We're gonna use astronomer math. We're only looking at orders of magnitude.
There are maybe 100 million bitcoin owners. There are soon going to be 10 billion people, increasing at an exponential pace. If we want 10% of people to trustlessly self-custody (this is a very low bar) we need to scale 100x.
An average tx on mempool right now is 37 cents. That's pretty cheap, actually. Imagine if the cheapest transactions ever got was $37.
That's 10% of people in 10 years. Extrapolate to 100% in 50 years (population doubles every 47 years), you need another 100x. That's $3700 for a one input, two output transaction. Used car money. You need to buy a new used car every time you interact on-chain. Wanna do a coinjoin? You're gonna need house money.
This assumes no new use cases for on-chain space. This could not be more wrong.
Quantum computers are coming. Post quantum signature schemes A) will require a soft fork and B) are also very large, so let's divide transaction throughput by 10-100x.
On-chain stablecoin transactions are just getting started. That's a huge market, multiple times larger than the lightning network today.
Ark service providers will be ticking out transactions at a steady clip. These are time-sensitive transactions so they will pay the prevailing fees to get on chain.
BitVM is building an entire computer using logic gates encoded in bitcoin script. Those will be very large scripts.
ZKPs are extremely promising and extremely large. We can't keep people from storing them on chain. Bitcoin is permissionless.
Don't forget about ordinals! Did you think they were just gonna go away? Lol, you silly goose. There will still be degens using up your precious blockspace.
Bitcoin simply does not scale well enough to serve the whole world. We need better tools on the base layer or else we are just LARPing about freedom technology.