Fusion Spacecraft

A fusion spacecraft uses controlled fusion for propulsion — either as a power source driving electric thrusters, or by directing the fusion products themselves out of a magnetic nozzle as exhaust. The energy density argument is real and enormous: deuterium–helium-3 fusion releases millions of times more energy per kilogram of fuel than any chemical reaction and roughly an order of magnitude more than fission per unit mass.

The framing under test is a syllogism, and it is the most common thing said about this subject: fusion power is nearly here; a rocket is a power plant with a nozzle; therefore fusion rockets follow. Every step of it is wrong in an instructive way, and this brief takes them in order. Fusion power is not nearly here on the numbers its own laboratory publishes. A power plant is not a rocket with a nozzle, because the two optimise opposite quantities. And the fusion propulsion concepts that actually exist run on the harder branch of fusion, not on a spin-off of the easier one.

Where this brief stops. Commercial Fusion owns terrestrial fusion in full: the National Ignition Facility, ITER, SPARC, the private sector and its funding, tritium breeding and the tritium stock, wall-plug economics, cost per kilowatt-hour, grid integration, the gain ladder, and the question of whether fusion power arrives at all. The rule this page applies is simple and it is applied strictly: if a fact would still matter if the device never left the ground, it belongs to that brief. The National Ignition Facility appears here only to establish that the transfer to propulsion does not happen.

What is left for this brief is the part nobody else owns: specific power in kilowatts per kilogram and why no power programme measures it; direct conversion of fusion products to thrust and the physics of a magnetic nozzle; the distinction between fuel and reaction mass, which a rocket budgets separately and a power plant does not budget at all; neutron shielding as a vehicle penalty rather than a building penalty; the aneutronic branch that propulsion needs and power programmes largely do not pursue; and the small set of concepts that are propulsion-native. Frontier This brief also inherits an entire dependency. A fusion vehicle needs everything Advanced Nuclear Propulsion needs — radiators at scale, megawatt power conversion, launch approval for a nuclear system, ground-test infrastructure, appropriations that outlast a cancellation — and then adds a reaction nobody has run at net power in a flyable package.

Established Start with what has been measured, because it is genuinely impressive and it is not what the framing needs it to be. The National Ignition Facility achieved ignition on 5 December 2022 — 2.05 MJ of laser energy on target, 3.15 MJ of fusion yield, a target gain of about 1.5 — and has repeated it many times since. The published record runs 3.88 MJ on 30 July 2023 at gain 1.9; 2.4 MJ on 8 October 2023; 3.4 MJ on 30 October; about 5.2 MJ on 12 February 2024 at gain 2.4; 5.0 MJ on 23 February 2025 at gain 2.44; 8.6 MJ ± 0.45 on 7 April 2025 from 2.08 MJ, a target gain of 4.13, the eighth ignition event; and 7.9 MJ ± 0.4 on 20 June 2025, the eleventh. Established That is a real, replicated, monotonically improving experimental record and this brief does not diminish it.

Established Now the laboratory's own statement of what remains, which is the citation this page is built on. In coverage of its inertial fusion energy programme, LLNL experts state that a plant would need fuel capsules ignited “at a rate of about ten times per second”, and forecast that “a target gain approaching 15 would be required to operate a test facility with zero net electric demand”. Established Set that against the achieved 4.13 and the comparison is exact: a factor of about 3.6 in gain, and many orders of magnitude in repetition rate, to reach a facility that merely breaks even on its own electricity draw — not a power plant, a test facility. Frontier This is a far better statement of the gap than the usual argument about wall-plug efficiency, because it is a number the laboratory published about its own facility, and its interest runs against it. The wall-plug arithmetic itself belongs to Commercial Fusion, which carries it properly.

Established The decisive quantity for propulsion is not gain at all. It is specific power. The ESA Advanced Concepts Team's assessment of open magnetic fusion for space propulsion (Romanelli, Bruno and Regnoli, ARIADNA study 04/3102 under ESTEC contract 18853/05/NL/MV) states that specific power values in the range 1–10 kW/kg “are adequate for the exploration of the solar system”, and that this requirement “fundamentally drives the entire technical assessment”. It further finds that a fusion power density in excess of 1 MW/m³ is needed. Frontier No terrestrial fusion programme optimises for either quantity, and there is no reason it should. ITER, SPARC and the National Ignition Facility are trying to make electricity affordable. A power plant can be arbitrarily heavy and arbitrarily well shielded, and being heavy is one of the largest cost savings available to it. A rocket cannot buy anything with mass.

