The year
Mostly Hardware That Is Flying or Being Built Now
The near term is extension, not invention.
The 2050 picture is a mature Earth-orbit industry, two lunar bases at the south pole, the first crewed landings on Mars and the start of a cislunar propellant economy. Almost all of it is hardware flying or in development in 2026: Starlink and the Starmind compute satellites, Starship, Tiangong and the commercial stations, Artemis and the International Lunar Research Station (ILRS).
What changes most is where computing happens. Orbital compute is the first customer large enough to absorb Starship's cadence - a terrestrial industry that ran out of grid, water, land and permits. The modeled migration rate, 15-20% a year, has to start in the 2030s and is still running in 2050.
Past the Moon, 2050 is science. Psyche, Europa Clipper, JUICE, Tianwen-4 and Dragonfly report; no outer-system outpost exists, because none is honestly crewed or supplied until megawatt-class nuclear electric propulsion flies.
Destination by destination
Every Destination in 2050
What exists, what is still building, and how firm each row is. Detail lives on each destination page.
| Destination | What exists | Building | Rated |
|---|---|---|---|
| Earth Orbit | Starlink; first Starmind tranches; Tiangong; commercial stations | Orbital compute at scale; Golden Dome layers | Building |
| Cislunar Space | Queqiao relays; navigation and relay at the libration points | EML-1 propellant depots | Projected |
| The Moon | Artemis and ILRS bases at the south pole; ILRS nuclear power station | Water plant; first mass-driver segment | Projected |
| Mars | First crewed landings; Tianwen-3 samples returned | Nuclear-powered outpost and propellant plant at Arcadia Planitia | Modeled |
| Venus | Science orbiters and probes | No industrial presence | Observed |
| Mercury | BepiColombo science | Solar-pumped laser research on the ground | Modeled |
| The Asteroid Belt | Psyche and Tianwen-2 science | Prospecting; nuclear electric cargo design | Modeled |
| Jovian System | Europa Clipper, JUICE, Tianwen-4 | Callisto surveys | Building |
| Saturnian System | Dragonfly flying on Titan | Nothing industrial | Building |
| Neptune System | No mission arrived | Flagship studies | Projected |
| Pluto & the Edge | New Horizons outbound | Nothing | Observed |
| Interstellar | Rubin census of interstellar objects | Sail and lens studies | Modeled |
The human
From Instrumented Astronaut to Adapted Workforce
The human body becomes part of the system design. Periods are capability transitions, not product dates.
| Period | Transition | What changes | Rated |
|---|---|---|---|
| 2026-2035 | Instrumented astronaut | Better dosimetry, AI medical support, early loading garments, dynamic circadian lighting; rodent torpor work | Building |
| 2035-2045 | Managed space worker | Multi-year lunar partial-gravity data; shared centrifuges; personal dose budgets set duty; one person supervises tens of machines | Modeled |
| 2045-2055 | Adapted workforce | Layered radiation countermeasures; Mars reproductive and developmental protocols; intermittent torpor on selected long transits | Modeled |
Gravity is treated as a dose - magnitude, duration and frequency - delivered by work garments and scheduled centrifuge time rather than a rotating city. Detail on habitats, suits, dose and settlements is on The Human Layer.
China
China in 2050
| Program | Position | Rated |
|---|---|---|
| Tiangong · CMSA | Expanded station, crewed continuously | Projected |
| Mengzhou, Lanyue, Long March 10 · CMSA | Crewed lunar flights after the first landing | Projected |
| ILRS · CNSA and Roscosmos | Basic station 2035; expanded 2045; nuclear power on site | Projected |
| Tianwen-3 · CNSA | Mars samples on Earth since 2031 | Building |
| Tianwen-4 · CNSA | Callisto orbiter science | Projected |
| Chinese compute constellations | State-built successors to the Three-Body constellation | Modeled |
What has to be true
What Has to Go Right First
Ship-to-ship propellant transfer, a launch-cost floor near $100/kg, lunar ice at a usable grade, and orbital compute that pays. Each has a test on the Experiments Board and a row on the Falsifiers page.
Contested
Deliberately Not Claimed
The load-bearing uncertainties, tagged with the confidence they actually earn.
Program dates will slip. The 2050 rows assume the announced sequence holds roughly; the record since January 2026 says it will not hold exactly.
Mars in 2050 is an outpost, not a settlement. Population stays below evacuation capacity; the crossing is a 2075 question.
Record
Provenance and Data Sources
- Confidence
- Observed for flown, fielded or measured - Building for hardware and programs in development - Projected for announced plans - Modeled for original analysis
- Status date
- Q4 2026
- Weakest link
- Any destination row past the first crewed landing