Why here
The physics never failed. The sponsor did.
In April 2016 Yuri Milner, Stephen Hawking and Mark Zuckerberg announced Breakthrough Starshot: gram-scale craft on 4-meter lightsails pushed to 20 percent of light speed by a ground-based laser array, reaching Alpha Centauri in about 20 years. The pledge was $100 million against an eventual mission cost Milner put at $5 billion to $10 billion.
By late 2025 the program was on indefinite hold, with no announcement and no closing statement. Public accounting suggests roughly $4.5 million was actually disbursed across about 30 contracts. Nine years of study identified hard engineering problems and solved none of them, and declared none of them impossible. The materials work got better, not worse.
It died because in 2016 it was philanthropy attached to nothing. There was no industry whose ordinary operations produced gigawatt beam arrays, no supply chain whose ordinary output included square kilometers of thin-film collector, and no institution with a 50-year balance sheet that had any reason to care.
That is the whole argument of this section, stated in advance. The interstellar question in 2016 was a physics question asked by people who had no industry. Asked in 2080 by people who own a beam-capable power infrastructure inside 0.4 AU, a fusion fuel chain in the ice giants, a fabrication base on the Moon, and autonomous systems that have run unsupervised past Ceres for 30 years, it is not easy. It is a procurement decision rather than a research program, and procurement decisions get made.
Envelope
What bounds it.
The constraint fingerprint. Two of these are irreducible — they are physics rather than engineering — and the rest follow from them.
| Parameter | Value | Consequence |
|---|---|---|
| Kinetic energy at 0.2c | 515 GWh per kg | Irreducible. Relativistically corrected; no design choice moves it |
| Beam-to-kinetic coupling | 20% at 0.2c | Irreducible. Coupling equals 2β, so 4/5 of every joule leaves as light |
| Cruise velocity | 0.2c | Set by human attention span, not physics: below it no target fits a career |
| Mission time | 6× distance in ly | Transit plus one-way light time. Proxima is 26 years, Teegarden 75 |
| Payload | Grams, beamed; tons, nuclear | Two branches that share no hardware |
| Deceleration | Mass ratio e⁴ at 0.1c | The exponent is where programs die. A round trip is e⁸, about 2,981 |
| Round-trip light time | 8.5 years to Proxima | Roughly 7,000 times the step from Ceres to Pluto. Full autonomy is the only mode |
| Solar gravity well | Irrelevant above ~0.005c | Escape from Pluto saves 0.06% of a 0.2c velocity budget |
Three families
There is no such thing as “an interstellar mission.”
The deceleration exponent splits the problem into three programs that do not resemble each other.
| Family | Mass | Speed | Deceleration | Verdict |
|---|---|---|---|---|
| The flyby | Grams to kilograms | 0.1–0.2c | None; one pass | Achievable within the century |
| The arrival | Tons | 0.05c or slower | Nuclear stage or passive brake | Plausibly started this century, arriving next |
| The settlement | Thousands of tons | 0.01c or slower | Not the binding problem | Energetically modest, institutionally impossible |
The settlement row is the counterintuitive one. A 10,000-ton generation ship at 0.01c needs about 7 percent of current world annual primary energy and arrives at Alpha Centauri in 425 years. Velocity is quadratic and mass is linear, so patience is worth vastly more than mass reduction: a civilization that could tolerate a 425-year transit could send a small city for a fraction of a year’s energy budget.
The barrier to the generation ship is not energy. It is a chartering institution that persists for 425 years and a crew society that stays functional and loyal to an original purpose across about 17 generations, with no possibility of correction and no meaningful communication. Nothing in human institutional history suggests that is achievable.
Hazards
What damages it.
| Hazard | Character | Defense |
|---|---|---|
| Dust at 0.2c | Continuous, cumulative | Statistical, not structural: 5,000 craft, 90% loss accepted. The zodiacal crossing alone is 10–35% of the total dust column |
| Sail absorption | Instantaneous | Absorptivity below 10⁻⁵ across a 22% Doppler band. Above 10⁻⁴ the sail reaches 1,134 K and fails |
| Radiation over a 21-year cruise | Cumulative | Destroys unshielded silicon. No maintenance, no reboot from the ground |
| The array as a weapon | Political | The same aperture holds steel at melting point to 60 AU. Mutual vulnerability and trivial attribution |
| Pointing loss | Continuous during burn | Feed-forward only. The beam must stay smaller than a 4-meter sail for the whole acceleration |
| Institutional decay | Over decades | The real one. An answer arrives from a machine designed by people who have retired |
Four readings
The same problem, four questions.
