teleo-codex/inbox/archive/colonizing-the-solar-system.md

---
type: source
title: "Planet Ships (MISMATCH: filed as Colonizing the Solar System)"
author: "Isaac Arthur"
url: https://www.youtube.com/watch?v=oim7VvUURd8
domain: space-development
format: video-transcript
status: null-result
processed_by: astra
processed_date: 2026-03-10
priority: low
tags: [planet-ships, generation-ships, interstellar, isaac-arthur]
notes: "TRANSCRIPT MISMATCH: Contains Planet Ships episode about moving entire planets between stars, NOT colonizing the solar system overview. Out of scope — too far-future for investment lens."
---

## Transcript

This episode is sponsored by Brilliant
We so often talk about building spaceships to visit and colonize new planets, but what
about making a spaceship out of planet? So today we’re back to the Generation Ships
series to discuss building spaceships that are of a planetary scale or even outright
moving entire planets between stars or even galaxies. And incidentally, if you’re new to the channel,
welcome to SFIA, probably the only place on the internet where a serious discussion about
moving entire planets would qualify as a fairly mundane. Why would you ever want to move an entire
planet? That's a good question, but it turns out that
we might have a few good reasons. First, we might realize that something is
about to go terribly wrong with our Sun, as will happen when it begins to run out of fuel
and slowly heats up and expands, and we'll want to move Earth. Even as early as about 1 billion years from
now, the Sun’s luminosity may have increased sufficiently to render the Earth uninhabitable. We’ve talked before about ways of extending
a star’s lifetime but those are very time and labor consuming tasks. Moving a tiny rocky planet like Earth is,
comparatively, a weekend project, so you might decide to just migrate Earth to a new, younger
solar system. Of course we could face a more immediate solar
emergency, perhaps of an artificial variety, such as an artificial black hole being dumped
into the Sun. You could also imagine disputes resulting
in a planet being kicked out of its native system. We'll discuss this possibility later on. However, the most likely scenario under which
you'd want to move a planet would be to serve as a colony ship. Something we noted earlier in the series,
in the Million Year Ark, is that for very long voyages - such as to another galaxy - you
need vast ship-space and resources, enough to be stable and redundant for timelines longer
than any civilization has lasted thus far, let alone any machine we’ve built. We do however have an example of a spaceship
that can last billions of years. It's the Earth. We don’t really know what the minimum size
and mass of a ship needs to be to survive very long journeys, beneath which it can breakdown
mechanically, genetically, or socially, but we know Earth will do the trick since it already
has. Is it even possible to move a planet, and
how would we go about doing it? Well, it doesn't violate the laws of physics. There are a few challenges to overcome, but
it turns out that moving a planet is fairly easy and really just involves the brute force
application of vast amounts of energy. Moving a planet on human time-scales is, however,
another story, and the first of our challenges. Planets aren’t really designed for rapid
acceleration, even less than an O’Neill Cylinder converted into a spaceship, which
we found out was quite a pain earlier in the series in Exporting Earth. Consider, the Moon exerts gravity on the Earth,
it’s about 60 times further from us than the Earth’s radius, and about an 80th of
the mass, so it exerts about 3.5 microgees of acceleration on us, compared to the Earth's
1 gee. If it’s on the opposite side of the planet
from you, you’re about 3.5 millionths heavier than normal, if it’s above you in the sky,
3.5 millionths lighter. Yet even this tiny force and acceleration
is still enough to cause the tides, a significant disruption to the surface of the Earth. The Sun, 400 times further away but much more
massive, actually does about double that, and sometimes the two will be in about the
same direction and combine force to about 10 microgees. We know the Earth can handle an acceleration
on par with this since it does it everyday, but going much higher would potentially cause
a lot of severe tidal effects as water and air migrated toward one side of the planet. You could mitigate this somewhat with engineering,
like coastal walls. That’s a big project of course but small
compared to moving a planet, but you probably can’t take this too far since you'd also
have to worry about the tectonic plates, mantle, and core shifting around on you and that would
be much harder to deal with. Still you can probably do a lot better than
10 microgees, but for now we'll assume that's the maximum acceleration the Earth can handle. To put 10 microgees in intuitive terms, it
takes about one day to accelerate an object at this rate to the same speed as one second
