What Utility Grids Teach Us About Lunar Power
Modern grid management is built on assumptions that don't survive contact with the Moon. The lessons worth keeping are the older ones.
The Invisible Assumption Stack
Most people don’t think about electricity very often. When they think about it at all, they think about it the way they think of gasoline — something produced, stored, and consumed at your convenience. The reality is far more precarious. Unlike virtually every other commodity, electricity can’t be meaningfully warehoused. Generation must match demand almost instantaneously, every second of every day. When you flip a light switch, a generator somewhere spins a little harder.
This delicate balancing act plays out across an interconnected physical system spanning thousands of miles, connecting hundreds of millions of endpoints, orchestrated in real time by operators watching frequency deviations measured in hundredths of a hertz. A sustained deviation means something has gone wrong: perhaps a station tripped, a line faulted, or demand spiked beyond forecast. The system must respond in seconds or cascading failures begin. In 2003, a single error in Ohio resulting from a sagging line touching a tree branch triggered a cascade that plunged over 50 million people into a blackout across the US and Canada in under two hours, including the entirety of New York City. The subway stopped running, water systems lost pressure, cell service failed, AC stopped working in the middle of summer. It took four days to restore power fully.1
I’ve spent years working inside this system. What strikes you after enough time isn’t how fragile it is — it’s how reliably it works despite that fragility. The grid runs at five-nines reliability because an enormous apparatus of human judgment, automated controls, and institutional discipline keeps it online.
Now imagine building a system like this on the Moon. In such an austere environment, there is no interconnection to borrow from when a unit trips, no gas pipeline or coal yard to feed backup generation. Every watt must be produced, balanced, and delivered within a closed system where a sustained outage means death. Most writing about lunar power focuses on generation technology and logistics: solar power, nuclear reactors, battery storage. These are interesting and important problems, but they are essentially already solved in that the architecture is understood. The harder, unsolved problems are not technological but operational and institutional. They are problems that the modern grid does not reckon with at all.
Modern grid management is built on an invisible stack of assumptions - scalable generation, interconnection, the regulatory compact - that is taken for granted because these assumptions have been true for a century. As a result, they have become entrenched, defining terrestrial electrical generation, transmission, and distribution.
Scalable generation means you can always get more power if you need it. Peaker plants supply extra capacity when demand exceeds forecast, and market purchases can offset shortfalls. Interconnection means the grid is not one system, but many systems linked. This allows failures to be contained and surpluses to be shared. The regulatory compact means the interests of the utility, the regulator, and the customer are balanced. The rate case mechanism ensures capital recovery and enables adequate capacity at the most affordable rates. This framework allowed the emergence of the “regulated monopoly” model that has enabled utilities to expand the grid and operate it reliably for over 100 years.
The Moon reveals and breaks all of these assumptions at the same time. Not because they are wrong - they’re right, for Earth. But on the Moon, there is no excess supply, and energy waste is not viable given the constraints of lunar settlement. There is no interconnection to borrow from or sell to when demand deviates from forecast. The regulatory compact does not exist because there is no regulator. In fact, it’s likely that in the early stages of lunar settlement, utility, regulator, and customer may be the same entity. The assumptions that have defined the modern grid do not apply on the Moon. What we have to contend with is what to replace them with.
Geography Is Destiny
On Earth, geography shapes power grids, but doesn’t determine them absolutely. Interconnection allows utilities to abstract over terrain, compensating for poor generation siting with transmission. On a small, isolated grid with no interconnection, where you put generation relative to load is not something you can easily hedge against later. Terrestrial grids have evolved over decades, allowing for corrections to early suboptimal decisions. But the engineer’s instinct to “optimize later” does not survive the lunar context. On the Moon, initial design decisions are irreversible, and the stakes asymmetrically higher, with no modern terrestrial parallel. Lunar power grids must therefore be designed around permanent constraints from the outset.
Decisions made about where to place the first lunar settlement will permanently impact lunar development. There are clear pros and cons to any potential settlement location. One proposed settlement location in particular appears most attractive in terms of overall habitability as measured by energy cost and resource accessibility. It should come as no surprise to readers that this location is the Shackleton Crater at the Moon’s south pole. The primary reasons it is optimal are that the crater’s rim receives near-constant sunlight, providing a reliable source of solar energy and reducing, though not eliminating, the reliance on batteries for energy storage; moreover, the crater’s interior contains permanently shadowed regions that are believed to harbor water ice, which can be extracted for oxygen and hydrogen fuel in addition to potable water.
