Yes, the launch is the real cost driver for space economics. Launch costs alone are in the neighborhood of $10,000/kg for large American providers and somewhat less in other providers/countries. I believe no one has come close to the $1,000/kg mark yet.
Because launch is so expensive, people then spend a lot more money making sure the payloads are ultra-reliable (launching a replacement is just too expensive). Because costs are so high, there isn't enough business to really get good economies of scale (yet...go SpaceX!) - so a lot of money goes into design and testing of new rockets that will only see 10-100 units built.
As for launch itself, fuel is pretty cheap - turbopumps are expensive. It is the high-precision machining of parts that ends up costing you. If you can reuse the hardware, then you get savings for future launches and everyone is happy.
Because launch is so expensive, people then spend a lot more money making sure the payloads are ultra-reliable (launching a replacement is just too expensive). Because costs are so high, there isn't enough business to really get good economies of scale (yet...go SpaceX!) - so a lot of money goes into design and testing of new rockets that will only see 10-100 units built.
As for launch itself, fuel is pretty cheap - turbopumps are expensive. It is the high-precision machining of parts that ends up costing you. If you can reuse the hardware, then you get savings for future launches and everyone is happy.