I know that they used to have this idea, back in the day, of a "space gun" to fire somebody into space, and I understand why that was not safe or practical/feasible.

However, once they had come up with modern rockets, why have them start with the maximum amount of resistance, pointing straight up? Why not have them go like an airplane at first, and then point it up once it had got off the ground and gained some momentum? It seems like such a massive waste of resources and money to have it stand perfectly still and just fire away straight down and have it slowly ascend into the air and further out into space.

  • 8
    $\begingroup$ An airplane has wings, a rocket not. $\endgroup$
    – Uwe
    Apr 26, 2020 at 9:42
  • $\begingroup$ I think it's just more feasible and not so costly. Some missiles do have wings and launch in an angle. Certain rocketplanes even launch horizontally from other planes: the X-15 from the B-52 and the SpaceShipOne and Two from WhiteKnight and WhiteKnightTwo. $\endgroup$
    – user35272
    Apr 26, 2020 at 9:54
  • $\begingroup$ Petrel and Skua = barrel launched at an angle; Pegasus = wings and near horizontal launch; Super Strypi = rail launched at an angle $\endgroup$
    – Puffin
    Apr 26, 2020 at 10:19
  • $\begingroup$ These are all small-ish ventures so perhaps there is a structural disincentive to launching a large distributed mass at an angle. Also with a vertical launch the points of contact during launch preparation are relatively concentrated at the base of the launch vehicle. $\endgroup$
    – Puffin
    Apr 26, 2020 at 10:25
  • 5
    $\begingroup$ You always want to minimize the time in atmosphere because travelling in dense air is expensive. Launching is expensive because you need 7.9km/s delta v. Lifting the rocket to 20km and mach 2 doesn't do much. And finally, building a rocket that can handle lateral force is expensive too, in money and in weight. $\endgroup$ Apr 26, 2020 at 12:07

5 Answers 5


One reason large rockets are launched directly up is structural. Cylinders are strong under compression, stacking cylinders on top of each other means the weight is symmetric, you need less structural weight to hold it all up. Launch it straight up and make gentle changes in direction and the forces are equally distributed through the structure all the way up. Turn it all onto its side and suddenly you need a whole lot more structure to support the weight, add wings and you need even more structure to distribute the aerodynamic load.

Another is that air resistance is the enemy. Launching sideways would mean the rocket would have to overcome air resistance for longer. On an airless planet a sideways launch might make more sense, with an atmosphere the most efficient route to orbit means getting high enough to be past much of the atmosphere before tipping over and accelerating to orbital velocity.

  • 3
    $\begingroup$ Imagine the landing gear for a sideways Saturn V shudder. $\endgroup$ Apr 26, 2020 at 14:46
  • 3
    $\begingroup$ OMG yes @OrganicMarble! Imagine the runway too! $\endgroup$
    – GdD
    Apr 26, 2020 at 14:55
  • 1
    $\begingroup$ @SolomonSlow if you know of a term for "gear that is only used at takeoff" please inform! $\endgroup$ Apr 26, 2020 at 17:41
  • 3
    $\begingroup$ @OrganicMarble Its called a "ground carriage" if there's no positive affixment, or a "jettisonable landing gear" if the pilot has to do something to drop the wheels. The Luftwaffe AR234 and ME163 had this, and landed on a skid. GroLaS and GABRIEL are a current ideas along this line about landing a craft with no wheels, but we all know Thunderbirds did it first :) $\endgroup$
    – Criggie
    Apr 27, 2020 at 0:19
  • 1
    $\begingroup$ @Criggie thanks, epecially for the Thunderbirds reference :) $\endgroup$ Apr 27, 2020 at 0:55

It's a matter of optimal trajectory - pitch maneuver/gravity turn which depends on characteristics of the rocket, the atmosphere, gravity etc. In particular, for rockets with lower initial thrust-to-weight ratio, the trajectory starts almost vertical; "rounding" the angle to perfectly vertical makes the launchpad infrastructure and preparation process easier; the "inefficiency" is so minuscule any saving here would be completely swallowed by costs of complication of the infrastructure for diagonal launch.

And what about rockets with very high initial TWR? Well, these that exist, like SS-520-4, launch diagonally or in other "interesting" configurations. But they are exceptionally rare, for simple economical reasons. "Fuel is cheap, engines are expensive" is the old saying of the rocket industry. It's more cost-efficient to add more fuel, increasing the initial mass, than to try to reduce gravity losses by shortening the time to orbit, using more, stronger engines. So nearly all common launcher systems have low initial TWR - and as result, vertical launch position.

As for wings - wings make a good sense until Mach 2-3 maybe, after that they only add weight and drag. And the rocket must reach Mach 21. Whatever early savings you'd get from winged flight early, they'll be soon turned into massive loss.

