I want to ask you if is lightning network potentially enough scalable to achieve stable Bitcoin in the future? If we admit that only new members open a new payment channel (and that LN is fully functionable and save) to connect into lightning network (and the amount of channel closing will be minimal) and set the funding transaction, is it enough little to avoid 8GB blocks and mining centralization?
In short: LN won't solve the scaling problem on its own, but it is a very important part of the puzzle.
Today, you can open roughly 100000-200000 channels per day. This would obviously not be enough for a really massive adoption, but is more than enough for current Bitcoin usage "as a currency". There is no reason for not using LN for small transactions up to $100 or so, which make up to about 25% of current BTC on-chain transactions.
For a really massive adoption (millions of new users/day) you need either "big blocks" or an intermediate layer, or both. The "intermediate layer" could be a pegged sidechain or a second Lightning-like layer, like the Channel factory guys are proposing. CFs are multi-user channels, where e.g. 10 users open a LN channel, and all members of the channel can open and close "sub-channels" between individual members with off-chain transactions.
Sidechains are basically independent blockchains pegged to BTC - if there is a reliable mechanism has to be studied.
Even if neither sidechains nor CFs work (I hope that's not the case), a "rough" solution could be simply to use altcoins as alternatives. If we had several relatively price-stable blockchain currencies then this could make up an ecosystem which could have enough capacity for a big part of the world population, and LN could be used for "lightweight" swaps between them.
The sidechain idea as an intermediate layer is however the solution which I consider the best one, at the moment, because it shares most properties with "normal" blockchains, so for example the "offline problem" isn't present. Channel factories could complement them.
Admit that not only financial services will run blockchain technology (include autonomous vehicles, factories, robots, smart homes etc. who all need to put a funding transaction on the blockchain).
I don't see why these things need to be connected to a public blockchain. They can all be connected to a private network with a gateway to the public blockchain. I don't think we need much more than 1 node per household or company.