Market, plan, and unit economics behind Stelaris, the shared swarm that moves anything in orbit. Grounded figures are sourced; forward figures are labeled as targets or scenarios.
● Confidential · behind login · for prospective investors
01 the market
A market that has to exist.sourced
As launch costs collapse and orbit fills, moving and servicing what’s already up there becomes essential infrastructure. Today it’s priced out: one bespoke vehicle per target. A reusable swarm converts latent demand into a real market.
Serviceable satellites vs. those that will actually be serviced through 2032: the gap we open.
$1.8T
space economy · 2035
Nearly 3× today (WEF / McKinsey).
~$8B/yr
on-orbit servicing · 2034
~12% CAGR, the closest existing market.
Sources: WEF & McKinsey, Space: The $1.8T Opportunity (2024) · NSR / Analysys Mason, In-Orbit Services · ESA Space Environment Report 2025 · FCC 5-year deorbit rule · Novaspace. Ranges are analyst-derived and directional. scenario Long-range top-down forecast from the internal memo (illustrative, not sourced line-by-line): TAM $2.1B (2025) → $12.8B (2030) → $47.3B (2035) → $284B (2045), ~28% CAGR.
02 the plan
Two raises. Prove it in software first.
The hardest, most differentiating risk (multi-robot coordination) is already de-risked in simulation, solo. So the near-term raise is small and software-only. Hardware capital comes later, gated on a working demonstration.
Stage 1 · Seed: now
$2–3M
~18 months · ~5 people · sim only
Take the proven coordination kernel to an end-to-end simulated dry-run of the product: a designed craft, an RL policy flying it in a 3D orbital environment, and an operator console tasking the swarm. No hardware built.
Stage 2 · Series A: later future
$50–75M
gated on the Stage-1 demo · first hardware
Satellite design, manufacturing, launch, and first orbital demonstrations. This is the large hardware raise described in the full investment memo, deliberately deferred until the software case is made.
scenario Full illustrative ladder from the memo: Seed $2–3M → Series A $50–75M → B $125M → C $250M → D $350M → IPO. Beyond Stage 1, these are planning figures, not commitments.
03 stage 1 · use of proceeds
$2–3M → a playable simulation of the whole system.
Roughly 18 months of runway for a five-person founding team, weighted to engineering, plus the in-house compute cluster and AIOS platform the team develops on. A spacecraft designer starts the hardware and supply-chain thinking in parallel.
Compute cluster, AIOS & simulationIn-house GPU cluster running local models via AIOS; 3D orbital sim + RL training
~20%
Spacecraft designerRobot design + early supply-chain, manufacturing & prototyping scoping
~16%
Prototyping, design contracts & legal/G&AEntity, IP, early hardware studies
~14%
Operations & supply-chain (part-time)Vendor scoping, program coordination
~8%
the milestone: a sim dry-run
00Compute cluster + AIOS. An in-house GPU cluster running local models through AIOS, our agentic development platform. The engine that builds the policies and the sim faster and cheaper, independent of external API providers.
01Craft design in sim. An initial SV-class servicer/worker robot, modeled with real geometry, actuation, and sensing.
02Policy in a 3D orbital environment. An RL policy trained on that vehicle’s dynamics under real orbital mechanics, beyond today’s abstract transport task.
03An operator console (C2). Sit down with several assets and several payloads in orbit and task the swarm to act on them.
→Together: a playable end-to-end simulation of exactly what the real system will do, the credible proof that de-risks the concept before hardware.
04 unit economics
Software leverage on high-value missions.targets
The model: high-value mission services from a capital-light, reusable fleet. Figures below are model targets, to be rebuilt bottoms-up as hardware is designed.
These are forward-looking projections and targets, not results. Hardware has not been built; figures will be re-derived from a bottoms-up model as the Stage-1 design matures.