I think you're missing the point. These would be Tesla Semis, not diesel semis. Perhaps eventually in a platoon with a single driver in the first rig!
My electric <> diesel efficiency conversion for class 8 semis works out to 2 kWh per mile electric vs. 6 mpg diesel. Tesla has said <2 kWh per mile and I'm using 2 kWh until we have more volume of data. Fleet average for diesel semi's has been 6 mpg from sources that I've looked up.
https://afdc.energy.gov/fuels/fuel_comparison_chart.pdf
This source provides us a comparison between gas and electricity (1 gallon of gas = 33.7 kWh electricity), with other fuels ranged out on that scale. Diesel weights in at 113% of the energy content of 1 gallon of gas - I make that 33.7 * 1.13 = 38 kWh of energy in 1 gallon of diesel.
38 kWh of energy in 1 gallon of diesel moves you 6 miles - let's call that 6 kWh per mile worth of energy in a diesel semi. This accords with my own intuition that electric drivetrain converts roughly 90-100% of energy into motion, and fossil fuel drive trains convert maybe 30% of energy into motion (that's not a precise scientific statement - that's a cross check to confirm we're in the right ballpark with the previous calc).
So electric miles are more efficient by a factor of 3 or so (these are all round number to establish order of magnitude or maybe even factors - not intended to be exactly precise).
So I'm not missing the point - yes, semi's in platoon will be more efficient than semi's not in platoon. Diesel and electric semi's, at least as a physics principle, have equal access to platooning (two trucks can drive close together). The software that does that might not have as quick of a vehicle response for diesel vs. electric, but it's not an inaccessible proposition. Of course, if it's only been written and available on Tesla's semis, then the efficiency gain is only available to Tesla using platoon on the highway.
Let's keep clear about what's being measured and compared. The start of this topic was energy efficiency. The big order efficiency gain is the way electric motors convert energy into motion, vs the way that Diesel engines convert energy into motion. There is approximately 0 energy efficiency gain by removing a driver from a second truck (but heck - that second truck can carry an extra ~200 lbs load!) - it's also obviously highly beneficial to the economics of moving things via truck.
More broadly, there are at least these important metrics that measure different important dynamics.
energy / mile to move stuff
fuel cost / mile to move stuff
cost / mile to move stuff (as measured by driver cost + fuel cost)
complete cost / mile to move stuff (amb driver, fuel, maintenance, etc..)
And realize that when we're comparing cost/mile to move stuff via truck compared to train, the trains (at least in the US) are operating in an environment where they pay their fuel, driver, and vehicle maintenance themselves IN ADDITION TO their infrastructure build and maintenance costs.
Truck companies and trucking pay a small fraction (I haven't found a source that will put a number to it) of their infrastructure build and maintenance costs. The study I did find previously modeled the cost of building and maintaining the road infrastructure as a function of two variables - weather and heavy truck loads (they might have modeled it in terms of axles > XX weight, so an individual truck would count as 5-10 axles).
The point is that all of the personal vehicle use of the road rounded to 0 for the maintenance of the road. That is the effect of road degradation being a function of the 4th power of axle weight.
On the fourth power law, here's a readable blog post that illustrates the idea:
The Fourth Power Rule
And seems to be a pretty good summary of a more complete survey such as this:
http://www.nvfnorden.org/lisalib/getfile.aspx?itemid=261
Where we measure activity / load on a road in ESAL units (equivalent standard axle loads) where the reference heavy vehicle axle of 10 ton load has ESAL10=1 (you need to squint to get a reading from light duty vehicles).