully conducted, raising concerns about misallocated public resources.

Skarbek highlights that broad-based taxation should support public goods like nuclear safety, ensuring costs align with beneficiaries. However, TMI’s financial burdens were unevenly distributed among stakeholders, suggesting future nuclear policies should include mechanisms to better allocate financial responsibility.

The reopening of the plant faces problems with fiscal equivalence as well; Microsoft being the sole power buyer should, in theory, mean achieving fiscal equivalence should be easy—have Microsoft pay for all of it. However, federal money being involved does put some risk and cost onto taxpayers. Constellation Energy reportedly seeking a $1.6 billion federal loan would, potentially, shift risks onto taxpayers in the event of a default, possible in the event of a failure to reopen; this is in addition to federal tax credits for the sale of energy, potentially worth $200 million per year.

Redistribution is also complicated to assess when examining the 1979 incident, which disproportionately affected low-income residents in Pennsylvania, who had limited evacuation options and faced long-term health concerns. Over 35,000 residents were surveyed for health impacts, yet many studies on fetal and maternal health were unpublished or lacked follow-up. This failure in long-term health monitoring shifted risks onto affected communities while prioritizing industry recovery. Ostrom warns that policy reforms must systematically consider marginalized groups rather than focus solely on institutional stability.

Accountability

Accountability assesses whether institutions are transparent and responsible to the public. After the TMI accident, public trust in the NRC was seriously undermined due to poor communication and a lack of clear leadership. Although reforms followed, many were superficial and failed to address structural problems like poor oversight and management disconnects. The NRC’s internal confusion, as highlighted by the Kemeny Commission, raised concerns about its capacity for self-governance.

Ostrom argues that public institutions must include mechanisms for oversight and sanction. However, in the TMI case, accountability was limited: few were held responsible, and corporate liability remained weak due to the Price-Anderson Act. Temples shows that NRC reforms were shaped more by political calculations than by true responsibility. As Skarbek emphasizes, genuine accountability requires both institutional transparency and meaningful consequences—both of which were lacking.

The NRC publishes documents related to the plant reopening, including details on licensing applications and the slide presentation from the February 19, 2025 meeting where Constellation proposed its emergency plan for the reopening. This ensures some level of transparency, and responsibility to the public through the NRC. These public hearings do offer opportunities for comment, including from opponents of the project, but as the plant is opening for a business-to-business deal, public accountability can only exist within the regulatory structure.

Adaptability

Adaptability reflects how institutions adjust to new risks and lessons. After TMI, the NRC introduced reforms like improved training and monitoring. Yet Ostrom stresses that adaptive governance requires ongoing feedback and learning, not one-time changes.

Temples argues that deeper reforms were resisted, especially proposals to restructure the NRC for more effective leadership. While some technical upgrades occurred, bureaucratic inertia persisted. As Robert S. Walker warned in a congressional hearing, industry refusal to engage openly risked further erosion of public trust. Without honest communication and cultural change, adaptation remains shallow. His warning—that “ostrich-like attitudes” could endanger the future of nuclear power—highlights the stakes of institutional inflexibility. This warning applies to the current situation in more ways than one, considering current concerns about AI. Adapting to novel technology in AI will be vital, as demand in the future will determine whether the plant reopening is justified.

This reopening is only the second of its kind, after the still-ongoing reopening of the Palisades Nuclear Plant in Michigan. Adaptability will be key in this novel situation, and, while the NRC has established an inspection program for this new circumstance, it remains to be seen if the regulatory body will be sufficiently adaptable.

Lessons Learned

Three Mile Island looms large over the landscape of nuclear energy in the US, and, fortunately, no disaster of similar scale has occurred since. Many of the lessons of the 1979 incident have been taken into account, resulting in regulatory and technological improvements. The reopening, despite being in its early stages, does offer some insights of its own that may be useful to examine.

The single-customer model only highlights the wide-ranging impacts infrastructure can have. Even without using power generated by a plant, the surrounding area is affected; Constellation cites job creation, a positive impact, while residents have raised concerns about fuel storage, changes to traffic, and water usage, the latter especially a concern with climate change affecting water scarcity and flood risks. The footprint of any project of this scale will be larger than its physical bounds, and if evaluation is solely focused on direct parties to the project, significant impacts may be disregarded.

The reopening also offers lessons on how market trends can impact infrastructure. The plant was shut down due to "economic reasons" and is reopening due to AI demand; demand fluctuations can affect whether an infrastructural element operates for its full lifespan and, in this case, bring it back into operation afterward. Infrastructure's footprint lasts long after it ceases operation; it's important to consider what happens should a facility need to shut down before it's expected to.

Further Reading

• The NRC's backgrounder on the 1979 meltdown
• The NRC's inspection manual chapter on reactor restarts
• Constellation's presentation on the proposed emergency plan

Discussion Questions

• How does the single-buyer model for the reopened plant change evaluation of its infrastructural effectiveness? Should a nuclear power plant sell its power to a singular entity in the first place?
• The plant's history makes for provocative headlines, as noted in an MIT Technology Review article on the subject. How should lawmakers, regulators, and corporations approach situations reported on in this dramatic fashion?
• Three Mile Island's reopening reflects a broader trend of increasing demand for power by data centers. Does the reopening reflect a viable solution to this need?

Acknowledgement and Works Cited

This casebook is a case study on the Three Mile Island nuclear plant reopening by Rawia Elhassan, Yihan Gao, Sage Kerr, Xiaomeng Li, and Jonathan Mulyar, created as part of Infrastructure Past, Present and Future: GOVT 490 (Synthesis Seminar for Policy & Government) and CEIE 499 (Special Topics in Civil Engineering) at George Mason University. This project is under the instruction of Professor Jonathan Gifford at the Schar School of Policy and Government and the Volgenau School of Engineering’s Sid and Reva Dewberry Department of Civil, Environmental, and Infrastructure Engineering