Established The second transfer failure is neutrons, and it is why the propulsion concepts diverge on fuel. Deuterium–tritium fusion — the branch every major power programme pursues, because it is the easiest reaction to ignite — releases about 80% of its energy as 14 MeV neutrons. A neutron carries no charge, cannot be steered by a magnetic nozzle, and must be stopped by shielding mass the vehicle carries. Frontier So every serious fusion propulsion concept goes aneutronic, to deuterium–helium-3 or proton–boron-11. Frontier And the ESA study is explicit about what that costs: for aneutronic reactions such as p-¹¹B and p-⁶Li, “the system cannot achieve a positive fusion gain except far from thermal equilibrium” — that is, with electron and ion temperatures deliberately decoupled. Propulsion's fuel of choice is the branch the power programmes are not solving, and it is harder.

Speculative The best-known propulsion-native concept is the Direct Fusion Drive, and its numbers are all design projections. It is built on the Princeton field-reversed configuration, invented by Samuel Cohen at the Princeton Plasma Physics Laboratory in 2002, using odd-parity rotating-magnetic-field heating. The claimed performance is 5–10 newtons of thrust per megawatt of fusion power, about 2 MW of electrical power available to the payload, specific impulse around 10,000 s on D-³He, from a unit roughly 2 metres in diameter and 10 metres long, with a claimed power split of 35% to thrust, 30% to electricity, 25% lost as heat and 10% recirculated. NIAC Phase I studies were published from 2012 and a Phase II simulation phase was entered by 2018. Established The essential caveat, which this brief will not bury: the Princeton field-reversed configuration has never produced fusion. The experiments are plasma heating and confinement experiments — by 2017, studies of electron heating with the odd-parity method had surpassed theoretical predictions, which is a real result about heating and not a result about fusing. Every thrust, specific-impulse, mass and power-split figure above comes from a machine that has not fused anything, and each is flagged speculative for that reason.

Speculative The most experimentally grounded concept in the NASA portfolio is Pulsed Fission-Fusion, and it is the one this brief treats at length — because its own report is unusually candid. PuFF, from Robert Adams at NASA Marshall with Jason Cassibry at the University of Alabama in Huntsville, David Bradley, Leo Fabisinski and Geoffrey Statham, took a NIAC Phase I award in July 2013 and Phase II in 2018. A Z-pinch compresses a deuterium–tritium plasma column; the fusion neutrons induce fission in a surrounding uranium or thorium liner; the fission energy boosts the fusion yield. The team calls it “a fortuitous cycle” in which neither process alone reaches the required energy. Phase I claims specific impulse of 6,500 s from nozzle modelling, a Mars sprint delivering 25 t to Mars in 37 days from 190 t in low Earth orbit, and an outer-system probe delivering 10 t to 1000 AU in 36 years at 0.075 AU per day from 150 t. Phase II claims 30,000 s with “thrust levels sufficient to travel to Mars in a month and to interstellar space in a few decades”.

Established And then the same report's own list of what it did not resolve, which is the most valuable set of sentences in this brief. The time-variant models were not completed — moving to a solid target “complicated our time-variant model, now requiring … phase change algorithms and stress-strain calculations”. The unresolved core problems are named as “reaching criticality and avoiding plasma instabilities”. Only 5% of available capacitor-bank energy is assumed to convert to thermal heating, with actual performance unknown. No breakeven has been demonstrated; steady-state models suggest it might be reachable with the Charger-1 facility, with no experimental confirmation in Phase I. Whether compression targets can achieve the required uranium shock compression “is not yet known”. And the results are explicitly from a “non-optimized” point design. Established That is a NASA study reporting the limits of its own concept, in its own funded report. Interest runs against it, which is what makes it the highest-weight source on this page after the ESA requirement analysis.

Frontier PuFF also settles the dependency question, and it does so by construction. The most concretely developed fusion propulsion concept in the NASA portfolio works by driving fission in a uranium liner. It therefore needs a fissile supply, a fuel form, a launch authorisation under the nuclear tier structure, and a ground-test regime — the entire constraint set that Advanced Nuclear Propulsion records. Established The practical path to a fusion drive currently runs through fission rather than past it, which is exactly why this brief depends on that one.

Frontier On configurations, the published assessment has a view and it is worth stating. The ESA study evaluated magnetic mirrors in tandem, gas-dynamic and field-reversed variants, field-reversed configurations proper, spheromaks and levitated dipoles, and concluded that open magnetic configurations are “particularly suited” to propulsion because they permit direct thrust rather than requiring a closed confinement volume and a separate conversion stage. Frontier The difficulty is achieving the required power density, and performance hinges on the achievable plasma beta — the ratio of plasma pressure to magnetic pressure. This is the one place where propulsion and power research genuinely part company on the machine as well as the fuel: the configurations propulsion wants are not the ones the tokamak programmes have spent fifty years developing.

