And shipping cargo to Mars doesn't need to be very expensive if you are patient - you send the cargo towards a low-perihelion trajectory and deploy a solar sail closest to the Sun. It may take a good couple years for the shipment to get to Mars, but the delta-V on the way up will be free. If you send the shipments well ahead of the landing crews, they'll have lots of toys when they arrive.
But I agree with the SpaceX folks. Nuclear is the way to go if we are to even go back to the Moon.
That, and the living quarters. If it's going to take a couple months to put stuff together and test, you don't want to be living out of an Apollo-sized capsule.
It's not possible to launch a 3720 ton ship to Mars from the surface in one shot, it has to be assembled in space from parts and fuel shipped up in multiple launches.
But now that we are having to have this talk, personally I think multiple ships assembled are still the way to go. A one-shot XX launch means not enough weight for proper radiation shielding, and enough food and spare parts to make it all the way there is a problem, and enough space to prevent the astronauts from killing each other is also an issue. We still have to have enough fuel to get there and land even if there is a bunch of fuel on site waiting in pods both in orbit and on the surface. The Dragon capsule is not really big enough to prevent mission failure from mass homicide en route, something considerable larger, and larger than the XX can lift in one go with fuel and food and water is needed to avoid this.
The space assembly step doesn't have to be something done in a giant station over a period of months. It can be send 4 modules up one after another, dock them together and go. We assembled two space ships together en route to Luna using this method on the Apollo mission. Adding a few more modules isn't a big deal, if we could assemble (not build from scratch) stuff in space 40 years ago, we can do it again now.
http://news.discovery.com/space/will-a-commercial-flight-be-...