of 1 gee acceleration would take. Remember that 1 gee is the acceleration you
experience when you fall on Earth. It takes about a year of accelerating at 1
gee to get to about light speed, so it would take 100,000 years to reach about the speed
of light at 10 microgees, or 1000 years to get to 1% of light speed. That by the way is still way faster than a
standard Shkadov Thruster can push a star up to interstellar speeds, and again you can
push it faster if you do some heavy modifications to deal with tidal issues. If you wanted to move your planet 10 light-years
away, and you did it under constant acceleration, it would take about 1000 years to hit the
midway point and your maximum velocity at turnover would be 1% of light speed. You’d arrive 2000 years after you left. That’s actually not bad for moving a planet
but a more conventional spaceship able to handle higher accelerations would seem preferable,
and you could build a planet’s surface area worth of O'Neil cylinder ships using a lot
less mass than a natural planet has. However, if you’re moving a planet because
you want that specific planet elsewhere, this might be a viable timeline. The longer the distance of the voyage the
less acceleration rates and times really matter. Push out to 1000 light-years, a hundred times
further away, and you’d hit a maximum speed of 10% of light and take only 10 times longer
to arrive, 20,000 years. And if you wanted to go 100,000 light years,
which is to say across the entire galaxy, that would get you pretty close to light speed
and that would be a 200,000 year journey, assuming that we ignore special relativity
and time dilation for the moment. That’s only twice as long as a conventional
spaceship would take, even one with no organic crewmembers that can handle lethally high
acceleration rates. For distances beyond that, at the intergalactic
scale, acceleration rates are essentially irrelevant. Though for slower accelerations you would
need longer distances to get up to speed, which will matter when we get to discussing
where we are deriving the energy to accelerate our Planet Ships. Energy is a big deal in three other ways besides
finding it for thrust though. First, planets are very good at storing heat
and a lot of the ways you’d be applying energy as thrust will lead to excess planetary
heat. Earth normally emits a couple hundred million
gigawatts of waste heat mostly from absorbed sunlight, and even adding a few million more
gigawatts of energy to that flow is going to have a noticeable effect on surface temperature. If you’re trying to push Earth up to a decent
percent of light speed you’re talking about adding somewhere around 10^40 or 10^41 Joules
of kinetic energy to it. If even a tiny fraction of that is being absorbed
as heat, say 10^38 joules, and you can only let a few million gigawatts extra into the
system, or about 10^23 Joules a year, you’d need a quadrillion years to push it up to
speed without roasting your planet. So you either need to use a method that produces
almost no new heat being absorbed by Earth, or you need a faster way to pump heat off. Fortunately we have a slight edge here, since
moving a planet through the interstellar void is going to result in a loss of sunlight. Even then, though, it’s too much energy
to deal with in a reasonable amount of time, so you'll need to make sure you’re absorbing
very little of your thrust-energy as heat, but you do want to absorb a bit of it because
your planet is going to freeze otherwise. This would defeat the purpose of sending a
living planet in the first place. Basically you need to add the sun’s light
worth of energy to the planet for the duration of the trip, since the Sun is not coming along…
this time anyway. We’ll discuss moving entire solar systems
in the next episode of the series, Fleet of Stars. Radiation and collision are our other energy
concerns. The interstellar void has the potential for
both of these in abundance when you’re traveling at high speeds. Even in intergalactic space, and with good
point defense to blow up objects in the way, these are going to be a bit much for a planet’s
magnetosphere and atmosphere to handle. You do not necessarily need to englobe the
planet with some big shell though, since the vast majority of the dangerous stuff is coming
from directly in front of you as you move. For instance a big disc-shield a bit larger
than the planet could be placed in front, and you could probably set that up at a distance
that made it look no bigger than the moon or Sun in the sky. You could, perhaps, engineer the planet-side
of the shield to serve as an artificial Sun, similar to what we discussed in the episodes
in our Megastructures series, Flat Earths and Making Suns. Of course, you’d probably want to bring
the Moon along if we were moving Earth. Conveniently, the Moon could serve as a Gravity
Tractor, which is the simplest method of moving a planet, though not ideal for high speeds. The Moon orbits the Earth, and as mentioned