For comparison, consider Mare Tranquillitatis, which has abundant mineral resources and constant line-of-sight to earth, not to mention historical and cultural value, but faces the brutal 14-day lunar night, complicating energy system design. Sitting near the lunar equator on the near side, Mare Tranquillitatis is accessible and would offer stable communication without the need for relays. However, it spends over 350 hours per cycle in darkness, when solar power is unavailable and temperatures drop below -100°C. A settlement here would require a completely different energy system architecture, necessitating nuclear baseload and massive battery energy storage.
Given that the stakes of initial design decisions are so high, geographic site selection becomes more than a critical factor in the success or failure of any lunar habitat. It becomes the defining factor for subsequent human activity on the Moon. The location of the first colony on the Moon will determine not only its own energy architecture and economic viability, but that of future projects on the Moon for decades to come. Geographic site selection is the founding decision. Everything else follows from it.
The Financing Problem
The modern utility system relies on the regulated monopoly model that emerged in the early 20th century. This model grants utilities exclusive territory in exchange for a commitment to provide reliable service to customers at fair rates. This framework is known as the regulatory compact - a tripartite agreement between the utility, regulators, and customers (though how much power customers have in this arrangement is debatable). Under this compact, utilities must deliver service to customers that meets safety and reliability standards. These standards of service are determined by regulatory bodies, and prevent utilities from cutting costs at the expense of service quality. The compact also enables utilities to recover the cost of their operations and grid investments and allows for reasonable profits. Regulators determine which costs are recoverable and how much profit is acceptable. Despite valid criticisms of this model, it has enabled the infrastructure to be built that provides reliable and affordable power to billions of people around the world.
This model is only possible, though, by balancing the competing economic and legal interests of utilities, regulators, and customers as separate entities. On the Moon, this separation is not possible from the outset. It is likely that the entity managing the grid will be the same entity that is overseeing the settlement and the same entity to which the settlers consuming power belong - the regulatory compact collapses into a single entity. What this means is that there is no rate base to recover capital investment and the cost of operations. Traditional utility financing models are therefore not possible on the Moon, at least not in the early stages of settlement.
What replaces utility grid financing on the Moon will more closely resemble colonial and early industrial financing models. Such capital structures were long-horizon, illiquid, and speculative. They had to be tolerant of massive upfront capital investment with extremely deferred returns and high rates of failure. The British East India Company, for example, financed infrastructure for years in places no market existed, on the bet that one would exist. Another characteristic of these models was sovereign-commercial entanglement. Colonial ventures were almost never purely private or purely state - they often had quasi-governmental powers or even directly governed territory. Whoever finances early lunar infrastructure is going to end up with governance authority because, in that environment, infrastructure is governance. It simply isn’t possible to separate management of life support systems from political power in early remote settlements.
There are clear limits to the colonial model analogy, though. Most notably, on the Moon, there is no indigenous economy to exploit. All economic value must be created from scratch. Biological self-sufficiency also looks significantly different, with much higher ongoing operational costs for life support systems.
Other models to consider are the company town or military base model. These are single-operator environments where utility, employer, and landlord are unified. As bootstrap mechanisms, they are highly efficient. Additionally, in the latter, there is no expectation of direct economic return. Rather, these are strategic positions that enable resource extraction, funded by sovereign budgets. Either or both of these models are likely to influence future lunar settlement.
There are well-documented failure modes for these models. Most notably, they enabled exploitation where structural checks that markets and regulators normally provide were absent. On the Moon, these checks must be included in the design from the start rather than hoping they emerge. One way to do that may be an infrastructure concession model, or build-operate-transfer. In such a model, a private entity could develop the critical infrastructure, operate it as long as necessary to recoup costs, and eventually transfer it to whatever governance body emerges.
The funding sources that establish lunar settlement and the governance structures that emerge within will look quite different from modern terrestrial models. They will draw on historical, speculative models for inspiration. At the same time, they should seek to avoid the pathologies of early models. New pioneer models will move lunar settlement beyond base-building and enable prosperous enterprises and thriving communities.
Demand Response When Everything Is Life-Critical
Terrestrial demand response is an economic optimization strategy employed within a resilient system. Lunar demand response, on the other hand, is survival engineering within a brittle one. The tools available to grid operators - energy storage, load management, and forecasting - are similar. However, the core design philosophy of the demand response protocols will be inverted.
