Reading the field
Capability Is More Concentrated Than Activity
The list of who is active in space is long. The list of who holds a load-bearing capability is short, and it is the second list that decides what gets built.
Two distinctions sort the field. A capability class is what an organization can actually do — put mass in orbit, run a deep-space link, keep a reactor alive on a surface. A program is the institutional effort that directs those capabilities toward a goal, such as NASA’s Artemis or the Chinese state effort. The two cross: Artemis buys heavy lift from SpaceX and deep-space communication from the NASA Deep Space Network, so a program is a route through the capability classes.
The space economy is narrow at the load-bearing points. SpaceX holds Western heavy lift, and Starship underlies nearly every near-term plan for Earth orbit, the Moon and Mars. Deep-space communication runs through a small number of antenna networks. Surface fission power exists in one NASA and Department of Energy program and nowhere commercial. China runs a parallel effort across every one of these classes, integrated under CASC and CALT rather than spread across competing firms, which is why it is treated as its own program rather than folded into the Western one. Building
Capability classes
Six Things a Player Can Hold
Each class gates a different part of what gets built. The leaders are named West and China, because no class has a single-nation answer.
| Class | What it gates | West · China | Rated |
|---|---|---|---|
| Heavy Lift | Mass to orbit, and the cost per tonne every destination is priced in | SpaceX Starship & Falcon; Blue Origin New Glenn — CALT Long March 9 & 10, LandSpace Zhuque-3 | Observed |
| Comms & PNT | Deep-space links and the timing and navigation signal everything else rides on | NASA Deep Space Network, GPS, Galileo — Chinese Deep Space Network, BeiDou | Observed |
| Orbital Compute | Data centers in orbit, and the power and cooling that terrestrial grids cannot give AI | SpaceX Starmind (AI1) — Chinese state constellations | Building |
| Commercial Stations | Crewed volume in low orbit as the ISS retires | Vast Haven, Starlab, Axiom — Tiangong (state, operational) | Building |
| ISRU | Making propellant, water and materials off Earth — the closure every settlement plan turns on | NASA demos, early commercial — Chang’e and Tianwen ISRU experiments | Projected |
| Space Nuclear | Surface fission power and nuclear propulsion — the capability gate at Mars and the outer system | NASA/DOE fission surface power, HALEU; SR-1 Freedom — Chinese reactor and RTG programs | Building |
Launch cost prices everything downstream, which is why heavy lift leads: a change in dollars per tonne to orbit moves every other class. No class has a single-nation answer, and none has more than a few credible holders.
The load-bearing few
Remove Four Organizations and Much of the Plan Stops
A short list of single points carries most of what is planned for the next decade.
Western heavy lift is SpaceX. Western deep-space communication is the NASA Deep Space Network. Surface fission power is one NASA and Department of Energy program. Orbital compute at scale is one company’s constellation, Starmind, and it has not flown. Each is a single point that a swath of planned activity leans on, and none has a fielded Western competitor of equal weight today. China is the exception: not four firms but one integrated state stack that spans every capability at once, so no single Chinese entity carries the same concentrated risk.
Programs
Two Stacks and One Cross-Cutting Constraint
Programs route through the classes toward a goal. Workforce is not a program — it is the labor limit that binds all of them.
| Program | What it is | Rated |
|---|---|---|
| Artemis | The US-led Moon-to-Mars program — SLS and Orion, the Starship and Blue Moon landers, Gateway, and the international and commercial partners under it. The route through Western heavy lift, comms and, eventually, space nuclear | Building |
| Chinese Program Structure | The state-integrated stack — CNSA policy, CASC and its CALT launch arm, the crewed program behind Tiangong and the Long March 10 lunar roadmap. Spans every capability class under one roof rather than a market of firms | Building |
| Workforce · cross-cutting | The engineers, technicians and trades every program competes for. A constraint on all of the above rather than a program of its own, kept here because it binds capability as hard as hardware does | Projected |
Contested
Deliberately Not Claimed
The load-bearing uncertainties, tagged with the confidence they actually earn.
The class list is a working cut, not a fixed scheme. Six classes — Heavy Lift, Comms & PNT, Orbital Compute, Commercial Stations, ISRU, Space Nuclear — cover what today’s plans for each destination lean on. Robotics, on-orbit servicing and launch insurance could each earn a class later; the cut will move as the field matures.
Concentration is a snapshot, and snapshots of this market move fast. Blue Origin New Glenn, the commercial stations and the Chinese reusable vehicles are all closing gaps that read as single points of dependence today. The finding is that capability is concentrated now, not that it stays that way.
Record
Provenance and Data Sources
- Confidence
- Observed for fielded capability — Building for programs and capabilities in development — Projected for capability that is announced or nascent
- Status date
- Q3 2026
- Heavy lift & recovery
- SpaceX, Blue Origin; CALT Long March 9 and 10, LandSpace Zhuque-3
- Comms, power & stations
- NASA Deep Space Network and GPS, ESA Galileo; NASA and DOE fission surface power; Vast, Axiom, Starlab; China’s BeiDou, Chinese Deep Space Network and Tiangong
- Programs
- NASA Artemis; the Chinese state stack — CNSA, CASC, CALT
- Weakest link
- Orbital compute and off-Earth ISRU are the least fielded and carry the most uncertainty; the concentration finding is a snapshot of a market that is moving