A city on Mars

Source: Inverse

A while ago, Elon Musk stated that creating a self-sustaining city on Mars would cost between $100 billion and $10 trillion, and that at a minimum, it would require 1 million tons of cargo, which means that it will have to cost as little as $100,000 per tonne to ship cargo to Mars, if this city is to be feasible.

So, let's redo the sums.  Musk has said that the cost per launch including all elements (fuel/depreciation/sunk costs) of the Starship plus its booster will be $2 million, provided both Starship and Super Heavy are fully re-usable.   According to Wikipedia, the weight of the propellant (methane plus oxygen) needed to fill Starship's tanks is 1,200 tonnes.  Each Starship will be able to lift around 150 tonnes to LEO (low earth orbit).  So that would mean that 8 launches of Starship's tanker version would be needed to provide enough propellant for one journey to Mars.  This means the cost of just getting enough propellant into orbit for a Mars mission will be $16 million.

However, the first two fleets of Starships sent to Mars won't return.  They'll be useful as habitats while surface habitats/domes/caves are built, but in any case, there won't be enough propellant to bring them all back.  Here, I estimate that 3-5 Starships will be needed just to bring the generation equipment (KRUSTY nuclear generators, solar panels and wind turbines) needed to produce the electricity to make the propellant to send one Starship home.  Each expedition will bring more generators, so the ratio of ships able to return home will rise each time.  If half the ships contain cargo or passengers, and the other half power generation equipment, it will take something like 10 expeditions, or 20 years, for enough spare generating capacity to exist on Mars to send the whole fleet of 10 ships home.  Maybe, that time line can be improved, because I'm sure colonists will quickly start making their own wind turbines and solar panels and KRUSTY nuclear generators.  All the same, in all likelihood, for the first decade or so, only some of the ships sent to Mars will return. 

Musk has said that the capital cost of each Starship will be $5 million, which won't be spread over many launches for the Mars fleet, because in the beginning they'll only be used once.  So the total cost for each ship going to Mars will be $16 million + $5 million, or $21 million.  At 150 tonnes payload, that works out to $140,000 per tonne.  Still 40% too high. 

Yet there is a very steep learning curve here.  SpaceX chose a totally unexpected material to build Starship and Super Heavy—stainless steel.  They've had to learn how to build tanks that don't rupture, a body that doesn't collapse under its own weight (the steel is just 4 millimetres thick), structures that hold up to the stresses of launch and re-entry, the Raptor engines that burn methane rather than paraffin (kerosene).  These are all totally new technologies. 

To cut costs, production has to be automated.  Musk has stated that the difficulty lies not in building one Starship (!), but in creating an assembly line that will build hundreds of Starships, efficiently and cheaply, as well as hundreds of the Raptor engines.   Building an efficient assembly line, he maintains, will be 10 times as difficult as building one Starship.  We've never had a rocket assembly line before.  Talk about a learning curve!

Perhaps the biggest cut in costs per tonne has yet to come.  Each increase in the size of the rocket causes an exponential increase in the load it can lift.  Falcon 1 (with one Merlin engine) could lift 180 kg to LEO.  Falcon 9 (with 9 Merlin engines) can lift 15,600 kg (and that's re-using the booster).  Starship's booster, Super Heavy, will lift 150 tonnes (150,000 kgs) using 25-37 Raptor engines.  Raptor engines have twice as much thrust as the Merlin, so that's equivalent to 50-74 Merlin engines.  The ratio of payload in tonnes to number of engines goes from 0.18 →1.73→2.4.  Each increase in the size of the rocket results in a more than proportionate increase in payload, and therefore a fall in cost per tonne.  Musk plans a larger version of the Starship/Super heavy combo, with a 12 metre diameter as opposed to the current 9 metres, a more than doubling of the space inside the hull.  This will increase payload to perhaps 450 tonnes, which will cut launch costs per tonne to about $100K to Mars.  My guess would be that the new giant Starship will be in operating in time for the 2027 expedition.

Who will pay for this new city, these new cities?  They will quickly start paying for themselves.  As I discuss here, the first exports will likely be tourism, both recreational and scientific.  And the settlers on Mars will start making their own hardware, because the cost of bringing stuff from Earth will be so high: solar panels, wind turbines, nuclear generators; rovers; dome construction materials; life support systems (water purification, CO2 extraction, N2 extraction); space suits; robots (for mining water and minerals, for exploration); food production.  All these will be much cheaper to make on Mars than to import.  The Mars city/cities will be a ferment of new ideas and new technologies, and some of those will be sellable—ideas/patents cost nothing to send to Earth.  The first expedition of 10 ships (cargo only) , will cost $210 million, ignoring equipment like rovers/solar panels, etc.  The crewed expedition will cost the same.  But after that cost will start to decline, as the Mars learning curve falls.  Larger Starships will take over, local enterprise starts generating income and cutting costs, infrastructure on Mars will allow Starships to be re-used.  Even without that, these sums are well within SpaceX's capacity to pay (Starlink, their high-speed broadband internet system should earn them billions annually), but it is very probable that governments, scientific organisations, millennial cults, and immigrants will be willing to pay the price of a ticket to Mars.

As I said here, the timetable still looks intact.  First cargo mission in 2022, first crewed expedition in 2025, second in 2027.  With each mission, the costs of sending one tonne of cargo or one person to Mars will fall, and as they do, the movements between the worlds will increase exponentially.  Within two decades, the Mars settlements will be paying for themselves.  And there will be thousands of Martians.