KSTTAI / Autonomous Industrial Ecosystems
Autonomous seeds that build the factories that build the future.
Tai Probes designs and operates self-replicating industrial ecosystems — autonomous systems that land on a body with resources, mine and refine locally, manufacture their own tooling and power infrastructure, and use the surplus to build another one. Each generation is a measurable engineering milestone, not a metaphor: the program is structured as five independently fundable MVPs on a path toward Kardashev Type II industrial capacity.
What a Tai Probe is
One seed, three functions, one closed loop.
A probe is not a single machine — it's an integrated energy, compute and manufacturing system that has to close its own material loop before it can grow. Every later capability (organizations, settlements, interstellar launch) is this same loop, run longer and at greater scale.
Resource extraction
ISRU mining and refining of local regolith, metals and volatiles into usable feedstock — the step every later capability depends on and the most common point of program failure.
Closed-loop fabrication
Additive and subtractive manufacturing that turns refined feedstock into structure, power systems, robotics and — critically — the tools that make the next generation of tools.
Directed self-copy
Once local industrial mass exceeds the mass of the original seed, surplus capacity is committed to building a new, independently operating seed rather than expanding the first indefinitely.
Everything past this point — infrastructure construction, organizational autonomy, interstellar seeding — is this loop composed with itself at increasing scope. The version ladder below is how we track and fund that composition.
Capability roadmap
Every version is a measured, funded engineering milestone.
Each generation is defined by one breakthrough capability with its own KPIs, budget and deadline — not a marketing label. The ladder below is the full v0.1–v8.0 program as currently costed.
| Version | Major breakthrough | Duration | Deadline | Incremental | Cumulative |
|---|---|---|---|---|---|
| v0.1 | Autonomous Mission Planning | 2 yrs | Yr 2 | $10M | $10M |
| v0.2 | Digital Twin Platform | 2 yrs | Yr 4 | $40M | $50M |
| v0.3 | Autonomous Mission Operations | 2 yrs | Yr 6 | $100M | $150M |
| v0.5 | Closed-Loop Self-Optimization | 2 yrs | Yr 8 | $350M | $500M |
| v1.0 | Autonomous Space Factory — operates for years unsupervised: ISRU, mining, refining, manufacturing, fault recovery | 5 yrs | Yr 13 | $5B | $5.5B |
| v1.5 | Self-Replicating Factory — first genuine von Neumann milestone: the factory builds another factory | 5 yrs | Yr 18 | $15B | $20.5B |
| v2.0 | Autonomous Infrastructure Builder — power, compute, habitats, transport, comms, no humans on-site | 7 yrs | Yr 25 | $30B | $50.5B |
| v2.5 | Autonomous Organizations — the system runs engineering, manufacturing and governance functions, not just machines | 5 yrs | Yr 30 | $20B | $70.5B |
| v3.0 | Self-Sustaining Civilization Seed — permanent settlement independent of Earth resupply | 10 yrs | Yr 40 | $130B | $200.5B |
| v4.0 | Interstellar Seed Probe — the civilization manufactures and launches new seed probes | 10 yrs | Yr 50 | $300B | $500.5B |
| v5.0 | Multi-System Network — independently governed civilizations with shared scientific knowledge | 15 yrs* | Yr 65* | mission-dep. | — |
| v6.0 | Autonomous Scientist — forms hypotheses, runs experiments, publishes verified engineering improvements | parallel v3–v5 | Yr 55 | $100B | ≈$600B |
| v7.0 | First-Principles Civilization Engine — improves its own engineering, manufacturing and AI without waiting on predefined procedures | 15 yrs | Yr 70 | $400B | ≈$1T |
| v8.0 | Fully Autonomous Civilization Seed — arrives, characterizes, builds, sustains and re-seeds without further support from Earth | 10 yrs | Yr 80 | $1T | ≈$2T |
* From v5.0 onward, "deadline" is a weaker concept than "capability complete" — interstellar transit time, not engineering readiness, dominates the later schedule. All figures are order-of-magnitude planning estimates, not predictions.
De-risking program
Five MVPs, each independently fundable and verifiable.
Rather than one monolithic bet, the path to Kardashev Type II industrial capacity is staged as five minimum-viable proof points. Each one de-risks the next phase of technical progress and unlocks the next round of capital on its own merits.
Closed-Loop Ground + Near-Earth Demonstration
A fully integrated terrestrial (and optionally LEO/cislunar) test system proving closed material cycles and partial self-replication under controlled conditions — before committing to expensive lunar delivery.
First Partial-Replication Cycle on the Moon
Delivery and operation of industrial seeds that grow total industrial mass beyond their own delivered mass using predominantly lunar resources — the single most important de-risking event in the roadmap.
Multi-Body Network + First Productive Output
A functioning network across the Moon plus at least two near-Earth or main-belt asteroids, proving the system can expand beyond one body and generate real economic value beyond pure self-copying.
First Operational Outer-System Industrial Node
Delivery and activation of an industrial seed in the Kuiper Belt (or equivalent), proving the architecture holds under low insolation, long communication delays and volatile-rich environments.