Speculative One published mission study shows what a working drive would buy, and it is a useful comparator precisely because its premise is conditional. A 2025 feasibility study of a mission to Sedna compares a 1.6 MW D-³He Direct Fusion Drive against an advanced solar sail using thermal desorption of its coating. The fusion drive “could reach Sedna in approximately 10 years, with 1.5 years of thrusting” and — the part that matters — permits orbit insertion. The sail with a Jupiter assist could arrive in seven years but only as a flyby. Conventional propulsion needs up to thirty years. Sedna reaches perihelion in 2075–2076. Speculative The fusion leg of that comparison assumes a drive that does not exist, and the study is cited here as a legitimate published trade study rather than as evidence that the drive works.

Established One thing in this subject is settled and should be said before the open questions: fusion releases energy and can be ignited in a laboratory. Eleven ignition events through June 2025 at a single facility. Nobody serious disputes this and the brief does not treat it as contested.

Frontier Open question one belongs mostly to another brief: can inertial fusion reach a gain sufficient for net electricity? The laboratory's own figures put the requirement near 15 at about ten shots per second against 4.13 achieved at a rate of roughly one shot per campaign. Frontier The reason this brief carries it at all is to state the falsification condition for its own thesis: if a gain-15 shot happened tomorrow it would be a landmark for Commercial Fusion and would not move this page. A fusion propulsion timeline moves only if someone demonstrates specific power, not gain.

Frontier Open question two is the real one: is 1–10 kW/kg reachable by any fusion device? This is a published requirement from an agency-commissioned assessment, not a measurement, and no fusion machine has ever been characterised in those units because no fusion machine has ever needed to be. Speculative The honest position is that nobody knows, because the question has not been asked of hardware.

Speculative Open question three: can a field-reversed configuration burn deuterium–helium-3 and produce direct thrust? Held by Cohen at Princeton and by Princeton Satellite Systems, which is a commercial vendor and an interested party on its own drive. The configuration exists, the heating method works better than predicted, and no fusion has occurred. Speculative Open question four, downstream of it: does the Direct Fusion Drive give roughly 10,000 s at 5–10 N/MW? A vendor projection from a device that has not fused. Both are carried here because the directive is to name hypotheses rather than omit them, and both are flagged at the weakest level the evidence permits.

Speculative Open question five: can aneutronic fusion reach positive gain at all? The ESA assessment says not at thermal equilibrium, which pushes the whole aneutronic branch into non-equilibrium plasma regimes that are themselves an open research problem. Frontier This is the deepest structural difficulty in the subject and it is almost never mentioned in popular treatments, which tend to present aneutronic fuel as a design choice rather than as a harder physics problem.

Speculative Open question six: does Z-pinch fission-fusion reach 6,500 to 30,000 s? Adams and colleagues say it might; their own Phase I report says criticality and plasma instabilities are unresolved and no breakeven has been demonstrated. Frontier What would settle it is a Charger-1 campaign showing the predicted uranium shock compression, and the report says whether that compression is achievable “is not yet known”.

Handwave Open question seven is not really open, and naming it is the point of this page: is fusion propulsion a spin-off of fusion power? No named technical holder advances this in the literature; it is a popular inference, and it is contradicted by the figure-of-merit argument, by the fuel divergence, and by the fact that the propulsion-native concepts import fission. Established Fusion propulsion is not downstream of fusion power. It is a divergent branch that cannot inherit the power programmes' central result.

Speculative Open question eight: will commercial fusion investment pull propulsion along? This is implicit in a great deal of fusion-startup enthusiasm and this brief found no evidence for it — no fetched record of a fusion power company running a propulsion programme. Speculative The mechanism by which it might happen is plausible in principle: magnet technology, plasma control and power electronics developed for a plant would transfer. The mechanism by which it would not is that a plant developer has no reason to spend money reducing mass, which is the only figure of merit that matters here.

Frontier Bottleneck one is specific power, and it is the bottleneck the whole brief is organised around. The requirement is 1–10 kW/kg and a fusion power density above 1 MW/m³. Nothing has been measured against either. Speculative A reactor that produces net energy at any mass is a triumph for the grid and useless for a rocket, and the gap between those two outcomes is not addressed by any funded programme anywhere.

