Why nothing leaves as mass
10.5 km/s, and nothing worth paying it with.
Every other destination on this site ends by asking what leaves. The Moon answers with mass through a driver. The belt answers with platinum-group metals and, more honestly, with propellant. The outer-system pages answer with almost nothing. Mercury is the strongest confirmation of that pattern anywhere in the architecture.
Surface to an Earth transfer costs 10.5 kilometers per second, of which only 4.25 is escape and the rest is climbing outward against the Sun. The crust is iron-poor, has no platinum-group story anyone has proposed, and holds nothing a near-Earth asteroid does not supply at a tenth the velocity. Even the polar water is worth nothing as an export, because every customer who wants water is further out, which means shipping it uphill twice.
Light leaves because photons pay no velocity cost. The question is entirely about collimation.
The aperture product
One relation sets the whole trade.
For a transmitting aperture Dₜ at wavelength λ, the diffraction-limited spot at range L is 2.44 λL/Dₜ. Setting that equal to a receiver of diameter Dₕ gives
Dₜ × Dₕ = 2.44 λL
The product of the two apertures is fixed by wavelength and range, and everything else is a trade. The important property is that it is a product: transmitter can be bought with receiver, and the reverse. That is what keeps the transmitting structure modest even at outer-system range.
The reference case throughout this site is 100 gigawatts delivered at 1.06 micrometers, declared once and used everywhere. The wavelength is not arbitrary: neodymium-doped yttrium aluminum garnet, the gain medium in every serious solar-pumped laser built to date, lases at 1.064 micrometers. The same reference beam is used by the interstellar architecture, which shares this collector field and builds its own beam line.
The receiver is sized first
By survivable flux, not by diffraction.
The naive move is to fix the receiver at something plausible and solve for the transmitter, and it gives the wrong architecture. A beam of 100 gigawatts into a one-kilometer receiver is 127 kilowatts per square meter, 14 times the flux at Mercury’s own orbit. Nothing survives that as a receiving surface without becoming a rocket engine.
So the receiver diameter is set by the flux the receiver can survive, and only then does diffraction set the transmitter. Cap the receiver at 10 kilowatts per square meter — roughly Mercury’s own orbital flux, which the architecture has already established is manageable — and 100 gigawatts needs 10 square kilometers of receiving area, or 3,568 meters across.
This is the single most consequential design inversion on the page. Receiver flux tolerance, not diffraction, sets the architecture.
What it takes to reach
Transmitter diameter by destination.
| Destination | Range | Transmit aperture | Note |
|---|---|---|---|
| Earth / cislunar | 0.6–1.4 AU | 108 m | Competes with free sunlight; worth nothing |
| Mars | 0.6–2.1 AU | ~170 m | Will refuse on sovereignty grounds |
| Ceres | 2.2–3.3 AU | ~290 m | Serves one rock and strands the rest |
| Callisto | 4.2–6.2 AU | 564 m | The highest-value customer in the solar system |
| Titan | 8.5–10.5 AU | ~960 m | Largest displacement, least to do with it |
| Triton | 29–31 AU | ~3,000 m | Same, further |
| Pluto (46.3 AU in 2075) | 46.3 AU | 5,020 m | The design case for a 5 km array |
A 5-kilometer phased array inside Mercury’s orbit delivers 100 gigawatts to a 3.5-kilometer receiver at Pluto. The same array reaches Callisto with an aperture smaller than a large building. Neither structure is exotic. Both are thin-film-and-truss objects of the kind the industrial base is already building for other reasons.
Diffraction is not the hard part. Two other things are.
Pointing is a metrology problem
50 nanometers across a 564-meter baseline.
Holding the spot on a receiver at Jupiter takes about 95 microarcseconds of accuracy — one to two orders of magnitude finer than Hubble or the James Webb Space Telescope. Stated that way it sounds impossible.
But a phased array does not point mechanically. It steers by controlling the relative phase of its elements, so the requirement converts into knowing and holding element positions across a 564-meter baseline to roughly 50 nanometers — one part in 10¹⁰ of the baseline. That is a class of problem gravitational-wave interferometry has already solved at more demanding relative precision, with LIGO holding four-kilometer arms to a fraction of a proton width.
What has never been done is that metrology on a structure absorbing kilowatts per square meter with a 2.3-fold forced thermal swing across an 88-day orbit. The thermal-mechanical case, not the phase control, is the honest gap.
Everything is feed-forward
No servo loop can close at interplanetary range.
Light takes 43 minutes to reach Jupiter, during which the receiver moves 33,900 kilometers — 10,000 times its own diameter. The beam must be aimed where the receiver will be, which needs its state vector known to about a kilometer and 0.4 meters per second, continuously. That is routine for a tracked station and impossible to close in a servo loop, because the round trip is 86 minutes.
Retrodirective control — the pilot-signal technique in which the target tells the transmitter where it is, and which is what makes short-range power beaming safe on Earth — is unavailable at interplanetary range.
A receiver in a known orbit can be served indefinitely. A receiver that maneuvers unpredictably cannot be served at all. More of Mercury’s architecture follows from that one sentence than from any other: it is why the power contract goes to a fixed installation, why the departure corridor works only for vehicles on filed profiles, and why the weapon cannot hit anything that dodges.
Where the art stands
800 watts over 8.6 kilometers, and a megawatt in orbit by 2030.
Power beaming exists and is nowhere near this scale. The current record is DARPA’s Persistent Optical Wireless Energy Relay demonstration in 2025, which delivered more than 800 watts over 8.6 kilometers at about 20 percent end-to-end efficiency, moving roughly a megajoule across the campaign. The Mercury system is 100 million times the power over 10 billion times the range.
The largest program pointed anywhere near this is Chinese. The China Academy of Space Technology leads a four-phase space-based solar power roadmap: a 10-kilowatt low-Earth-orbit test in 2028 carrying both a microwave transmitting antenna and a low-power laser transmission payload and transmitting across 400 kilometers; up to one megawatt from geostationary orbit around 2030, over 35,800 kilometers, requiring on-orbit assembly; then 10 megawatts by 2035 and 2 gigawatts by 2050. Contributors include the State Key Laboratory of Space Microwave Technology, the Qian Xuesen Space Technology Laboratory and Chongqing University, with test facilities under construction in Chongqing, and the China Academy of Launch Vehicle Technology has proposed a reusable Long March 9 to launch it. Earlier work included a 300-meter microwave transmission from an airship-mounted payload.
That program is microwave-first and Earth-directed, so it is not the Mercury architecture. It is the only national program building institutional competence at power beaming at scale, and the 2050 rung at 2 gigawatts is within four orders of magnitude of the reference beam rather than eight.