An engine with both high thrust and high exhaust speed fires in space

The Epstein drive in plain terms: an engine whose exhaust is fast enough and whose thrust is high enough that, scaled up, a ship could accelerate the whole way and reach the Belt in weeks rather than months. Nothing like it exists. Chemical rockets have the thrust but not the exhaust speed; electric thrusters have the speed but push like a breath. The node measures the engine physics (both at once, from fusion, for an hour), not a transit: a kilonewton at a hundred kilometres a second is fifty megawatts of exhaust power, which nothing has approached, and a ship that crosses in weeks needs that engine many times over. This is the node that separates the Slow Expanse from the Full Expanse, and it is the project's clearest long shot.

The test: what counts as it having happened

A propulsion system operating in space, whatever powers it, sustains, for at least one hour, an exhaust speed above one hundred kilometres per second (chemical rockets manage under five) together with a push of more than one kilonewton (roughly the weight of a hundred kilograms on Earth, which today's high-speed electric thrusters miss by a thousandfold), confirmed by the operator and by independent measurement of the vehicle's acceleration.

Settled by: The operator's test record and independent tracking of the vehicle's acceleration

The odds, by how long I live

baselineto 2071 ± 8	moderateto 2080	strongto 2095	radicalto 2136	openno deadline
38%	50%	75%	88%	91%

Each column is a different guess at how long I live, from current trends (baseline) to open-ended. The number is the chance this happens before that deadline, assuming everything it depends on happens first. As of September 25th, 2026.

Why these numbers: The criterion stopped requiring fusion on October 3rd, 2026 (my decision, on Claude's proposal); the numbers stand, because a power source this light is far likelier to be fusion than fission, so the clause was carrying almost none of the difficulty. Re-estimated September 25th, 2026 by Claude with the new dependency, I accepted October 2nd, 2026: the step now waits on a hundred-megawatt power source at one kilogram per kilowatt, which carries the weight gap this number used to price. Given that source, the remaining steps are a fusion burn in a flying reactor and a nozzle that turns it into thrust; a source that light is far likelier to be fusion than fission, so the fusion clause costs little here. The numbers are set so that the chance through the fusion-plant route alone stays where it was (0.08 times 0.38 is about the old 0.03 by 2071), and the fission route adds to it. Earlier reasoning, September 20th, 2026: re-estimated against the renamed node: the fusion-powered clause makes the criterion harder than it read (a non-fusion engine no longer counts) and dropping the transit promise makes it easier than the old name implied, and the first number was already priced for a fusion engine, so the two changes cancel and the numbers stand. A hundredfold weight gap and no thrust ever produced from fusion; a coin flip with no deadline is the founding rule at work. Proposed by Claude September 20th, 2026, I accepted.

What this one is worth

Nothing. The headline number does not rest on this step at any remove, so however it turns out, the chance that I live and work off Earth does not change. That is not a judgement that it does not matter. It is what the tree says: this step feeds a tier above the one the headline asks about, or sits beside the equation rather than in it.

What the ledger says about this step

The fast drive no longer has to be fusion-powered: any engine that meets the thrust and exhaust-speed test counts 2026-10-03 The tree itself changed
Forty-one steps had their criteria tightened so that two readers would agree when each has happened, with no number moved for that alone 2026-09-20 The tree itself changed

How it could happen, and where physics would have to give way

Long shots stay in the tree on purpose, because they make the goal vivid. The rule is that each one has to be written plausibly all the way to the edge of the impossible, naming how it would work and exactly where it would break.

The mechanism: A fusion reaction whose products are steered out the back directly, rather than used to boil water and spin a turbine, gives exhaust speeds of thousands of kilometres per second in principle. The path runs through fusion plants on the ground (proving the reaction can be sustained), then compact magnet-based reactors made cheap by superconducting wire, then a reactor light enough to fly, then a nozzle that turns its output into thrust. Pulsed designs that detonate small fuel pellets are the leading route on paper because they avoid holding a steady plasma.

The breaking point: The reactor has to be light as well as working. Ground fusion designs come to tens or hundreds of tonnes for each megawatt of electricity; paper drive concepts need about a kilowatt for each kilogram, a gap of a hundred- to a thousandfold that no built device has closed, and fission-electric systems top out near a fifth of a kilowatt per kilogram. No fusion device has ever produced thrust. Everything past a heavy fusion plant on the ground is engineering nobody has done.

The last 3 checks

Once a week the tree is read against the news, node by node, against the test above. Open a line to see what that week held. Every check ever made on this node is on the Radar page.

