The big thing with a big truck is torque. It takes monster huge amounts of torque to get a load moving. Torque is also very important when going up hills.
Horsepower is how fast you hit the wall. Torque is how far the wall moves after you hit it.
A series hybrid big rig would work, as long as the electric end has the torque to get up to 80 kilopounds moving from a standstill, and to keep it moving at least as good as a regular drivetrain up a long hill. It'd also need the battery power to keep it moving up a long hill. The smaller engine would still end up running flat out to keep putting power into the battery.
Fortunately, most long climbs are paired with long runs down hill, often as steep and long, so regenerative mode combined with the engine and generator to charge the battery up for the next climb.
What could be fairly easily done with today's technology is programming the drivetrain with a 3D route map to control the engine for the most efficient operation over every hill. Assume hill one is a 2,400 foot climb, a descent down to about 2,000 feet, followed immediately by a long slog to over 4,000 feet, followed by a descent down to about 3,000 feet. (A rough guesstimate of the run up Whitebird south of Grangeville, Idaho.)
Up AND down the first hump you'd want the engine going full throttle to ensure the battery is topped up so it wouldn't run out of juice partway up the big mountain. Once at the top of Whitebird, the engine could idle because the Camas Prairie is pretty darn flat, lots of downhill, most of the way to Lewiston, which is the lowest point in Idaho.
3D route map/control using GPS and repeated route automatic tuning could optimize fuel use. Live measurement data could be copied from one truck to another to speed up the learning/optimization process. That'd work best with identical trucks. With cellphone internet connections, the data could be streamed live and downloaded by other trucks following so their performance could be optimized in real-time for economy optimization when it's the first time a truck has taken the route.
How about a service selling route optimization data to owner/operators and fleet operators? Could be a sort of brokerage where truckers could upload their data and those who download it pay if they get better MPG than a previous run.