pulls on the Earth too, indeed with the exact same force the Earth exerts on the Moon. If we push on the moon with something, it
will move, and if you push too fast it will fly away from Earth, unless you’re pushing
it toward Earth. If you just want to push the Earth further
from the Sun, you could push on the moon when it’s between the Earth and Sun, moving it
away from the Sun but toward Earth. You would then push on it again when it was
on the far side of Earth from the Sun, again away from the Sun but also now away from Earth. That will cancel out its motion relative to
the Earth, but not the Sun, and the Earth will have been nudged by gravity away from
the Sun. The fastest you can push is equal to the force
the Earth exerts on the Moon, otherwise it will fly off, but as mentioned, that’s the
same force the moon exerts on Earth and as you’ll recall, more or less the maximum
acceleration the planet can handle since we used the tidal effects of the Moon and Sun’s
gravity to place that limit. The gravity tractor approach certainly works,
and isn’t limited to using the moon. For example, You could also send a string
of asteroids by the Earth so that each exerted a small gravitational pull on the Earth as
they flew past, or you could place large but weak engines in orbit around Earth that didn’t
produce enough thrust to break out of orbit, yet produced it for a very long time in only
one direction. One problem though is that tidal forces cause
tidal heating and while gravity lets us avoid touching the Earth while we move it, that
gravity is still producing some heat. So this method is fine for slowly moving planets
further out in their own solar system with low inputs of tidal energy, but not ideal
for moving a planet quickly, which would require high inputs of tidal energy. It is also heating up the moon with whatever
you are using to move the Moon, and that’s not ideal either. So why not just apply force directly to the
Earth instead of the Moon? That is an option, but it's problematic. In theory you can put giant rockets on the
Earth, or detonate nukes on the planet's surface, but the Earth has an atmosphere that’s going
to absorb almost all that energy as heat. This is the problem with using something like
the Fusion Candle, our trick for moving gas giants, where huge platforms in the atmosphere
suck in hydrogen, fuse it, and blow it out into space as propellant. We don’t really care if those planets get
hot, and they typically are fairly low density with more effective radiating surface area,
so this works better for them. In fact, you could move a rocky planet of
your choosing into a stable orbit around a gas giant that has a Fusion Candle; then,
when the fusion candle moves the gas giant, that rocky planet could come with it for the
ride. Of course, you'd have to get the rocky planet
into an orbit around the gas giant by using some other method of moving it, which brings
us back to the problem at hand. This is also a good approach if you wanted
to move Venus further from the Sun and get rid of a lot of its atmosphere, which could
be used as a propellant. But there are two engineering options for
enabling the use of direct rocketry on a planet without heating the atmosphere. The first is to selectively remove the atmosphere
in a small area around the rocket, which can be done by building a very high thin wall,
like a big rocket nozzle, that goes up above the atmosphere so that the rocket is basically
in a vacuum. This is sort of the reverse of the partial
terraforming trick that we often see in science fiction, where a high wall is built around
the area intended for habitation, and that area is filled with air. Or, if you want to take this even further,
you could actually englobe the entire planet, giving it an exterior spherical shell that
had no atmosphere. Indeed it’s quite likely worlds looking
to move beyond Ecumenopolis levels into being full-blown, many-layered Matrioshka Shellworlds
might leave their top layers airless anyway, to facilitate off-planet transport and trade,
so this would be a great option if they wanted to move their planets. The second option for enabling direct rocketry
is related to the construction of the shell world. A giant sphere around a planet will require
some sort of support, which could involve the active support system we call, appropriately,
an Atlas Pillar. It's basically a big space tower. You could just put your big rocket on the
top of a space tower that reached safely above the atmosphere. Of course you’re not using any sort of conventional
rocket, not for moving planets, chemical fuel ain’t gonna cut it. Even fusion is only going to work if you’ve
got a hollow shellworld full of fusion fuel rather than molten metal, and even then it
will only allow speeds good enough for moving over interstellar distances. What we really want are technologies enabling
travel over intergalactic distances, because we’re interested in Planet Ships, not just
moving planets we want elsewhere. A shell around the world full of fusion fuel