The energy never leaves home. A fixed array does the work and the craft is a sail, which is why the coupling ceiling of 20 percent is the number that matters and not the efficiency of anything aboard.
The output of the whole system rather than a node in it. Every component was built somewhere else for another reason.
Δv stops being the axis above 0.005c. What binds is beam power, sail absorptivity, and receive aperture.
Attrition beats redundancy here, and nowhere else on this site. The vehicle is cheaper than the environment is hostile.
The corridor
Not a launch site. A line that already points outward.
The intuitive move is to pick a departure point high in the Sun’s gravity well. That intuition is imported from Earth launch and does not survive the change of scale: departing from Pluto rather than 1 AU saves 35,420 meters per second of solar escape, which against a 0.2c mission is 0.06 percent of the velocity and five parts in 10 million of the energy.
What the arithmetic produces instead is a corridor, and every node on it was placed for reasons that had nothing to do with the stars.
| Node | Solar flux | Collector for 100 GW | In-situ volatiles | Branch served |
|---|---|---|---|---|
| LEO / 1 AU | 1,361 W/m² | 368 km² | None | None |
| Mercury orbit, 0.39 AU | 8,948 W/m² | 56 km² | Regolith, polar ice | Beamed: optimal |
| Jupiter, 5.2 AU | 50 W/m² | 10,000 km² | Helium-3, water at Callisto | Fuel supply |
| Neptune, 30 AU | 1.5 W/m² | 333,000 km² | Helium-3, deuterium | Fuel supply |
| Pluto–Charon, 39.5 AU | 0.87 W/m² | 573,000 km² | Water, N&sub2;, CH₄, tholins, D | Nuclear: optimal |
| SGL nodes, 550+ AU | 0.0045 W/m² | n/a | None | Targeting and downlink |
The corridor runs Mercury to the ice giants to Pluto to 550 AU. Mercury for power, the ice giants for fusion fuel, Pluto for the volatile stockroom and a near-free elevator, the 550 AU line for astronomy. The interstellar architecture does not require building anything new. It requires noticing that the corridor already points outward, and deciding to use it.
In this section
The Interstellar pages.
14 pages. Three of them carry the argument: the irreducible number, which says what cannot be engineered around; the solar gravitational lens, which changes a mission parameter by eight orders of magnitude and never leaves the solar system; and interstellar objects, which is already returning data.
Contested
Deliberately Not Claimed
Nothing arrives this century. Even an optimistic 2085 launch at 0.2c reaches Alpha Centauri in 2106, with the first data on Earth around 2110. The century closes with the corridor built, the beam operating, several thousand craft in cruise, and nothing arrived. That is what the first quarter of a crossing looks like, and it should be said without softening.
The sail absorptivity specification is the hardest single number here. Below one part in 100,000, across a 22 percent Doppler band, at a gigawatt per square meter, in a membrane 100 nanometers thick, strong enough to survive 10,000 g. It is a thin-film photonics problem rather than a propulsion problem.
The interstellar-object number density is known to a factor of about 50. Published Rubin detection-rate predictions span 0 to over 100 per year, all extrapolating a galactic population from three detections. Everything downstream depends on it — intercept frequency, how much parked Δv is worth holding, and how violent planet formation is on average.
Image tracking at the gravitational lens focus is unsolved. Following Proxima b’s projected image costs about 20 km/s per year, sustained, forever. That is a propulsion program rather than a station-keeping budget, and it is a consequence purely of the planet being close to its star.
Record
Provenance.
- Presence
- Uncrewed. Crewed interstellar flight is not in hand by 2100 by any architecture at any speed
- Reference beam
- 100 GW at 1.06 µm, shared with Mercury; the sail beam line is dedicated hardware
- Nearest target
- Proxima Centauri, 4.2465 ly; Alpha Centauri AB at 4.34 ly, 0.21 ly from Proxima
- Reviewed
- Q3 2026