Any fully decentralized crypto at the scale of use that the Web Monetization API would need would have enormous tx prices. There are ways to scale this, ie the lightning network, but those are essentially centralized solutions to scale.
There is however no technical reason we should limit the blocksize to 1mb. We could have 10mb or even 100mb blocksizes easily. Realistically a 100mb block would be large enough to handle all transaction data our species currently generates.
The transaction fee is a considerable portion of miner's revenue. Miners ultimately are responsibility for making changes to the bitcoin protocol. I think it's unlikely bitcoin miners will vote for a higher blocksize because it will cause short term decreases in revenue. However, they are missing a potential boon from the Jevons paradox -- the cheaper a resource becomes to use the more of the resource we use.
So in summary, it's not really a technical limit to have a high transaction volume but we aren't likely to see it from the current big coins.
Edit: The problem is not 100mb is too small for the future. The problem is that just like we cannot go from 1mb to 100mb today, we would have no ability to go from 100mb to 1000mb when needed.
Otherwise it would've been long superseded by other chains.
Bitcoiners still arguing about the solidity of the network guaranteed by so much mining, yet virtually all the mining is in the hands of few specialized operations all knowing each other (so it's not really decentralized) and despite Ethereum proving new protocol to be a valid and safe alternative.
None of Bitcoin cultists at the end of the day understood that it wasn't about algorithms nor computers, but consensus among people.
And that consensus voted to have Bitcoin, the oldest and least technologically developed chain to be the "store of value" of cryptocurrencies.
Nonsense.
With the current block size and average transaction size, the bitcoin network processes ~7 transactions per second. A 100X increase in block size gets you to 700 transactions per second. A quick google says global credit card transactions currently average more than 21,000 per second.
You're still almost two orders of magnitude off, and that's just existing credit card transactions.
Other networks were designed for speed and others are working on it (etherum with layer 2 networks, and more speed focused networks, for instance as was mentioned Nano was doing 1200-1500tps years ago, with plans for increases not sure if thtey eer went further.
The problem is crypto has a few solid real projects and a billion loud useless scam/spam projects that make the idea of "crypto" look like its all trash, finding the networks that actually have something to really give to the world is hard.
Some nodes are more popular than others for open a channel to, but there are still several thousands, so calling it centralized is misleading.
The reason adoption has been so low is that with current hardware you cannot run a full node in a phone, so you either get a server to run a full node or you trade-off some control/security of your coins
There are still other theoretical problems, mostly things like denial-of-services, but adoption has been what has stagnated it.
The hardware/control tradeoff is a real PITA, we need a phone with like 50x more battery or efficient to be able to run lightning in a completely self-custodial manner
And there are still some people advocating for blocksize increase, which will make the problems of lightning even worse