Established Bottleneck two is heat rejection, inherited and made worse. The nuclear electric analysis gives the shape: radiators of roughly 1,500–3,000 square metres single-sided for one to two megawatts electric at around 500 K, and for one earlier reactor programme a heat-rejection subsystem projected at 10.1 kg/kWe — about half the entire mass budget. Frontier A fusion drive at tens of megawatts inherits a worse version of the same problem, and direct thrust from a magnetic nozzle only helps for the fraction of energy that leaves as charged particles.

Established Bottleneck three is megawatt power management, and it is stated flatly by an independent committee: radiation-hardened power electronics for power management and distribution at megawatt electric levels “have never been developed”. Frontier A fusion vehicle that converts thermal or direct-conversion output to electricity for thrusters needs exactly this hardware, at higher power than the fission case, in a harsher radiation environment.

Established Bottleneck four is test infrastructure and it is shared with the fission brief. Ground testing of electric propulsion is limited today to under 50 kWe. There is no facility anywhere for testing a multi-megawatt space propulsion system as an integrated article, and a fusion drive would need one plus tritium handling, neutron activation control and a vacuum chamber that can absorb the exhaust.

Speculative Bottleneck five is the magnetic nozzle, which is where propulsion physics and fusion physics actually meet and where the least work has been done. Direct thrust requires detaching a magnetised plasma from the field that confined it, efficiently, at high power density, without eroding the nozzle. Frontier The ESA assessment's preference for open configurations rests on this being tractable; the achievable plasma beta is the parameter it hinges on and it is not settled.

Frontier Bottleneck six is fuel supply, and it splits by concept. Deuterium is abundant. Tritium is not, and the world civil stock and its breeding problem belong to Commercial Fusion. Helium-3 is the one this brief owns, and it is worse: the terrestrial supply is a byproduct of tritium decay in weapons stockpiles, and the alternative sources proposed for it are lunar regolith mining and gas-giant atmospheric extraction, neither of which exists. Speculative A drive whose fuel argument ends in “mine the Moon” has moved its bottleneck rather than removed it.

Established This brief depends on Advanced Nuclear Propulsion, and the dependency is substantive rather than thematic. A fusion spacecraft needs, before any fusion happens, every capability that a megawatt-class space fission vehicle needs: radiators at the thousand-square-metre scale; power conversion and distribution at megawatt levels that has never been developed in a radiation-hardened form; a ground-test facility for multi-megawatt propulsion that does not exist; a launch authorisation under a nuclear tier structure; and appropriations that survive a change of administration.

Established And the dependency has a second, sharper leg: the most developed fusion propulsion concept in the NASA portfolio carries a uranium liner. Pulsed Fission-Fusion is a fission-assisted concept by design. It inherits the fissile supply constraint, the fuel-form question and the launch-approval tier structure directly, and its own report names criticality as one of its two unresolved problems. Frontier A reader who wants to know what stands between us and a fusion rocket should read the fission brief first, because most of the answer is there and none of it is about fusion.

Established What this brief does not depend on is the thing everyone assumes it depends on. It does not wait on Commercial Fusion, and no depends_on edge is recorded to it. That absence is deliberate and is the page's thesis in machine-readable form: the result commercial fusion is working towards — economically viable net electricity from a large stationary plant — is not the result this brief needs, and achieving it would not discharge this brief's requirement. Frontier The two subjects share plasma physics, magnet technology and a name. They do not share a figure of merit.

Speculative One genuine transfer does exist and should be granted. High-field superconducting magnets, plasma diagnostics and control, tritium handling and neutron-tolerant materials are all developed by the power programmes and all usable by a propulsion programme. Frontier That is a real and useful flow of technology. It is not the flow the framing claims, which is that a demonstrated power result would demonstrate a propulsion result.

Frontier The experiment that would move this brief is a specific-power measurement, and nobody is running it. Take any fusion device that produces net energy, weigh it including magnets, cryoplant, shielding, power conversion and radiators, and report kilowatts per kilogram. Speculative No fusion programme reports that number because no fusion programme is graded on it — and until one is, every propulsion claim is a projection from a machine designed for a different objective.

Speculative The most decisive near-term experiment is the one the Princeton concept has never done: fuse something. The field-reversed configuration exists, the odd-parity heating method has exceeded predictions on electron heating, and the step from a heating and confinement experiment to a device producing fusion reactions is the whole question. Frontier A field-reversed configuration producing measurable D-³He fusion at any gain would be a genuine landmark for this brief in a way that no inertial-confinement result would be.

Speculative On the Z-pinch side the required experiment is named in the concept's own report: demonstrate the uranium shock compression the design assumes, on the Charger-1 facility, and show that criticality can be reached without the plasma instabilities the team lists as unresolved. Established The report says explicitly that whether the compression targets can achieve it “is not yet known”, and that no breakeven has been demonstrated. That is a well-specified experimental programme with a stated unknown, which is more than most concepts in this cluster offer.

