~
Even assuming all that, then at most you can open 9,000 channels per block, meaning it would take 17 years just to let everyone in the world open a single channel. As soon as you consider that obviously some people need to have multiple channels open for Lightning to work, and obviously people will want to close channels, open new ones, top up their channels, and so on, then that number increases exponentially.
Lightning is great, but it cannot support global adoption without further changes to the base layer.
Let's assume the improvements Carlton Banks mentioned, let's assume that not everybody in the world will use LN, more like, 100 million at maximum and probably half of that to be realistic and when we're considering this we arrived at the point when we might have to ask ourselves if LN won't be solving a problem that's no longer there and will come to hunt us more likely in a distant future!
4 years since the last spike in price and the activation of segwit and despite periods of high fees look where we are now, it's the middle of the week, blocks are coming in at
90% of the pace , 44 blocks behind in 3 days, and the mempool is empty, less than an hour ago a 700kb block was mined, even if we have been experienced 1-3 sat/vbyte days for almost a month now, with the difficulty finally not taking those damn swings usage is at the lowest. Of course, the usage might grow again if we're entering in a bull run but those that rush coins then won't be using LN, until the thought of making 10x with our coins in a year won't fade away I don't see the demand for transactions growing that much, especially with a lot of users opting for centralized solutions, debit cards, internal wallets off-chain transfers, and many more.
So, I guess we have quite a lot of time to seek improvements, and if those problems arise when a 12TB ssd costs 200$ and nobody produces drives with capacity counted in GB, I can see how this will be solved.
