Given current evidence on candidate microgravity products and orbital compute workloads, what concrete, falsifiable unit-economics milestones over the next 5–10 years (e.g., $/kg to orbit, $/module for standardized factory hardware, $/kWh and $/rack-year in orbit) would show that Wright’s-law learning is actually steep enough for any orbital manufacturing or compute line to approach cost crossover with Earth options, rather than remaining permanently niche demos?

Answer

Indicative, falsifiable milestones that would signal at least one orbital manufacturing or compute line is plausibly on a crossover path (not just demos) in the next 5–10 years:

1. Launch & logistics

• L1: Recurrent LEO launch at ≤$300/kg for bulk cargo, ≤$600/kg for time‑critical return pods, sustained for ≥2 years.
• L2: End‑to‑end logistics (factory/compute module delivered to working orbit, serviced, and safely deorbited) at ≤$1,000/kg average over a 5‑year program for standardized payload classes.

2. Standardized orbital factory hardware (microgravity manufacturing) For a first‑wave product like ZBLAN fiber or a narrow semiconductor step:

• M1: Fully robotic, standardized production module (no routine crew time) with all‑in capex (flight hardware + test + integration) ≤$20M for ≥100 kg/year nameplate output.
• M2: Measured Wright’s‑law learning on identical modules of ≥15% unit‑capex or unit‑throughput‑cost reduction per cumulative volume doubling, over at least three doublings (e.g., from 1→8 modules or equivalent throughput).
• M3: Achieved effective production cost (excluding launch) ≤$5,000/kg in orbit for the target product, with demonstrated yield and quality matching the premium niche that pays ≥$20,000–$50,000/kg on Earth.

Concrete falsifier: If after deploying >8 essentially identical modules cumulative, per‑kg in‑orbit production cost (hardware amortization + ops) has not fallen below ~$10,000/kg despite launch at ≤$300/kg, Wright’s‑law learning is likely too shallow for that line to reach crossover.

3. Orbital compute racks Assuming niche, high‑margin workloads (rad‑hard testing, secure enclaves):

• C1: Standardized compute rack or module (including power, thermal, comms) delivered and installed in LEO for ≤$2M per 10 kW / 1 physical rack equivalent.
• C2: Verified energy cost in orbit ≤$0.5–1.0/kWh at rack input (including solar array amortization, storage if used, and orbital servicing), with at least 3× price/performance advantage vs rad‑equivalent Earth or ground‑based alternatives for that niche.
• C3: Effective cost per rack‑year (capex amortization + opex + rights/insurance, excluding workload‑specific software) ≤$300k/rack‑year for LEO, and a measured learning rate of ≥10% cost reduction per cumulative rack‑capacity doubling over three doublings.

Concrete falsifier: If, with launch ≤$300/kg, delivered rack‑year cost cannot be pushed below ~$500k/rack‑year after three doublings of standardized capacity, orbital compute is likely to remain a thin specialty testbed rather than approach crossover for any steady commercial workload.

4. Robotics and servicing as enabler

• R1: Standardized servicing/inspection robot class with all‑in capex ≤$15M and per‑intervention marginal cost (visit, repair, or swap) ≤$200k by the time ≥50 interventions have been logged.
• R2: Observed learning rate in per‑intervention cost of ≥15% per cumulative intervention doubling, sustained over at least three doublings (e.g., 10→80 interventions).

These are enabling milestones: if robotics cannot reach this regime, both factory and compute cost floors rise materially, pushing crossover further out.

5. Evidence that any line is near cost crossover A specific orbital line (e.g., fiber, a semiconductor step, or a secure compute enclave) looks plausibly near crossover if, by year ~10:

• Total delivered cost to customer (including launch, amortized hardware, ops, rights, insurance) is ≤2× the best terrestrial alternative for that niche, and
• Measured learning rates on at least two of: launch, modules, servicing, are at or above the thresholds above, so simple extrapolation over another 2–3 doublings brings cost to parity.

If none of the above thresholds are hit under a regime with launch ≤$300/kg for multiple years, it is strong evidence that current microgravity and orbital compute candidates are structurally niche under realistic Wright’s‑law slopes.