Frontier A magnetic-nozzle experiment at relevant power density is the third. Detaching a hot magnetised plasma from its confining field, efficiently, without eroding hardware, is the step that converts fusion energy into thrust rather than into heat that must be radiated. Speculative Small-scale magnetic-nozzle work exists in the electric-propulsion community; nothing has been done at fusion power densities.

Frontier And an aneutronic gain experiment at non-equilibrium temperatures would be the most consequential of all. The ESA assessment's finding that aneutronic reactions cannot reach positive gain at thermal equilibrium is a theoretical statement with a clear experimental target: demonstrate a decoupled electron and ion temperature regime that produces net fusion energy from p-¹¹B or D-³He. Speculative Several private ventures claim to be pursuing this and this brief found no fetched result to cite from any of them.

Speculative The engineering requirement set for a fusion vehicle can be stated, and stating it honestly means labelling most of it as a specification for something that does not exist. A confinement device delivering 1–10 kW/kg at over 1 MW/m³; a magnetic nozzle that converts charged fusion products to directed exhaust; radiators sized for the fraction of power that does not leave as thrust; power conversion and distribution rated in megawatts and hardened against the device's own neutron flux; and a fuel and reaction-mass budget that keeps the two separate.

Frontier The fuel-versus-propellant distinction is the requirement most often missed and it deserves its own statement. A fusion rocket needs energy and reaction mass, and they are not the same commodity. Fusing a gram of deuterium releases enormous energy and provides almost no momentum; the momentum has to come from propellant heated by that energy, or from the fusion products themselves at very low mass flow. Frontier That is why specific impulse and thrust trade against each other in every one of these concepts, and why the Direct Fusion Drive's claimed 5–10 newtons per megawatt is a small number: it is what a very high exhaust velocity buys at a given power.

Speculative The physical scale of the propulsion-native concepts is modest and this is part of their appeal. A full-size Direct Fusion Drive unit is described as roughly 2 metres in diameter and 10 metres long. Frontier That is a claim about a machine that has not fused, and the comparison worth making is that terrestrial fusion devices producing far less useful output are very much larger — which is either evidence that the propulsion configurations are cleverer, or evidence that a paper design has not yet met the mass a real one carries.

Established Shielding is a vehicle penalty here and a building penalty on the ground, and that difference is the whole reason the fuel branches diverge. A power plant surrounds its neutron source with concrete and lithium and pays nothing in performance. A vehicle carries every kilogram of shield through every manoeuvre. Frontier For a crewed vehicle the shield must also protect people, which sets a minimum that scales with distance and burn time rather than with cleverness.

Frontier Tritium and activation are engineering requirements that also make this a nuclear vehicle in the regulatory sense. A D-T stage handles tritium; a fission-assisted stage carries uranium; both activate their own structure under neutron flux. Speculative Whether a pure-fusion drive would fall inside the launch-approval tier structure written for fission systems is not adjudicated in any source fetched for this brief — the memorandum's tiers name fission explicitly — and a design carrying a fissile liner plainly does.

Established The boundary that matters most is with Commercial Fusion, and this brief enforces it strictly. That brief owns the National Ignition Facility, ITER, SPARC, the private fusion sector and its funding, tritium breeding and the tritium stock, wall-plug economics, cost per kilowatt-hour, grid integration, the gain ladder, and the question of whether fusion power arrives at all. Established This brief owns only what does not transfer: specific power in kW/kg and why no power programme optimises it; direct conversion of fusion products to thrust and magnetic-nozzle physics; the fuel-versus-reaction-mass distinction; neutron shielding as vehicle mass; the aneutronic branch; and the propulsion-native concepts. The shot record appears on this page for exactly one purpose — to establish that the transfer does not happen — and then the page moves on.

Established The dependency on Advanced Nuclear Propulsion is the other tight neighbour. Anything true of a space reactor in general belongs there and is referenced rather than restated here: radiators, power conversion, the fissile supply chain, the launch tier structure, ground-test facilities, the programme-continuity record. Pulsed Fission-Fusion is the boundary case and is treated here, with the pointer back for its fissile-supply and launch-approval consequences.

Frontier Against Antimatter Propulsion the split is by trigger. Antimatter-catalysed micro-fission and micro-fusion sits exactly on the seam: the catalysis concept and its antimatter budget belong to that brief, the fusion burn physics belongs to this one. Frontier It is worth noting which way the dependency runs — the least-implausible antimatter concepts are fusion concepts with an antimatter spark, so that brief depends on this one and not the reverse.

