If you talk to enough founders, investors, and program managers across commercial spaceflight, one constraint comes up again and again, regardless of mission type. It's not launch cost. It's not regulation. It's not even spectrum.
It's power.
[HIGHLIGHT]Across orbital data centers, mass-manufactured satellite constellations, lunar surface infrastructure, and deep-space exploration, the limiting factor is increasingly the same: how many watts you can generate, store, and dissipate per kilogram of spacecraft.[/HIGHLIGHT]
This is what Spacecadet Ventures general partner Wiz Khuzai meant when he framed Arinna's $4 million seed round as the "unlock" for the next generation of space companies. Power has become the gating constraint that determines what is and isn't economically possible above the Kármán line.
Where the Constraint Bites
The pressure on space power is showing up across at least four distinct domains, each pulling the supply chain in a different direction.
1. Orbital Data Centers
The most extreme power demands in commercial spaceflight today come from a category that barely existed three years ago: orbital data centers. A growing cohort of startups and incumbents is pitching megawatt-class compute clusters in low Earth orbit, with longer-term roadmaps pointing toward gigawatt-scale on-orbit compute infrastructure.
A meaningful orbital data center is a multi-megawatt machine. Reaching that scale with conventional space solar would require massive, rigid arrays whose mass alone makes the launch economics painful. Bringing the watts-per-kilogram number down — and doing it with cells that survive a decade or more in radiation — is non-negotiable for the orbital compute thesis to work.
2. Mass-Manufactured Mega-Constellations
Starlink, Project Kuiper, the Space Development Agency's transport and tracking layers, and a long tail of imaging and communications constellations are all pushing satellite production rates that the legacy supply chain was never designed for. Today's mega-constellations have largely shifted from gallium arsenide to silicon for cost reasons — but silicon's faster degradation under cosmic radiation has real implications for satellite design life and replenishment economics.
[STATS]
Conventional GaAs Cell | Expensive, rigid, supply-constrained
Silicon Cell | Cheap, but radiation-degraded
Arinna TMD Cell (target) | Flexible, durable, roll-to-roll
[/STATS]
For a constellation operator, every percentage point of efficiency loss is a kilogram of extra solar mass per satellite, multiplied across thousands of vehicles. Power technology is not a side feature; it is one of the dominant variables in mega-constellation unit economics.
3. Lunar and Cislunar Infrastructure
The next phase of NASA's Artemis program, combined with growing commercial lunar interest, is creating demand for surface power systems unlike anything in current production. The lunar surface is brutal: 14-day nights, dust contamination, extreme thermal swings, and radiation environments that vary dramatically between equatorial and polar deployments.
Surface power for the Moon — and eventually Mars — needs to combine very high specific power, mechanical flexibility (to be packaged and deployed by limited human crews or robotic systems), and durability over years of operation without maintenance. The current bespoke gallium arsenide supply chain is simply not designed for this kind of deployment volume or operational profile.
4. Electric Propulsion and Power-Hungry Payloads
Even within the world of conventional satellites, the power envelope keeps expanding. Electric propulsion — used by Starlink, many GEO communications satellites, and a growing list of national-security platforms — is fundamentally a power-conversion device: the more watts you can deliver to the thruster, the more thrust you can produce.
The same is true of next-generation radar, high-throughput communications, and on-board AI inference. Every new capability tends to be a power-consuming capability, and each one nudges spacecraft designers toward bigger, heavier, more expensive arrays.
Why Solar Is Still the Answer
The natural question is whether the answer to a space power crunch is something other than solar — nuclear, perhaps. NASA and DOE have steadily advanced fission surface power and radioisotope concepts that point toward a serious nuclear future for deep space and surface infrastructure. But for almost every commercial application in Earth orbit, solar remains the only viable architecture for the foreseeable future:
- **Reliability and safety.** Solar arrays do not require fissile material, regulatory approval cycles, or specialized handling.
- **Cost.** Photovoltaics are a mass-manufacturable commodity in a way that nuclear power systems are not.
- **Operational flexibility.** Sunlight is abundant in low Earth orbit, GEO, and most cislunar trajectories — the question is how efficiently you can harvest it.
What space needs is not a different power source. It needs a fundamentally better solar cell.
The New Wave of Space-Power Startups
Arinna is the most prominent recent example, but it isn't alone. A small but growing cohort of startups is attacking different parts of the space-power stack:
- **Photovoltaics.** Arinna's TMD-based cells are the most ambitious materials-science play, but other groups are working on improved thin-film, perovskite, and inverted metamorphic concepts.
- **Storage.** Several startups are working on space-rated batteries with higher energy density and better cycle life, particularly for missions with long eclipse periods.
- **Power management and conversion.** As payloads get more power-hungry, the electronics that route, convert, and protect spacecraft power are becoming a differentiator in their own right.
- **Wireless power transfer.** A handful of programs are exploring beamed power for lunar surface and orbital applications — still early, but increasingly attracting serious investment.
What to Watch in the Next 18 Months
For founders, investors, and operators tracking space power, three milestones will tell us whether the bottleneck is starting to break:
1. **First on-orbit TMD photovoltaic flights** — Arinna's planned 2026 qualification campaign will be the first real test of whether 2D-semiconductor cells survive the space environment.
2. **Orbital data center power architectures** — As Starcloud, SpaceX, and others reveal more of their orbital compute roadmaps, the choice of power technology will reveal which approaches are considered investment-grade.
3. **Lunar surface power awards** — NASA's evolving lunar program, including any contracts that flow from the new Moon base plan, will indicate whether incumbents or new entrants are positioned to supply the next generation of surface power.
For now, the bottleneck is real, the demand is growing, and the supply of fundamentally new solutions is still catching up. The companies that solve space power in the next five years won't just enable a few exciting payloads — they'll set the upper bound on what the entire space economy can do.