Another statement you cannot back up with facts.
A minimum size-multifarm with willow trees (which require a grand total of one breeding cycle and no expensive treealyzer, something you can finish long before you even start to make steel) will produce roughly 110 buckets of ethanol per hour (source, data describes two-hour test runs). Let's assume you have one fermenter and two stills, which won't be under full load for this setup. At most 15+1 MJ/t for the fermenter, and each still takes what, 5+1? I forget, but the values show ingame if you run 1.6.x. The multifarm takes less than 1 MJ/t and has no power perdition. Let's eyeball it at ca. 25 MJ/t for the entire production stage over the full runtime, which might be overestimating it, depending on whether or not I got the stills wrong.
At 8,000 HU per bucket, a 36 LP boiler will consume one bucket of ethanol every 50.5 seconds (source). Meaning for every hour it needs between 71 and 72 buckets. Conveniently the setup described above produces 110 buckets, taking about 25 MJ/t to do so. This leaves a positive energy output of 47 MJ/t - roughly two thirds of the boiler's output remains available for use, without factoring in the ca. 52% ethanol overproduction that you could be using to run another 12 LP boiler. Which could be powering the production stage pretty much by itself, leaving the entire 36 LP output as net gain.
Using HP boilers changes nothing about this, it simply requires twice as much fuel for twice as much output in the same timeframe.
Additionally you can increase a fermenter's output by 50% without making it require more energy by using fruit juice. And if you use a higher sappiness trait - since the willow is only in the middle of the pack - you also get increased output for zero extra energy cost (and it's even multiplicative with the fruit juice bonus).
So your plan requires us to first stockpile all the Ethanol it would take to heat the 36 HP boiler to start, so we need something like 3 max size steel tanks right out the gate, then we apparently have to breed trees now so we need to spend some time doing that and finding a swamp, o and we'll need bee's to pollinate it too great. Now your willow tree here gives .750 buckets of biomass per tree (have no fruit juice nor will we invest even more power to get it or this becomes even worse) so for every say 10 trees we still were getting 2.2ish buckets of Ethanol. We need 144 buckets to power a 36 HP. So that's 4 Stills 4x6MJ, 1 Fermentor 1x15MJ, Water Pump for the Fermentor 2MJ (No TE ATM), so 41MJ so far. Now comes the sapling problem, you claim a 'minimum size' multifarm can supply this Fermentor and I have no idea how because I have 2 of the largest sized ones running and the fermenter still doesn't have enough saplings coming in to keep the 4 stills running all the time, (all my original calculations were done with MFR farms, planters, standard trees) because these are going to cause another problem instead which we'll get into next. So we got I guess 2MJ from the 2 farms, 2 water pumps 4MJ there, 6MJ + 41MJ were up to 47MJ. Now comes the that problem which is Fertilizer since were using 2 Forestry Multifarms and a Fermentor, and were going thru alot all combined so we need a MFR Composter operating at nearly max to keep up with the 2 farms and the fermenter so there 50MJ, 47+50MJ were up to 97MJ now, you omitted this tiny energy cost because your system in not a closed loop; you require manually finding fertilizer. So doing it your way, yes we do barely produce a small MJ surplus. So small I couldn't to even bother constructing it and waiting the several days for 4 stills to produce the Ethanol to fill the initial steel tanks you'd need to heat the boiler in the first place. When you factor in the mj wasted with heating costs the small MJ surplus still would never give you positive energy unless you played for years maybe.