Established Elsewhere on the map: Interstellar Probes owns the mission rather than the drive and names this brief as one of three it waits on; Beam-Powered Propulsion is the alternative that avoids the specific-power problem entirely by leaving the power source at home, which makes it the cleanest contrast available; Advanced Fission owns the reactor physics a fission-assisted design would draw on; and High-Temperature Superconductors owns the magnet technology every magnetic-confinement drive assumes.

Frontier Off-map, the useful adjacency is plasma propulsion rather than plasma power. The magnetic-nozzle problem, plasma detachment, and high-power electric thruster engineering are worked on by the electric propulsion community at kilowatt scales, and that community — not the tokamak community — is where a fusion drive's exhaust physics would come from.

Frontier The institutional finding for this brief is that there is no institution. Fusion propulsion has no programme, no budget line, no facility and no anchoring mission anywhere in the world. What exists is a small number of NIAC studies — an early-stage grant programme designed to fund concepts precisely because nothing else will — and a university group with a commercial spin-out.

Established The budget environment it would enter is the one that just closed the fission propulsion line. NASA's FY2026 request states that near-term needs do not require nuclear propulsion and cuts Space Technology from $1,100.0M enacted to $568.9M requested, a 48% reduction. Frontier A fusion propulsion programme would be asking to start inside the same line that was just cut in half, for a technology further from flight than the one that was cancelled.

Frontier The programme-continuity record this brief would be drawing against is the fission one, and it is poor. SNAP-10A in 1965 remains the only US reactor operated in space; about a dozen start–stop–restart efforts have followed; one consumed over $400M before cancellation. Speculative A fusion drive is a longer development than any of those, which means the institutional requirement is strictly harder and the base rate is strictly worse.

Frontier There is an institutional asymmetry worth naming, because it explains why this subject is quiet rather than contested. Terrestrial fusion has a large, well-funded, politically supported constituency spanning national laboratories, an international treaty organisation and a private sector. Fusion propulsion has none of that and cannot borrow it, because the thing propulsion needs demonstrated — low mass per kilowatt — is a quantity the power constituency has no reason to fund. Speculative The most likely institutional route to a fusion drive is therefore not a fusion programme at all, but a space nuclear programme with a mission that fission cannot serve.

Speculative And there is a plain vendor-interest point. The Direct Fusion Drive's performance figures originate with a commercial entity that would build it. That is not a reason to dismiss them; it is a reason to label them, which this brief does throughout. The corresponding weight goes the other way for the Pulsed Fission-Fusion caveats, which are a NASA team publishing what its own concept has not achieved.

Frontier The honest first ethical observation is about claims rather than hazards. Fusion propulsion is a subject where the gap between published performance figures and demonstrated hardware is larger than almost anywhere else in this category, and the figures circulate widely without their caveats. Established A specific impulse of 10,000 s from a device that has never fused, and 30,000 s from a concept whose own report lists criticality as unresolved, are not dishonest numbers — they are design outputs, correctly published as such — but repeating them without that label is how a research concept becomes a public expectation.

Established The material hazards are real and are inherited rather than novel. A fission-assisted drive carries uranium and falls squarely inside the launch-approval regime. A D-T stage handles tritium, which is mobile, biologically active and difficult to contain. Neutron activation makes the vehicle's own structure radioactive over its life. Frontier The claim that fusion propulsion is inherently clean does not survive contact with any specific concept on this page.

Frontier Helium-3 introduces a resource-politics question that is usually raised in a fantastical register and has a mundane core. The aneutronic fuel propulsion wants is scarce terrestrially, and the standard proposals for obtaining it involve lunar regolith or gas-giant atmospheres. Speculative Long before that becomes a governance problem it is an engineering one, and this brief's position is that a drive whose fuel argument depends on an unbuilt extraction industry has relocated its bottleneck rather than solved it.

Speculative The dual-use question is milder here than in the neighbouring briefs and should be stated at its actual size. A pulsed fission-fusion device is a controlled sequence of small fission-boosted events, which is a sensitive technology, and the concept's inertial-confinement cousins share physics with weapon design. Frontier Against that, nothing in a magnetic-confinement propulsion device is a weapon, and the energy in a fusion drive's exhaust is far below the kinetic energy of the vehicle carrying it.

Frontier And an opportunity-cost point that applies to the whole cluster. The money a fusion propulsion programme would need is money not spent on the fission capability that is closer to flight, or on the mission that would justify either. Speculative The record suggests that the failure mode is not choosing wrongly between them but funding both intermittently and completing neither.