will provide shielding, though. You could use a big external shell that wasn’t
just a thin shield but a bunch of thin hollow tanks full of hydrogen, which is very good
at absorbing radiation and of course is a good way to store a lot of fuel, since mass
arranged around something as a spherical shell exerts no gravity on the inside. Or no net gravity anyway, a topic for another
time but you can use a very massive hollow sphere as a way to slow time down for those
inside it. Regardless, such a shell full of fuel is a
good way to store the fuel you need to slow down and to run life support for that planet,
which in this case is just artificial sunlight since it is an entire planet. We could also potentially use artificial black
holes both as a power supply and a gravity tractor, either via smaller ones emitting
hawking radiation, or bigger ones. We’ll be looking at that more soon, but
fundamentally, while they’d offer a higher velocity than fusion, as could something like
antimatter if you can make a planet’s worth of it and dare store that, they still have
that basic problem of the rocket equation. You still have to carry all your fuel and
pay the mass penalty for carrying it. We’ve spent a lot of time early in this
series specifically talking about alternatives to avoid the limitations of the rocket equation,
mainly light sails, laser sails, and the stellaser. This is going to be the method that lets us
really make planet ships viable for high speeds and intergalactic colonization, and inter-supercluster
colonization, which you can probably only do with a planet ship. You don’t necessarily need to use a planet,
but a lower mass object like a moon exposes you to slow material leakage as you have no
decent natural gravity well holding things together. So you do have to be looking at things on
that scale if you want to seriously contemplate trips that might take many millions or even
billions of years without resupply. This is also one of the few cases where you
might build a stellaser all the way up into the Nicoll-Dyson Beam, Death Star levels of
output. Converting an entire sun into a giant laser
cannon sounds cool, but it’s overkill for pushing a ship and not the best way to weaponize
a star, either. You could just use a swarm of Relativistic
Kill Missiles, each accelerated by smaller lasers suitable for accelerating a normal
spaceship. This is something we’ve discussed a few
times before, most recently in the Dark Forest Theory episode if you want details on that. Lasers give us some big advantages, especially
for slow acceleration. Mirrors can be made highly reflective, so
they absorb very little of the light incident on them as heat, and indeed since we’re
accelerating quite slowly initially we can bounce that beam back and forth many times
to maximize the push. Now you can’t just push a planet with a
beam, not without melting it, as the atmosphere will absorb that light, but we’ve already
discussed some options for dealing with an inconvenient atmosphere. We can install big mirrors on the ground in
areas evacuated of air, we can put those mirrors on the Moon which then acts as a gravity Tractor,
or in orbit on hefty mirrors platforms, potentially O’Neill Cylinders to provide extra living
room. We can hang them above the atmosphere but
attached by space towers to the ground, or we can just build a big reflective sphere
around Earth, which is bigger than Earth itself, so also gives us more surface area to radiate
absorbed heat away. Now, we’ve discussed pushing with lasers
quite a bit before, and if you’re bouncing the beam off the target, you need 1.5 gigawatts
of laser for every ton you want to push at 1 gee, or 1.5 megawatts per kilogram. We want to do only 10 microgees though, so
we only need 15 watts per kilogram. The Earth's mass is 6 x 10^24 kilograms, so
we’d need 9x10^25 Watts of power, and conveniently the Sun produces about 5 times that, so we
don’t even need to the extra advantage of repeatedly bouncing the beam and we can still
get away with accelerating the Earth away at about five times our preferred rate. But if we did need to get the Earth moving
quickly, we could pour the juice on, with full sun-power and beam bouncing, and we could
continue targeting that beam quite far out since a planet is a pretty big target to keep
a lock on. And of course if it has a shell, that target
could be even bigger. It also means we don’t need to be too picky
about what other stars we use along the way, because we don’t need to limit ourselves
to the small fraction of stars as bright or brighter than our own Sun. Furthermore, we could also be boosting our
planet ship with lasers generated from multiple stars, since a target as big as a planet can
be hit from many light-years away. It also means you can send supplies along
the way, as we discussed doing with colonial fleets in a previous episode. This is an entire planet, so it could have
whole armadas of ships and habitats swarming around it that were jumping ahead, or off