Multi-GW Solar Factory Cluster
The transition point from "self-replicating industrial network" to "stellar-scale energy infrastructure builder" — the minimum proof that the system can climb the Kardashev scale, not just replicate indefinitely.
First-principles sizing
What a viable seed actually needs.
Order-of-magnitude minimums for a seed that can bootstrap itself, reach local self-sufficiency, and begin exponential replication in a favorable location (near-Earth asteroid or inner Main Belt). Requirements rise substantially in the Kuiper Belt or Oort Cloud.
Energy
| Parameter | Minimum viable | Robust |
|---|---|---|
| Continuous power | 100 kW – 1 MW | 5 – 50 MW |
| Energy storage | 1 – 10 MWh | 50 – 500 MWh |
| Target EROEI post-bootstrap | > 10:1 | > 30:1 |
Below roughly 5–10:1 EROEI, a seed cannot grow. Energy is the most common binding constraint the farther a seed operates from the Sun.
Computing
| Parameter | Minimum viable | Robust |
|---|---|---|
| Sustained compute | 10¹⁴ – 10¹⁵ FLOPS | 10¹⁶ – 10¹⁷ FLOPS |
| Memory / storage | 100 TB – 1 PB | 10 – 100 PB |
| Autonomy without contact | Days – weeks | Full local autonomy, months |
Under-provisioned compute is a common failure mode: seeds that can't close the decision loop locally can't adapt or recover from faults.
Infrastructure — mass & manufacturing
| Parameter | Minimum viable | Robust | Notes |
|---|---|---|---|
| Total seed mass | 50 – 300 t | 1,000 – 10,000 t | Structure, robots, power, compute, tools |
| Initial robotic workforce | 20 – 100 units | 300 – 2,000 units | Mining, transport, assembly, maintenance |
| Manufacturing capability | Own spare parts + simple structures | Complex parts + replication-ready | Critical-path item — must make the tools that make the tools |
| Mobility | Local (km–10s of km) | Regional (body-to-body) | Must reach new resource deposits |
Replication doubling time by location
| Location | Early-stage doubling | Mature-system doubling | Main limiting factors |
|---|---|---|---|
| Near-Earth asteroids | 8–18 months | 4–9 months | Best energy + metals, short transport, easy thermal control |
| Main asteroid belt | 12–24 months | 6–12 months | Good metals, moderate energy, higher delta-v |
| Kuiper Belt | 2–5 years | 12–24 months | Low solar energy, mostly ices, long distances |
| Oort Cloud | 5–15+ years | 2–6 years | Extremely low energy, sparse resources, cryogenic operations |
A frequently cited 6-month doubling time is realistic only in the inner system under highly optimized, mature conditions — it becomes progressively harder moving outward.
Growth that has to earn its own energy budget is the only growth that doesn't stop.
Every version on the ladder above exists to prove one thing at increasing scale: that the loop — acquire, manufacture, replicate — closes on its own, without a subsidy from Earth.
Simulation-first engineering
Simulation is the source of truth. Visualization just reads it.
Every seed generation is designed, tested and optimized in simulation before any hardware is committed. The simulation core (TaiSim) computes state — energy flows, mass and recycling, replication, KPIs. The 3D layer is telemetry: it reads TaiSim's state and renders it, never the other way around.
Level-of-detail hierarchy
Node / site
Sub-meter to meter scale. A single integrated seed site — power, storage, compute racks, heat exchangers, recycling loops, robots. Highest fidelity.
Orbital
Kilometer scale. Constellations, factories, beaming paths. Medium fidelity, instanced meshes.
Stellar
AU scale. Distributed solar collector arrays and energy-network topology. Aggregate flows, not individual objects.
Master KPI at every scale: the Closed-Loop Efficiency Index — recycled matter % × compute per joule × energy-capture % — shown as a single live gauge.
Program structure
Seven synchronized R&D programs, one systems-integration product.
Rather than one monolithic project, the ~$2T v0.1–v8.0 program is treated as seven large, independently valuable R&D efforts with synchronized milestones. The Tai Probe is the systems-integration product that combines their outputs — which means intermediate technologies deliver commercial value long before the full civilization-seeding system exists.
Illustrative allocation across the full v0.1–v8.0 program, order-of-magnitude. Tai Probes is the integration layer that draws on each program's output rather than duplicating it.
Where the technical risk concentrates
| Milestone | Main unknown |
|---|---|
| v1.5 | Reliable self-replication from local resources, unsupervised |
| v3.0 | Complete autonomous civilization operation without Earth support |
| v4.0 | Practical interstellar probe deployment at meaningful scale |
| v6.0 | AI capable of expert-level autonomous scientific discovery across domains |
| v8.0 | End-to-end verification of a fully autonomous civilization seed |
Company
Part of the KSTTAI ecosystem.
Tai Probes is a daughter company of KSTTAI Holding (ksttai.ir), the systems-integration layer that draws on the group's energy, manufacturing, computing and simulation programs. Founded and led by Matin Khabari, CEO across all KSTTAI companies.
Get in touch
Request a technical briefing.
Tell us which phase you're evaluating — Phase 0 ground demonstration, Phase 1 lunar seeding, or the program as a whole. We'll follow up to scope a technical walkthrough of the roadmap and MVP gates.