Speculative If specific power in the 1–10 kW/kg band were achieved with a fusion drive, the change would be genuine and it would be a solar-system change rather than an interstellar one. Outer-planet missions with orbit insertion rather than flybys; crewed transits measured in weeks rather than months; and the practical opening of the region beyond Neptune to instruments that arrive slowly enough to stop.

Speculative The arrival capability is the part worth emphasising, because it is what the published trade study actually isolates. A drive that reaches Sedna in ten years and can insert into orbit is doing a different science than a sail that arrives in seven and flies past. Frontier Fast flybys are comparatively cheap; slowing down is where the velocity budget explodes, and it is the capability that continuous high-specific-impulse thrust uniquely buys.

Handwave The interstellar register that fusion propulsion is usually discussed in is not supported by anything on this page. The best documented outer-system claim from a NASA-funded concept is 10 t to 1000 AU in 36 years, conditional on a drive whose own report says criticality is unresolved. Speculative A thousand astronomical units is about 1.6% of the distance to the nearest star. The gap between a fusion precursor and a fusion starship is larger than the gap between a chemical rocket and a fusion precursor.

Frontier The most transferable civilizational lesson here is methodological rather than technological, and it generalises well past space. Two fields can share a name, a physics, a workforce and a public constituency, and still be pursuing incompatible objectives — one optimising cost per unit output, the other optimising output per unit mass. Established When that happens, progress in one is routinely reported as progress in the other, and the error is almost impossible to see from outside because the vocabulary is identical. That is the reason this brief exists as something other than an appendix to a fusion power page.

These horizons track specific power and the fission capability this brief depends on, not fusion gain. Each is stated with what would falsify it:

• 10 yr: Established No fusion rocket, and no funded fusion propulsion programme anywhere. Frontier Expect the inertial-confinement gain record to keep improving and expect that to be irrelevant here: the laboratory's own zero-net-demand figure of about 15 at ten shots per second is a target for Commercial Fusion and not for this page. Speculative The result that would matter at this horizon is a field-reversed configuration producing measurable D-³He fusion, or a Charger-1 campaign demonstrating the uranium shock compression the Z-pinch concept assumes. Frontier Also expect the fission dependency to remain unresolved: the propulsion demonstration line was zeroed and the surface-power programme that replaced it does not produce a drive.
• 25 yr: Speculative Small-scale fusion propulsion experiments are plausible at this horizon and a flight system is not. Frontier The pacing item is not gain but a specific-power measurement on any fusion device, together with the megawatt power-conversion and radiator capabilities that the fission dependency also requires. Speculative A plausible alternative trajectory is that fusion propulsion remains a NIAC-scale study line indefinitely while fission propulsion flies first, which is what the current institutional evidence points at.
• 50 yr: Speculative A first fusion drive at this horizon is a convention rather than a derivation, and this brief marks it as one. Speculative It requires three things in sequence that have never been done in any order: aneutronic or fission-assisted net energy in a compact device, a magnetic nozzle at fusion power density, and a vehicle mass budget that closes at 1–10 kW/kg. Handwave Any date attached to that sequence is an assertion. What can be said is which single measurement would move it, and it is specific power.
• 100 / 250+ yr: Speculative At this horizon a mature fusion drive is a coherent picture for solar-system transport and remains far short of a starship. Handwave Treatments that place fusion starships at this horizon are extrapolating from an energy-density argument while omitting the mass, heat-rejection and reaction-mass budgets, which is the same error the framing under test makes at the near end.

• Depends on Advanced Nuclear Propulsion, and substantively rather than by theme. A fusion vehicle inherits the entire fission constraint set before any fusion occurs: radiators at the thousand-square-metre scale, megawatt radiation-hardened power management that has never been developed, a multi-megawatt propulsion ground-test facility that does not exist, a launch authorisation under a tier structure written for fission systems, and appropriations that outlast a cancellation. And the dependency has a second leg that is not analogical at all: the most developed fusion propulsion concept in the NASA portfolio, Pulsed Fission-Fusion, works by driving fission in a uranium liner, so it imports the fissile supply chain, the fuel-form question and the launch tier structure directly. The practical path to a fusion drive currently runs through fission.
• Enables Antimatter Propulsion depends on this brief, because the antimatter concepts that survive analysis are fusion concepts with an antimatter trigger rather than antimatter rockets — without a fusion propulsion capability the antimatter buys nothing. Interstellar Probes names this brief as one of three drives it has no independent path around.
• Adjacent Commercial Fusion is the closest neighbour and the sharpest boundary: it owns the National Ignition Facility, ITER, SPARC, the private sector, tritium breeding, wall-plug economics and the gain ladder, and no dependency edge runs from this brief to it, because the result it is working towards is not the result this brief needs. High-Temperature Superconductors supplies the magnets; Advanced Fission supplies the reactor physics a fission-assisted design draws on; Beam-Powered Propulsion is the alternative that avoids the specific-power problem by leaving the power plant on the ground. Off-map: electric-propulsion plasma engineering, which is where magnetic-nozzle and plasma-detachment work actually lives.

