The cost of launch has been dramatically reduced by companies such as SpaceX building parts in house. Once that lowered the cost of launch as much as it could, they moved onto reusing the first stage, and finally started reusing fairings. With current programs that are already underway, such as SpaceX's Starship program, which will reuse the second stage, what is the lowest the cost of launch can get? What about with only using current technologies? What is the lower limit and how low can the cost of launch be within the known laws of physics?
I think launch costs can be broken into five general areas:
Cost of equipment you expend
Cost of maintaining and refurbishing equipment you re-use
and 2) and 4) have to be amortized over launches, while 3) & 5) are typically per launch costs. So to figure out your per launch costs you need to figure out 1 & 2 over a Starship's lifetime and divide by number of launches.
Falcon 9 has greatly reduced costs from category 1 and Starship is attempting to eliminate costs from it altogether. The big question is how much category 2 will cost for Starship.
The best comparison is airliners. Outside of daily maintenance checks, they have a series of planned maintenance activities, labeled A, B, C, and D checks.
An A check is most frequent, at every 200-300 flights, with a minimum 10 hours, so very inexpensive as a per flight cost. D checks are once every 6-10 years and taking 30,000-50,000 hours, or easily millions of dollars of mechanics time. But that's going to be a cost amortized over thousands of flights.
Lets assume a future version of Starship can achieve similar maintenance requirements as a large jetliner. That might not be entirely ridiculous, given in some ways its simpler than a jetliner. Specifically liquid rocket engines are simpler than jet engines. But life support is going to be much more complex, pressurization requirements much higher, and it's frame and body are subject to higher forces. Assume a combined A/B check (200 hours) every 6 months, C check (2,000 hours) every 2 years, and D check (50,000 hours) every 6 years. At 100 per hour, maintenance would average roughly $1M per year, without counting replacement part costs.
So how do we count replacement parts? Right now it appears most of the engines and thermal tiles will likely be replaced over a long life of a single Starship. So may other moving parts. How do we calculate that? One way is to use a percentage of build cost, say 50% as the budget using the SWAG technique (silly wild a** guess). This brings us to another critical question, how much will Starship cost to build?
Given a Falcon 9 is built out of far more expensive materials than stainless steel and costs roughly $40M, $$50M should be a reasonable budget for production cargo Starships. For crew starships life support and crew accommodations will likely add quite a bit, but I'll ignore that.
But every Starship requires a SuperHeavy in order to make orbit, and given their larger size, I'd budget $$100M for a SuperHeavy. One SuperHeavy will be able to launch multiple Starships, one each day, but to be conservative I'll pair each Starship with a single SH, doubling maintenance costs.
Now you have to guess at average life time. Given Space Shuttles flew for up to 25 years, 20 years seems like a conservative estimate. Given all our assumptions that would produce total manufacturing and maintenance costs for a Starship/SuperHeavy pair of 240M over their lifespan, or 1.2M a year. Assuming an average of 100 flights per year (two per week), that's a cost of $120,000 per flight.
Note that Elon has given 100 flight life-span as a goal for Starship, I'm explicitly assuming that future Starships will far surpass that number and average 2,000 flight lifespans. Substitute a 100 flight lifespan and these costs rise massively.
Now you have to figure out fuel, crew and pad costs. Fuel costs for a single launch have been estimated at roughly $1M, though SpaceX might be able to reduce them.
So lastly, how often does the pad and it's equipment need repair, update, maintenance, and how many people will need to be involved in each launch?
For the pad, this estimate is $5M-$10M a year plus $20M to $30M in construction costs.
The oil rigs SpaceX bought for future launches only cost $3M each, but they will likely have higher costs for maintenance and operation being that they'll be off-shore. So let's use the higher 10M a year for each pads yearly costs and assume they are used to launch almost daily. That's about 300k per launch.
So far we are roughly at 1.4M per launch, and I'm sure I've missed some things. So it might be possible that a future passenger version of Starship can reduce costs to 2M a flight, for 100 passengers, or 20,000 per seat costs. I don't know if we will see ticket costs this low in my life-time, but I'm pretty sure we won't see lower than that.
But please note how far away we are from anything approaching that. Starship hasn't made orbit, hasn't shown it's re-entry shielding tiles work yet or how much maintenance and replacement they'll need. They are still working out significant issues with Raptor 2. Solving all that is just enough to make it a useful commercial launch system. Beyond that it needs to demonstrate a workable in-orbit refueling system to be able to offer deep space flights.
It's unlikely Starships will fly without heavy refurbishment after every flight to start, to move from that to only once every 6 months will take decades. The shortest turnaround for a F9 at this point is (I believe) 27 days, being able to turnaround a much larger and more complex Starship in even 2 days is going to take a lot of development.
And there is a bit of a chicken and the egg problem with Starship at the moment. If it could fly every 2 days now, it would run out of commercial cargo for the year in a few weeks (outside of Starlink). It has to be cheap to build the level of demand that can give SpaceX the resources to build the level of cadence to fill a high level of demand. What is likely to happen is it will start flying no more often than Falcon 9 does today, and slowly improve and lower costs over a very long period of time before SpaceX can start offering daily flights at super low cost.