Batteries are one key difference. On Earth, storage is increasingly important, but still primarily a market instrument - charge when power is cheap, discharge when it’s expensive. The grid handles real balancing through dispatch and interconnection. On the Moon, batteries will do the job that interconnection and dispatchable generation do on Earth. They won’t be about optimizing electricity rates, but rather keeping people alive during generation fluctuations. Consequently, battery failure becomes a different category of event. Redundancy, degradation monitoring, and replacement logistics of energy storage will become first-order design constraints, less about asset management and more about life-safety.
Another key difference is demand predictability and the consequences of errors in forecasting. While terrestrial demand is cyclical, it has stochastic variation: weather, human behavior, economic cycles. I remember my first visit to a grid operator control room. It was one of the most interesting things I have ever seen. The control room supervisor showing me around told me an interesting fact that seems obvious in retrospect but nonetheless was eye-opening at the time. On the day of the Super Bowl, the control room operators can tell simply by monitoring the grid when the halftime show is over: a predictable spike in demand occurs when 100 million Americans get up from their couches and open their refrigerators at the same time - and the grid meets that demand instantaneously. A lunar habitat’s demand profile is actually more predictable in some ways. A fixed population on a controlled schedule lends itself well to a known life support baseline. But the consequences of a forecast error are asymmetric in a way that has no terrestrial parallel. Errors on Earth lead to dispatch problems and inconvenience for customers; on the Moon, they may be fatal.
Related to the forecasting difference is the trough problem and that of making waste productive. On Earth, excess generation can be curtailed, sold to adjacent grids, or accepted as inefficiency. On the Moon, there are no adjacent grids, and waste heat in a vacuum is a significant thermal management problem in addition to the economic one. The austere environment will push utilities to find productive loads for trough power. These may include water electrolysis for hydrogen and oxygen production, in-situ resource utilization (ISRU) processes that can run opportunistically, or even manufacturing or processing loads that are schedule-flexible. Grid management will become in part a scheduling problem where generation availability informs industrial loads, rather than the reverse.
Lastly, demand response is definitionally different in the context of a lunar habitat. On Earth, where electricity rates fluctuate with the time of day, utilities encourage customers to move heavy usage such as charging their electric vehicles to the night when prices are low. This not only helps utilities manage load by smoothing out demand, it also helps customers save money. But there is no discretionary residential load in the traditional sense on the Moon where all load is directly life-critical or productive. Thus, demand response becomes about operational triage. Life-critical loads will be protected absolutely; operational loads will be placed on managed schedules; discretionary loads, if and when they exist, will be the only real flex.
What Grid Experience Actually Teaches
Modern grid experience is of great value. Operational discipline around load forecasting and dispatch will be essential for lunar power engineers. Contingency planning and redundancy thinking, especially without mutual aid, will be critical elements of system design. Interrogating the potential event sequences that could lead to cascading failures will likewise be crucial. Understanding degradation, maintenance cycles, and replacement logistics will be of first-order importance.
However, the things to unlearn are just as important. Assumptions that the modern grid is built on, but which are not transferable to the Moon, need to be dismantled. There is not always more power available somewhere. Interconnection cannot bail you out. On an isolated, remote grid, what you can generate on site is what you have available. No regulatory compact exists to sort out financing and governance issues. Evolving the grid in an austere, hostile environment is not easy or cheap. Initial decisions around geographic and economic constraints will have permanent effects.
Old knowledge will become useful again. The most essential will arguably be pre-interconnection and pre-regulatory compact. Financing models hundreds of years old will need to be dusted off and reimagined. Operators from islands such as Hawaii and other remote communities have uniquely applicable experience. Locating generation close to load, as with pre-REA rural electrification, will be more relevant than modern transmission and distribution. Military base power infrastructure, characterized by a single operator with life-safety priority and no external grid, has useful and applicable parallels.
The Moon will need people who understand how power grids work. But it will need them to carry that expertise carefully. Modern grid management’s most valuable contributions will be as much from negative knowledge as positive, in that they will require operators to understand precisely which assumptions they need to leave on Earth.
One day in the not-too-distant future, a grid engineer will be watching a terminal from some lunar habitat as the 14-day night approaches. Storage will be nearly full, power management routines having done their job. Triage protocols will be ready for any contingency. The engineer will know exactly what to do; the settlers in the habitat will be safe. That knowledge will have come not just from modern utility operations, but, more crucially, from something older, something more innate to what power actually is.