The data in words
9 experiments run in parallel; the cloud-life survey decides the gate on 15 December 2028. If Ouroboros or the Photonic Exhaust fails, the plan falls back to the multi-century route. Seed foundry launches: 1 October 2028. Gate: no native life: 15 December 2028. Every experiment must pass: 31 December 2028. Seed lands on Mercury: 14 January 2029.
| Experiment | Where | Runs | Proves | Pass | Kill signal |
|---|---|---|---|---|---|
| Ouroboros-0 (no substitute) | Earth | 1 January 2027 to 31 December 2028 | A factory that builds itself from simulated Mercury regolith | 95% mass closure, doubling in under 120 days | Under 70% closure |
| Exhaust-1 (no substitute) | Laboratory | 1 January 2027 to 31 December 2028 | Electroluminescent cooling at useful power | 10 kW/m² of net cooling, within 50% of the Carnot limit | No net cooling above 100 W/m² |
| Lightpipe | Low Earth orbit | 1 January 2027 to 31 December 2028 | Coherent power beaming | 1 MW over 1,000 km between satellites at 30% end to end | Phase lock fails beyond 100 emitters |
| Veil-1 | Orbit | 1 January 2027 to 31 December 2028 | An electrochromic slat | A 1 km² slat in orbit dims from 0 to 99% and holds its attitude | The film degrades within 6 months |
| Railhead | Earth | 1 January 2027 to 31 December 2028 | An ice mass driver | 1 t of ice at 3 km/s, 10 shots an hour | Under 1 km/s, or under 1 shot an hour |
| Cryo-seal | Earth | 1 January 2027 to 31 December 2028 | Dry-ice storage under a 285 K sky | Under 0.1 W/m² of heat leak for 12 months | The cap cracks under thermal cycling |
| Mini-Aegis | Sun-Earth L1 | 1 January 2027 to 31 December 2028 | A magnetic shield | A 10 m coil's bubble deflects the solar wind | Repeated quenches |
| Cloud-life survey (gate) | Venus | 1 June 2027 to 1 December 2028 | No native life will be harmed | Balloon and droplet sample return show no biology | A credible biosignature: the whole program stops |
| Choir-sim | Earth | 1 January 2027 to 31 December 2028 | Fleet autonomy | 10,000 robots run a factory for 90 days unattended | More than 1 human intervention a day |
| Experiment | Proves | Pass by the end of 2028 | Kill signal |
|---|---|---|---|
| Ouroboros-0 | A factory that builds itself from simulated Mercury regolith | 95% mass closure, doubling in under 120 days | Under 70% closure |
| Exhaust-1 | Electroluminescent cooling at useful power | 10 kW/m² of net cooling, within 50% of the Carnot limit | No net cooling above 100 W/m² |
| Lightpipe | Coherent power beaming | 1 MW over 1,000 km between satellites at 30% end to end | Phase lock fails beyond 100 emitters |
| Veil-1 | An electrochromic slat | A 1 km² slat in orbit dims from 0 to 99% and holds its attitude | The film degrades within 6 months |
| Railhead | An ice mass driver | 1 t of ice at 3 km/s, 10 shots an hour | Under 1 km/s, or under 1 shot an hour |
| Cryo-seal | Dry-ice storage under a {earth-like-temp} sky | Under 0.1 W/m² of heat leak for 12 months | The cap cracks under thermal cycling |
| Mini-Aegis | A magnetic shield | A 10 m coil's bubble deflects the solar wind | Repeated quenches |
| Cloud-life survey | No native life will be harmed | Balloon and droplet sample return show no biology | A credible biosignature: the whole program stops |
| Choir-sim | Fleet autonomy | 10,000 robots run a factory for 90 days unattended | More than 1 human intervention a day |
The pass criteria are not the production targets. They are the smallest result that shows the idea works and that scaling it up is engineering, not a new discovery. The kill signals mark the opposite: a result so far short that no amount of scaling will close the gap in time.
What each experiment proves
Ouroboros-0
A factory on Earth that builds itself from simulated Mercury regolith. It tests the one invention the whole plan depends on: a machine that makes every one of its own parts, chips included, from rock (Ouroboros Foundry). The pass bar is looser than the plan's doubling, because a first loop only has to show it can close; speed is engineering. The kill signal is about closure, not speed. On Mercury there is nothing to import, and a factory that needs a large share of its own mass sent from Earth does not grow exponentially at all. Closure cannot be engineered in later. Speed can.
