Perhaps the serve every route mentality is the wrong one for success. Should it not be serve every profitable route. There are very few ultra long haul routes in the real world. Roughly under 50 routes flown above 7000nm direct, so not popular even without tech stops. Not even that many over 6000nm. Perhaps they are not so successful in AWS for the same reasons in the real world.
The primary reason for that is that the real world treats pax differently to AWS. The only way for an AWS pax to get from Melbourne to Amsterdam is to fly it, with a tech stop. A RL passenger will be much more likely to fly it as 2 separate legs, with some amount of time spent wandering around Dubai, Singapore or wherever. The demand is still there IRL, there are just other options beyond the direct flight.
The demand system is set up based on real life, on there being 300 people wanting to get from Melbourne-Amsterdam daily, or whatever the 2013 ingame number is. But the pax algorithm will only produce ~120 people willing to fly that route, because it requires a tech-stop. But since it requires some sort of stop IRL too, that shouldn't happen. 300 RL people want to fly the route with a stop, then 300 virtual people should want to do the same.
The biggest issue with this is that the demand graph is essentially off by 50% or more. If I look at a route that says 400 daily pax, and I stick a 300 seat plane, most appropriate to that route, on it, then when my RI and CI are high, competition is non-existent, I should get close to a full plane. But I don't. In fact, I get half the stated demand, or even less.
The problem isn't whether a ULH flight should be profitable, whether default pricing should be different (that's an entirely separate argument), the problem is whether an appropriate plane/s on a given route should be able to get roughly 100% of displayed demand when RI is 100, CI is 80+, and pricing is at the default. And right now, that doesn't happen. Players shouldn't be expected to look at the displayed demand, listed as 95% accurate, and instead treat it as being more than 100% optimistic compared to the true demand. Tech stops should only get penalised if there is a competing, direct flight.
At least until some future point where there is a much more complicated algorithm, and a virtual pax wanting to get from MLB-AMS has the same sort of options as RL, a 'direct', tech-stopping flight, or a connection in SYD, LHR, DXB, etc.