to the side, to colonize or set up new stellasers. Those smaller ships don’t have to arrive
around a waystop sun ahead of it either, since the planet isn’t stopping there and can
push them down to speed with planet-based lasers as it flies by the star so they can
slap together another stellaser that can shoot the planet ship and push it faster. And you do want to be building more pushing
stations along the way because that slow acceleration means your planet ship needs a very long laser
highway to get up to cruising speed. If you want to get it close to light speed
at that slow acceleration, it basically needs 100,000 years to get up to full speed, and
would have crossed a big chunk of the galaxy during that process. It’s also going to need the same to slow
back down again at the destination, and you will be needing to send out vanguards ahead
to build the necessary stellasers to slow it down. Now in truth you probably won't need to build
anything to get the planetship up to speed, since odds are you’re doing something like
this after you’ve already colonized a lot of other systems and already have a lot of
laser highways setup, though I’m sure you’ll need to do some modifications since a planet
ship decidedly qualifies as a ‘wide load’ for your laser highway. You will need to install new laser highway
infrastructure for the deceleration portion of the trip, since presumably you're travelling
to a new destination that hasn’t been colonized yet, unless you’re just shipping home bulk
matter for building something enormous like a Birch Planet, where the destination already
has the infrastructure to slow the planet down. But amusingly you won’t always need to do
as much slowing as you did speeding up. And with this we get to the real purpose of
these things. We don’t need them for colonizing our own
galaxy, and even our nearer neighbors like the Andromeda galaxy probably do not require
this level of effort. A planet ship is not intended to colonize
a single solar system; it’s the ultimate gardener ship. Its purpose is to sow a line through a galaxy
leaving a thick trail of seed colonies in its wake. Indeed, you might not even try to stop it,
just detach fleets of smaller colonial ships and push them to slower speeds with the planet
ship's own lasers. These seedling colonies can then grow, build
local stellasers and send fresh boosts of energy and materials to the mother ship for
its continued journey down the intergalactic road. And they’d grow fast too, because you don’t
have to make colonies with just a few thousand people, such a planet ship is probably a Ecumenopolis
peopled by trillions who can easily dispatch a billion colonists and trillions of tons
of colonial gear every decade or so as it passes by a good colony prospect. Indeed it could be dispatching whole fleets
to several systems in a fairly wide cylinder along its path. The planet ship need not stop in any galaxy,
but can fly right through, and get a course correction to intercept another galaxy down
the road. This is where the planet ships excels, because
it can contemplate multi-billion year journeys, and there are many interesting destinations
billions of light years away. And remember, all this stuff is moving away
from us as the universe expands. Hubble Expansion is about 7% of light speed
for every billion light years of space between locations, so a ship hoping to reach a place
a billion light years away needs to be doing more than 7% to ever reach it, and will arrive
seeming to be moving slower. So, if you send out a ship at 8% of light
speed, it will arrive at only 1% light speed, making it much easier to slow that ship down
on arrival. Of course it will take a hundred billion years
to get there, so you probably want to be going a good deal faster, even just jumping up to
9% of light speed would half your journey time. But we also don’t necessarily care about
ever slowing that planet ship down, indeed we might keep pushing it faster and faster,
because we can always evacuate the population off in smaller ships if we want, as we've
discussed, or just let it meander through the eternal void until it exhausts its onboard
energy supplies for lighting itself during trips. But, as we discussed back in the episode Dying
Earth, this will be many trillions of years if its fusion fuel supply was a decent fraction
of the planet’s mass. But why limit ourselves to 9% light speed? With our ultimate planet ship, accelerated
up to, say, 70% of light speed, we could conceivably travel to galaxies that are currently 10 billion
light years away. And at 99% light speed, or higher, folks on
the ship will experience relativistic effects, and time will slow down. The intergalactic void is quite thin so these
higher speeds are actually more viable than within galaxies. Again, a planet is a huge target for a laser,
and with its mass and atmosphere it can handle debris collision risks a lot better than a
ship, because it can absorb a much bigger whack without being critically damaged, meaning