Handwave “Fusion propulsion follows from fusion power.” This is the framing under test and it fails at three separate joints. Established The figures of merit diverge: power plants optimise cost per kilowatt-hour and can be arbitrarily heavy and shielded; propulsion needs 1–10 kW/kg and a power density above 1 MW/m³, and nothing in the inertial-confinement record, the ITER programme or private tokamak development is measured in those units. The fuel diverges: power fusion is deuterium–tritium and neutron-rich, propulsion needs aneutronic fuel to have anything to push against, and the published assessment says aneutronic reactions cannot reach positive gain at thermal equilibrium. And the concrete concepts smuggle fission back in. Frontier A gain-15 shot would be a landmark for Commercial Fusion and would not move this brief at all. That single sentence is the page's thesis.

Handwave “The National Ignition Facility achieved net energy gain, so fusion power is basically solved.” The gain is measured against energy delivered to the target. Established The laboratory's own published figure for a facility with zero net electric demand — a test facility, not a power plant — is a target gain approaching 15, at about ten ignitions per second. The best achieved is 4.13, at roughly one shot per experimental campaign. The full wall-plug accounting belongs to Commercial Fusion; this brief carries only the laboratory's own statement of its own shortfall, which is the strongest form the point can take.

Speculative “The Direct Fusion Drive gives 10,000 seconds of specific impulse.” It is designed to. Established The Princeton field-reversed configuration on which it is based has never produced fusion. The experiments are plasma heating and confinement experiments, and the heating result — electron heating exceeding theoretical predictions by 2017 — is a real result about heating. Thrust, specific impulse, mass, dimensions and the 35/30/25/10 power split are design projections from a machine that has not fused anything, and the projections come in part from a commercial vendor.

Speculative “Pulsed Fission-Fusion reaches 30,000 seconds.” That is a Phase II target. Established The concept's own Phase I report says the time-variant models were not completed, that the unresolved problems are “reaching criticality and avoiding plasma instabilities”, that only 5% of capacitor-bank energy is assumed to convert to thermal heating with actual performance unknown, that no breakeven has been demonstrated, that whether the required uranium shock compression is achievable “is not yet known”, and that the point design is explicitly non-optimized. A NASA team publishing that list about its own concept deserves to be quoted rather than summarised, and the performance number deserves to be read as a target.

Established “A fusion drive would be clean.” A deuterium–tritium stage handles tritium and activates its own structure. The most developed concept carries a uranium liner and lists criticality among its unresolved problems. Frontier Neutron shielding is a vehicle mass penalty rather than a building cost, which is the whole reason propulsion concepts chase aneutronic fuel — and the aneutronic branch is the harder physics, not the cleaner engineering choice.

Handwave “Private fusion investment is advancing propulsion.” This brief found no fetched evidence of any fusion power company running a propulsion programme, and does not assert investment totals it could not verify. Speculative Magnets, plasma control and power electronics do transfer. Mass reduction does not, because no plant developer is graded on it.

Speculative “A fusion rocket only needs fuel, and fusion fuel is essentially free.” A rocket needs energy and reaction mass, budgeted separately. Frontier Fusing a gram of deuterium releases enormous energy and provides negligible momentum; the momentum comes from propellant heated by that energy or from the fusion products at very low mass flow, which is precisely why the claimed thrust figures are single-digit newtons per megawatt.

Speculative “Fusion will get us to the stars.” The best documented outer-system figure from a NASA-funded fusion propulsion concept is 10 tonnes to 1000 AU in 36 years, conditional on a drive that has not been demonstrated. Established One thousand astronomical units is roughly 1.6% of the way to the nearest star. The 1978 British Interplanetary Society study that supplies the popular twelve-percent-of-light-speed figure could not be obtained for this brief and no number from it is printed here.

Frontier And the correction that reverses the usual direction of the argument. Fusion propulsion is not a downstream application waiting politely for fusion power to arrive. Established It is a divergent branch, on a harder fuel, in machine configurations the power programmes are not developing, measured in a unit nobody else reports, whose most concrete embodiment needs a uranium liner to work. Reading progress in one as progress in the other is the single most common error made about this subject, and correcting it is the reason this page is a brief rather than a paragraph.