Elon Musk at some point tweeted about the idea of an "air rocket" for the cars. But air is too light to give a lot of momentum. What would really work well for

Elon Musk at some point tweeted about the idea of an "air rocket" for the cars. But air is too light to give a lot of momentum. What would really work well for acceleration is a water rocket with an electric pump. Water mist is harmless beyond a short distance, as long as it's properly dispersed. To make a 1000 kg car accelerate at 4 gee or 40 m/s^2, you need 40000 N of force. A water rocket engine with 200 m/s exhaust velocity can deliver that with 200 kg/s of mass flow. Since your target velocity is only 28 m/s, you only need reaction mass for 0.7 seconds, or 140 kg or 140 L of water. That's one feed barrel. The back of the envelope kinetic energy of the water jet is only 3 million joules or less than a kilowatt hour. But the power is 4 megawatts. Pumps also have startup acceleration cavitation issues, so it'd need to start gradually etc. One could maybe do it with a lot of small independent systems in parallel: a ultra high C rate lithium iron battery connected via a transistor to a pump. Each has its own water pouch and nozzle. If you can achieve 400 N of thrust with one pumping 2 kg/s, then having 100 of them in a 10x10 grid would work. The transistors would be controlled centrally. A more boring alternative is, like in a real bottle rocket, also here, a compressed air tank half filled with water, a much more straightforward source of pressure energy.