  • $\begingroup$ This is interesting considering Falcon 9 first stage has a TWR between 1.28 and 5: space.stackexchange.com/questions/11972/… If you watch Falcon 9 launch compared to Saturn V launch the difference is startling. $\endgroup$ Apr 26, 2020 at 14:31
  • 3
    $\begingroup$ @TerrelShumway 1.28 is still mere 0.28g of acceleration at launch (5 is much later in flight when most of fuel is depleted). While it's still rather brisk as typical big launchers go, TWR of SS-520-4 is almost 7! 2.7m/s^2 versus 58m/s^2. Falcon is definitely deep into the 'underpowered, overweight' category of great most of launchers. $\endgroup$
    – SF.
    Apr 26, 2020 at 14:48
  • $\begingroup$ There was a good answer on the site here about why TWRs are what they are but I can't find it at the moment. $\endgroup$ Apr 26, 2020 at 14:52
  • 1
    $\begingroup$ Here it is: space.stackexchange.com/a/20052/6944 $\endgroup$ Apr 26, 2020 at 15:01

The atmosphere thins out exponentially as you go up. This means that the part of the atmosphere close to the ground is substantially more dense than higher up. In other words, air resistance is much higher when you're close to the ground.

In order to minimize the amount of fuel lost to fighting air resistance, you need to minimize the amount of time you spend in the thickest part of the atmosphere. Any initial tilt angle $\theta$ lengthens the amount of time you spend in the lower atmosphere by approximately a factor of $\frac{1}{\cos\theta}$, just from geometry considerations. As you might expect, the quickest direction to leave the lower atmosphere is straight up.


Besides the good answers you already got, note that for a rocket to launch diagonally you need the vertical acceleration to be more than 1g. In a Saturn V, initial acceleration was barely beyond 1g. If you put your rocket at a 45 degree angle, now you need 1.41g just to not fall to the ground.

There are also huge engineering issues. When you stand a major rocket vertically, it has to support its own weight, and you set it free as soon as it starts ascending. If your rocket is at an angle, you need to deal with supporting the weight of the rocket sideways (compare the strenght of the side of a beer can compared with top-to-bottom, and remember that a rocket is basically a stack of tanks of fuel each with engines at the bottom), plus the rocket would need to slide off its supports until it has enough speed. And, as a reminder, something like the Saturn V weights three thousand tonnes at lift-off.

  • $\begingroup$ I did the maths for this before I'd seen your answer. Saturn V weighed about $29.1\times 10^6\,\mathrm{N}$ and had a thrust of about $35.1\times 10^6\,\mathrm{N}$. For it to leave the ground at an angle $\theta$ you need $\cos\theta \ge 29.1/435.1$ which gives a maximum $\theta$ of about $34^\circ$. $\endgroup$
    – user21103
    Apr 27, 2020 at 11:50

Fuel Economics

When you say resources I will assume you are mostly focused on fuel costs, which is where you would presumably save money by taking advantage of a horizontal airplane-like takeoff. It turns out airplanes and rockets have very different economic operating costs. Lets put some numbers on it. For simplicity we will consider the SpaceX Falcon 9 and the Boeing 777 (in particular a 777-200 which is used for routes like NYC-London).

A 777-200 costs about \$300 million and has a lifetime of around 40,000 cycles. As a rough estimate this means the amortized cost of a flight from NY to London would be on the order of \$7,500. The fuel costs for the same flight will be around \$30,000. From quick googling it seems like the hourly total cost would be roughly \$30,000 / hour of flight, so the whole flight cost would be north of \$100,000. Regardless, it is easy to see that the cost of the aircraft itself is a small part of the total cost of a flight.

A Falcon 9, on the other hand, costs perhaps about \$60 million to build (exact numbers aren't available but most estimates put it in that range). Right now the reusability record sits at 5 times, but lets be a bit generous and say it's possible to be reused 10 times without refurbishment. That gives us a per-flight cost of \$6,000,000. According to SpaceX the fuel cost of a flight runs at \$200,000. We can quickly see that the vehicle cost is a very significant cost of a single flight. SpaceX puts the cost at 0.4% so our estimate of 0.033 appears to be in the ballpark. Even if we were extremely generous and figure it can fly 100 times, that still means the vehicle cost per flight is significantly more than the fuel cost per flight, a complete flip from how airlines operate.


So what would you gain with horizontal airplane-style takeoff? You might be able to reduce that fuel bill by a bit, lets again be generous and say you'd save \$100,000 per flight. What would you lose? As the other answers and comments illustrate: quite a bit. You would have a massive increase in the engineering challenges you need to address (which means more development costs), but more importantly, you are introducing so many more places for things to go wrong. Not only does it still need to perform like a rocket does for the upper atmosphere, you have to have it work like a airplane in the lower atmosphere. It would also need to flawlessly transfer between these flight profiles.

TLDR: It doesn't work because at best you save a little bit in fuel costs but add massive engineering complexity and reliability issues.

Note: the numbers used are rough estimates from quick searching, but I believe they illustrate the economics of the issue. Some sources include:

What is the cost breakdown for a Falcon 9 launch? https://aviation.stackexchange.com/questions/654/whats-the-typical-cost-and-its-breakdown-for-a-long-haul-commercial-flight https://en.wikipedia.org/wiki/Boeing_777 https://aviation.stackexchange.com/questions/2263/what-is-the-lifespan-of-commercial-airframes-in-general https://www.spacex.com/reusability-key-making-human-life-multi-planetary https://thepointsguy.com/guide/cost-of-fueling-an-airliner/


Not the answer you're looking for? Browse other questions tagged or ask your own question.