You can also use mobile structure architecture to make a cargo vessel. I honestly don't remember how well they performed but I think they were still not very fuel efficient compared to real life vessels. They are also capped at cruising at 2 (or 3?).
You're capped to 4 MP for naval mobile structures. And their fuel efficiency is utterly ridiculously good.
Emulating a Template E (>30.000t) Tech Level C large naval support vessel, you'd get a 6-level, 9-hex mobile structure. Now setting both at 4 MP with ICE engines:
- the large naval support vessel (depending on size) requires between 10,800.5 t at the lower end and 36,000 t at the upper end for the engine.
- the mobile structure requires ICE engines of 648 tons for the power system.
Note that at lower speeds the margin is less, since for the naval vessel it scales virtually exponentially.
One day of travel at full speed (=23.3 knots), i.e. 1036.8 km, costs the large naval support vessel 1119.8 to 3732.5 tons fuel; the same for the mobile structure would be 134.4 tons.
For the mobile structure this is in line and rather favourable in comparison with current ships - which burn around 140-150t per day at that high speed.
However for cargo, mobile structures are rather bad:
- The large naval support vessel comes with a deadweight of between 13,334.5 t (for a 30,000.5t vessel) and 44,450 t (for a 100,000t vessel)
- The mobile structure comes at maximum (CF150 fortress) with a deadweight of 7357.5 t
The remaining deadweight of a mobile structure with fuel loaded is
lower than that of a naval vessel as long as the desired range is less than 6288 km / 3395 nm (in comparison to 30,000.5t -> 6,542t remain) to 10688 km / 5771 nm (in comparison to 100,000t -> 5,971t remain).
P.S.: A naval mobile structure of that size is physically capped to a maximum mass of somewhere around 127,500 tons due to its water displacement, and for that it'd need a relatively odd shape filling all hexes.