2026-11-01 Nothing new

Pulsar Fusion reported first plasma in its Sunbird exhaust test system in March 2026 (no fusion, no measured thrust), and nothing in the window shows a kilonewton thruster above 100 km/s.
2026-10-04 Nothing new

No in-space test reached 100 km/s exhaust speed with over one kilonewton; Pulsar Fusion's Sunbird first plasma (April 2026) is a ground test, and thrust measurement is only a stated next step.
2026-09-19 Nothing new

Pulsar Fusion's Sunbird first-plasma milestone, a ground test with thrust and exhaust-speed measurement still pending, occurred on March 25th, 2026, before this window; nothing new appeared in the current window.

Revisions

Date	Field	From	To	Why
2026-09-19	name	A drive that could cross the solar system in weeks is fired in space	An engine with both high thrust and high exhaust speed fires in space	Step 27 outside-model review, item 68 (docs/reviews/2026-09-19-rulings.md): the criterion is an engine-physics milestone and the old name promised a transit (a kilonewton on a thousand-tonne ship is 3.6 metres a second in the hour); renamed to what it measures, with the transit test not adopted. Decided September 19th, 2026.
2026-09-19	resolution	any propulsion system meeting the thrust and exhaust-speed thresholds	a fusion-powered propulsion system meeting the same thresholds	Step 27 outside-model review, item 70 (docs/reviews/2026-09-19-rulings.md): the criterion did not require fusion while the node depends on a fusion plant, so a non-fusion drive that met the thresholds would have been blocked by a dependency it did not need; the edge is kept and the criterion now matches it. Decided September 19th, 2026.
2026-09-20	probability	0.03 / 0.06 / 0.15 / 0.35 / 0.50	0.03 / 0.06 / 0.15 / 0.35 / 0.50 (re-estimated, unchanged)	Step 27 outside-model review, item 69 (docs/reviews/2026-09-19-rulings.md), held for re-estimation with the renamed node: the fusion-powered clause narrows the criterion and dropping the transit promise widens it, and the first number was already priced for a fusion engine; the re-estimate lands on the same numbers. Proposed September 20th, 2026, I accepted.
2026-09-25	depends_on	E-fusion-plant-delivers-net-electricity	E-compact-power-source-at-fast-drive-scale	Candidate 4 of docs/proposal-function-over-route-2026-09-25.md, I approved September 25th, 2026: what the drive takes from the energy factor is a compact power source of the right class, and a grid fusion plant is one route to it; a direct fusion drive need not follow a grid plant, and fission-electric at very high power is a separate path. The grid plant is now one of two routes under the function step.
2026-09-25	probability	0.03 / 0.06 / 0.15 / 0.35 / 0.50	0.38 / 0.50 / 0.75 / 0.88 / 0.91	Re-estimated by Claude with the new dependency, which I have not yet accepted or moved: the conditional now starts from a flight-weight hundred-megawatt source, so the weight gap priced into the old number moved to the function step; set so the fusion-plant route alone gives about the old overall chance and the fission route adds to it.
2026-10-03	resolution	'A fusion-powered propulsion system operating in space'	'A propulsion system operating in space, whatever powers it'	One of the questions the function-over-route pass left open (docs/runs/2026-10-02-estimates-ruled.md). 'Fusion-powered' was added on September 19th, 2026 (step 27, item 70) only so the criterion matched a link to the fusion plant. Since September 25th, 2026 the step waits on a light power source by any route, and the clause meant the fission route under that source could never satisfy this step. Proposed by Claude, decided October 3rd, 2026. The thresholds are unchanged and no probability moved.

Notes

Weight-gap figures checked September 8th, 2026 against a European Space Agency assessment of open magnetic fusion for propulsion and NASA fusion-rocket studies: conceptual fusion drives are projected at one to ten kilowatts per kilogram, fission-electric near 0.2, and the practical requirement is quoted as above one kilowatt per kilogram. Step 27, item 68 (September 19th, 2026): Astra's arithmetic, one kilonewton at a hundred kilometres a second is fifty megawatts of exhaust power, recorded in the description. Function-over-route pass, September 25th, 2026 (candidate 4): the criterion still says fusion-powered, as decided at step 27 item 70 to match the old fusion-plant link; with a fission-electric route now under the power step, whether the criterion should admit any power source is left for me.

Think this estimate is wrong? Here's how to tell me, and what happens next. The record itself is data/D-drive.toml in the public repository.