Imagine a city ordering another hour of sunlight. Imagine a lunar power plant selling electricity across space. Now stop imagining that all these ideas have the same technical status. Some are demonstrated capabilities; others are credible engineering proposals; the grandest remain speculative systems assembled from much smaller successes. A September 21, 2026 research check found a genuinely fresh contribution: Hunter Williams’s September 15 paper on lunar electrical grids. It discusses infrastructure for sustaining activity on the Moon, not an operating lunar utility exporting power to Earth. No primary-source confirmation of a functioning lunar-to-Earth power station or continuous city-scale orbital illumination was found. That distinction is the foundation of this article, not a footnote. First, the mirror behind Earth. Yes, spacecraft can remain broadly on Earth’s nightward side while orbiting the Sun with us. NASA (National Aeronautics and Space Administration) describes halo orbits around the second Sun-Earth Lagrange point, called L2, roughly 1.5 million kilometers beyond Earth. A suitably chosen orbit can keep a spacecraft out of Earth’s shadow, although maintaining it requires corrections. But this is not a parking space directly above your house. Earth rotates beneath it, a particular city moves out of view, and a mirror needs both incoming sunlight and an unobstructed path to its target. A useful year-round lighting service would need carefully selected orbits, multiple spacecraft or relays, and allowances for clouds. One stationary cosmic bathroom mirror does not abolish night. Could you aim the reflected sunlight at a city? Yes, but think broad illumination, not a perfectly bounded spotlight. A June 2022 engineering paper explains that the Sun’s apparent width sets the approximate footprint: distance multiplied by about 0.01 radians. Applying that relationship gives a patch roughly six kilometers across from 600 kilometers away, before accounting for viewing geometry and optical imperfections. At L2 distances, the corresponding scale is roughly 15,000 kilometers. Those are illustrative calculations, not mission performance predictions. Moving the mirror farther away makes persistent positioning attractive but local illumination much harder. More importantly, a small mirror does not deliver daylight merely because its reflection is visible. Daylight-level illumination across a large patch requires an enormous collecting area or many overlapping reflectors. Steering photons is possible; multiplying their energy is not. The commercial ambition is real, but its scale deserves a cold shower. Reflect Orbital’s July 17, 2026 description of its proposed demonstration specified approximately full-moon-level light for up to five minutes, within two hours after local sunset. That is not perpetual noon. Its website checked for this article describes a staged expansion toward larger lighting and energy services, but a roadmap is not proof of deployment. The useful near-term question is whether a controlled pass can deliver measurable value without unacceptable spillover, rather than whether humanity has already acquired a second Sun. No primary-source confirmation of a successful orbital demonstration was found in the September 14-21 freshness search. Now replace the mirror with solar cells and a transmitter. This is a different machine: sunlight becomes electricity, electricity drives a microwave or laser beam, and a receiver converts the incoming energy back into electricity. The California Institute of Technology, known as Caltech, reported in June 2023 that its MAPLE (Microwave Array for Power-transfer Low-orbit Experiment) had demonstrated wireless power transfer in space and sent a detectable signal to a receiver in Pasadena. Detectable is the crucial word. It did not run a neighborhood. Separately, DARPA (Defense Advanced Research Projects Agency) reported on May 16, 2025 that a terrestrial laser experiment delivered more than 800 watts across 8.6 kilometers for 30 seconds. Those are substantial demonstrations of ingredients, not evidence of an orbital electricity industry. Could solar panels on the Moon charge storage and fire energy pulses back to Earth? In principle, yes: combine solar generation, storage, a transmitter and a receiving station. But pulsing changes when energy arrives, not how much energy exists. A receiver would have to handle the peak power and usually smooth delivery for customers. The Moon is also not permanently sunlit. Its roughly 29.5-Earth-day solar cycle means many locations face about two weeks of darkness, while its average distance from Earth is approximately 384,400 kilometers. These facts make storage, site selection and transmission geometry central design problems. Williams’s September 2026 paper considers solar arrays, storage, distribution and fault management for lunar users. My engineering inference is that supplying nearby lunar equipment is a more plausible early application than constructing a giant export industry before that local infrastructure exists. The beam itself is another brutal trade-off. Longer wavelengths generally require larger transmitting apertures to achieve a narrow beam; short-wavelength lasers can be more tightly directed but demand precise pointing and face cloud interference. NASA’s 2024 small-spacecraft technology assessment describes those underlying optical constraints. Applying them to power delivery adds demanding conversion, cooling and exposure-control requirements: a communications link need only convey information, while a power link must deliver useful energy safely. Any serious proposal needs a complete accounting from collected sunlight to usable electricity, not an impressive transmitter number. NASA’s January 2024 solar-power assessment, which examined possible systems operating in 2050, found its studied architectures more expensive than terrestrial alternatives under its assumptions. That was a dated scenario analysis, not a current quotation or a verdict that physics forbids the idea. Beyond orbital sunlight, here is the genuinely wild menu, with the readiness labels left attached:

  • Change an asteroid’s trajectory, demonstrated. The September 2022 DART (Double Asteroid Redirection Test) impact altered the orbit of Dimorphos around its companion. NASA’s March 2023 analysis reported a roughly 33-minute change in its orbital period. That demonstrates deliberate celestial intervention, not an ability to tow arbitrary asteroids wherever we please.