Exhaust-1
Electroluminescent cooling at useful power, in the laboratory. This is the plan's heart: nothing else removes heat from Venus faster than a plain radiator at the cloud tops can shed it (Photonic Exhaust). The effect is real but has only been seen at tiny powers: of light from of electricity (Santhanam 2012). The kill signal sits at the level a plain radiator already reaches: a tile that cannot beat passive radiation is not a heat pump worth building. The pass bar is well above that, and names a distance from the Carnot limit, because the model charges the pump work at the Carnot minimum, joules per joule. A real tile will do worse. The budget has to spare on the Lattice alone and once the Ring powers the Railheads, so it can absorb a tile somewhat short of ideal, not one far short. The production tile must reach by 2031.
Lightpipe
Coherent power beaming between satellites. The Helios Lattice needs emitters phase-locked into virtual apertures wide, so a beam stays across over . Each aperture is millions of emitters acting as one. If phase lock fails with only a hundred or so, it will not hold with millions: beams spread, receivers balloon, and the beam-safety rules, which rely on precise pointing and instant defocus, cannot be met.
Veil-1
One electrochromic, light-bending slat in orbit (Hesperus Veil). The Veil must keep of the sunlight off Venus from to the end of the cooling, and then serve forever as the planet's thermostat. A film that degrades within months cannot do that, and a Veil that fails after cooling lets sunlight boil the CO₂ back out. So the kill signal is about lifetime, not performance.
Railhead
An ice mass driver on Earth. The Ice Train fires Bricks at every from each of its coilguns (Ceres Ice Train). The test asks for a fraction of that speed at a modest rate; the kill signal marks a speed and a rate so low that the scale-up would be a new invention. The test must also show that ice survives being thrown: a Brick feels at , carried by its sock.
Cryo-seal
Dry ice kept frozen under an Earth-warm sky. The Cryo-continents hold all of Venus's CO₂ under caps that must leak no more than about and last for decades. The kill signal is cracking under thermal cycling, because a cracked cap is how a sealed sea gets back into the air.
Mini-Aegis
A small superconducting coil at the Sun-Earth L1 point that inflates a magnetic bubble against the solar wind (Aegis Shield). The full coil is in radius and carries per strand. If a small coil cannot stay superconducting in deep space without repeated quenches, a large one will not either.
Choir-sim
A fleet of robots runs a factory unattended (Choir). The plan runs about a trillion machines across three worlds, with up to of light delay between Earth and Ceres. Humans cannot be in that loop. Any need for intervention that grows with the number of machines is fatal, so the kill signal is set at a rate a small fleet should never need.
The cloud-life gate
The cloud-life survey is not a technology test. It answers whether Phosphoros should exist. Venus's clouds, between and up, are the only part of Venus temperate enough for life as we know it to be imagined there (Atmosphere of Venus), and cooling the planet would destroy that habitat completely. There is no partial version of the plan that spares it: the Exhaust canopies float in the same layer.
The survey flies balloons through the cloud deck and returns droplet samples to look for biology, from to . The gate falls on , after the seed foundry has launched but before it lands on Mercury. A credible biosignature stops the whole program. The real missions flying in that window cannot settle the question in time (Precursor Science), which is why the plan flies its own.
A negative result is only as strong as the survey is sensitive. What "no biology" means, what counts as credible, and who judges it must be fixed and published before the survey flies, not after. Phosphoros proposes that an independent international panel holds that decision, so that the people who want the program cannot also be the ones who clear it. How sensitive such a survey can be in two years is an open question.
If a gate fails
A kill signal stops the line it tests. Most lines have a substitute that costs a margin, a date or a redesign, and the model can price each one: change the assumption in the Explorer and see what follows.
Two lines have no substitute within a decade. If Ouroboros or the Photonic Exhaust fails, the plan falls back to the multi-century route: a sunshade, a planet that cools by radiating on its own over to , and a slow build-up of industry in space, close to Birch's plan (Birch 1991) and to Phosphoros's own first version (About). Most of the other proof experiments still pay off on that route, because its machines are needed sooner or later.
If the cloud-life survey finds life, there is no fallback. Venus is left as it is.