its various point defense and detection gear won't have to work as hard at the same speed
as a smaller vessel's would. A smaller ship can’t afford to miss even
a single pebble at near light speed because it will detonate with the energy of a nuke. A planet can take that strike, and its sheer
size can house much more detection gear, point defense, and whole armadas of tender ships
and vanguards. Planet sized ships need not be an actual planet
though, merely things closer to that scale than the classic spaceship we see in scifi. Of course when you’re at the point that
you’re thinking about colonizing other galaxies then planets or armadas in that general size
zone are not much of a problem to source, you’d have billions to spare, handy too,
since there are billions of galaxies we could colonize this way. And this gives us the approximate answer to
just how far off we can ultimately colonize without faster than light travel, at least
10 billion light years. Might as well think big. I always like pointing out that things like
this, which are totally allowed under known physics, tend to be the sorts of things even
scifi with lots of super-science and Clarketech won’t touch as plausible. But moving planets is just raw brute force,
not ultra-high tech, though doubtless more tech will help. And we might need to move Earth one day. It’s a pretty unique place we’d want to
save, rather than disassemble to be part of a Dyson Swarm. Amusingly you might even get kicked out of
your home system by that Dyson Swarm too. Planets have a lot of concentrated mass that
represents a lot of perturbation on neighboring objects. It’s manageable but a hassle, and I could
see the quintillions living in a Dyson Swarm telling the billions back on Earth to either
let them disassemble the place or pick up the planet and move, and more so for places
like Mars or Venus that aren’t humanity’s cradleworld. ‘Marxit’ or ‘Vexit’ scenarios might
arise, and some place like Saturn or Neptune, which would already be pretty artificial and
not dependent on sunlight anyway might be willing to pack up and leave. We’ve talked about how O’Neill Cylinders
in a Dyson Swarm might pack up if they didn’t like their neighbors, these artificial worlds
are already basically spaceships to begin with. It does make me wonder if future galactic
civilizations might have the equivalent of solar divorces, where one side keeps the sun
and the other gets most of the planets, or heck, they might starlift a chunk of the Sun
off to take with them to make a red dwarf out of. We’ll play around with moving solar systems
next episode in the series though, in Fleet of Stars, and dig in more to the Supernova
Engine we talked about last month in Dying Stars along with Starlifting Binary Shkadov
Thrusters, literal “Starships”. I do suspect in most cases a planet ship will
tend to be a much more artificial thing built to planetary scale but designed with space
travel and higher acceleration in mind. But now we can see that true Planet ships
are possible, and put a whole new meaning on ‘Spaceship Earth’. So I was talking a moment ago about how this
sort of endeavor is extreme even by scifi standards but is actually a fairly simple
process inside known physics, we can dream concepts like this up by knowing our math
and physics and so often can find truly amazing ideas that way. We rarely get a chance to dig into the details
of how that math and science works here, and partially that’s because learning them is
best done at your own pace in an interactive environment, not by lecture. That’s where our sponsor, Brilliant, really
excels. They have a wide range of courses on math
and science that let you pick the area you want to learn, start where you are ready to
start, and practice that knowledge through interactive quizzes before moving on to more
advanced topics when you are ready to. They also have daily problems in math, science,
and engineering, so that you can learn something new in 5-10 minutes every day. If you’d like to learn more science and
math, go to brilliant.org/IsaacArthur and sign up for free. And also, the first 200 people that go to
that link will get 20% off the annual Premium subscription, so you can solve all the daily
problems in the archives and access every course
So often on the channel we discuss dreams of more advanced technology in a bright future,
but that’s not everyone’s dream and next week will return to the Rogue Civilizations
series to look at potential colonies settled by Techno-Primitivists, and we’ll see how
that might work out. The week after that our episode will be on
National Pet Day, and we’ll take some time to look at what the future might have in store
for our furry friends. For alerts when those and other episodes come
out, make sure to subscribe to the channel, and if you’d like to support future episodes,
you can visit our sponsors or donate to the channel on Patreon. Until next time, thanks for watching, and
have a Great Week!