  • Manufacture oxygen on another planet, demonstrated at experimental scale. NASA reported in September 2023 that MOXIE (Mars Oxygen In-Situ Resource Utilization Experiment) had produced 122 grams of oxygen from the Martian atmosphere. Scaling that into dependable supplies for crews or propulsion is a major engineering project. Turning an entire planet breathable is a vastly different proposition.
  • Choose the apparent gravity inside a habitat, physically established principle, unbuilt at settlement scale. Rotation can press occupants against an outer floor. NASA’s 2021 study of deployable kilometer-scale structures explored architectures relevant to artificial gravity, while identifying deployment and validation work still required. A rotating habitat is not antigravity, and a promising structure is not a completed space city.
  • Edit cells to change the course of inherited disease, demonstrated medical technology. On December 8, 2023, the FDA (Food and Drug Administration) approved Casgevy, which modifies a patient’s blood stem cells using genome editing. This is a specific treatment involving substantial medical intervention, not evidence that arbitrary enhancements or immortality are available.
  • Put a dimmer between Earth and the Sun, a published megastructure proposal. Roger Angel’s November 2006 paper examined a cloud of spacecraft near the first Sun-Earth Lagrange point, called L1, intended to reduce incoming sunlight. The paper establishes a studied concept, not a deployed technology or proof that planetary consequences could be managed safely.

The mirror idea also contains a problem that a sales map cannot erase: illumination spreads beyond the intended customer. An August 2026 preprint by Miroslav Kocifaj, Gáspár Bakos and František Kundracik modeled atmospheric scattering from proposed orbital mirrors. Under the scenarios examined, it found significant alteration of the nighttime environment out to roughly 30 kilometers. This is modeling, not measurement of an operating constellation, and the preprint should not be treated as settled field evidence. Nevertheless, it directly challenges the comforting assumption that a sharply targeted service has sharply bounded consequences. Reflect Orbital argues that steerability and limited operating periods can control its service. Those claims and the researchers’ concerns need independent tests against the same measurable standards. Here is an original synthesis worth testing: combine electromagnetic beam steering with the fault-tolerant microgrid architecture discussed in the lunar-grid paper. Call it a Dark-Sky Energy Network. This is a proposed system concept, not a newly discovered physical theory, and I am not claiming nobody has considered related architectures. Instead of illuminating whole cities, orbital collectors would direct energy toward dedicated receiving sites connected to local storage and controllable electricity networks. The speculative advantage is that customers could receive power without purchasing visible light across the surrounding landscape. Caltech’s beam-steering experiment and the September 2026 lunar-grid research support the ingredients, not the performance or economics of this combined design. The thought experiment gets more interesting when you ask what would falsify it. Can a pilot deliver useful electricity after every conversion loss? Can it shut down reliably when pointing, tracking or receiver conditions fail? Can storage carry the load through interruptions? Does the complete system outperform a comparable terrestrial supply-and-storage alternative for the particular customer? Those are proposed tests, not demonstrated achievements. Passing them could make a limited wireless power service worth building; failing them would leave an expensive spectacle. The leap toward a more powerful civilization is not declaring that everything is possible. It is learning which extraordinary things remain valuable after the engineering audit.