Let's look at the attributes Elon has described publicly, and see how they compare to a launch loop:
* could be cheaper than CA high-speed rail? Check (no right-of-way).
* "theoretically fastest way" from A-B? Check.
* can't crash? Check (w/ well-designed rails).
* immune to weather? Check.
* twice as fast as a plane? Check.
* lower energy cost than a car? Check.
* self-powering with solar panels? Check.
* stores power w/o batteries? Check.
* requires aerodynamic expertise? Check.
* "ground-based Concorde – as fast as a Concorde, but on the ground"? Arguably. All weight is supported by the ground, and cars take a Concorde-like path.
* rails are not needed? BZZT!! However, he may be talking about the "right-of-way" aspect of rails here.
* leaves right when you arrive? Check (w/ individual pods).
* isn't a vac tunnel? Check.
Of all the ideas that have been kicked around, the launch loop seems to fit best. Elon would certainly have been exposed to the idea when brainstorming for SpaceX.
My only question for the past few weeks has been, "How do you build the thing?" It doesn't inertially support itself in the air until the rotor is spun up, and the rotor can only be spun up when it's fully constructed between the endpoints. Catch-22!
There are basically three ways to do it, each with their own challenges.
Build it in the air. You would need huge numbers of lifting balloons (presumably of zero-pressure solar Montgolfier design, ala ARCA), and the portions of the system in the troposphere would be incredibly susceptible to adverse weather while under construction. You would want to have the entire thing already built, then unreel it as fast as possible. Smart tethering would reduce this risk, and the whole thing would need emergency descent parachutes for operation anyway. This seems like the least unworkable scheme, imho.
Build it in the water. Ocean currents and tides are now your problem instead of wind. The requirement to survive the corrosive salt water environment puts additional constraints on material selection. Oh, and you have to shut down all boat traffic from LA to SF. And your base station either has to be mobile (to start out in the ocean) or you have to start with a big arc out from the coastline and "tip it up" vertically. Do the math on how big the rotor endpoints would be (given reasonable assumptions about bending magnets), and you quickly realize that building them in place is the only option.
Build it over land. This seems the most unworkable – all the risk of having it fall (or worse, lose containment), and now you have to build tall towers every few thousand feet and string a high-tension support cable between them. Imagine explaining to homeowners that the "rail" we're building above their house will have parts moving at 10 km/s inside before it lifts off. Ouch.
Whatever method is used to get it up in the air, it'll need to be safely reversible. Over water bringing it down is much easier. And you'll have to bring it down for repairs. Every five minutes it will do a full loop.. Iron is fairly ductile, but those stresses will add up. And wikipedia says the kinetic energy involved is close to a small nuclear bomb. Back of the envelope math suggests the iron cylinder would be very thin. It will wear out sooner or later.
My guess is that Musk's big breakthrough is for a cheap and reversible deployment system.
I'm double checking my numbers now (and reading the original papers) but the entire contraption might weight a lot less than we are implicitly assuming. Like, light enough that a single large plane/blimp could support the entire cable/sheath system. And a small fleet of planes could safely tow the cable through the sky. Or a fleet of blimps could support the entire cable while it was stopped for maintenance. (Both would need a large number of well trained pilots, it would make the skycrane maneuver look simple.)
edit: The original paper talks about a ribbon 5cm wide and 7.6mm thick weighing 15.6 Gg. I dropped a kilo in the back of the envelope calculations. So it would need at least 100 very large aircraft to hold up. Still feasible, but more than a little crazy.
Hydrogen is another possible lifting gas, with all its associated foibles. Obviously lifting bags would have to be isolated to prevent catastrophic failure, but you could get a lot more lift out of them.
This must take place in a La Niña year, when the jetstream is north. Reel out the track, drag it into place with two airships, join the parts and start 'er up. You could even name the airships Jupiter and Number 119. ;)
This simplified strategy that would avoid the need for aerial stations at both ends, intead using a gentle curve achieved by anchor cables and varying linear density. Acceleration would begin immediately upon departure, leaching energy from the rotor (generating eddy currents), then contributing it back by braking off the rail on the descending side.
Of course, maybe I'm just dreaming here…
Working on the principle that the speeds and efficiency mentioned means it has to avoid the friction of travelling through a medium. Since he had said "not a vac tunnel" I set my mind considering how the problem could be solved without relying on a vacuum. I figured, 'why not move the medium as well?'. Making a tube train travelling in a fluid medium would be interesting. Make the inner surface of the tube a hydrophobic surface. Use solar power along the length of the tube to accelerate the liquid. The speed would increase until the energy lost through friction with the sides reached equilibrium of the solar acceleration. If you can reduce the surface friction sufficiently, extremely high speeds should be possible.
Comparing against the checklist you did, a Hydrophobic torus full of high speed liquid gets a fair number of hits. The same aspect applies of energy storage as kinetic. It's a better match for 'ground based' but The same ground based aspect means it would get in the way. Building the liquid filled thing would be much easier, possibly implementable at smaller scales initially.
Mostly I think this just goes to show the interesting ideas you can come up with when given a set of constraints and a goal. I seem to vaguely recall a Mike Abrash anecdote about VGA fifos that applies here too.
I think I'd be a bit disappointed if it is a Lofstrom loop(but that's what it likely is) unless there is some significant innovation extending the idea. It's just like a vac tunnel with cool inertial skyhooks.