Ten years ago, processors were.. 3 GHz. And today, processors are.. 3 GHz.

Wait.

Let me restart.

The necessary storage/transfer will be mostly linear scaling due to the limits on the number of transactions.

So today, a full node takes maybe 300 GB. Ignoring the first few years, we can estimate 300 GB 5 years. In twenty years, all transactions will be encompassed within < 2TB of storage. Think how fast memory/HDD develops. You have 2 TB of HDD now, easily accessible. And that's the limit within 20 years. That's not really that much.

Yes, it might take a few weeks to download, but bandwidth will increase too.

Also, keep in mind that many people don't use a full bitcoin node to interact with the chain.

IPC increases.
Bitcoin at its current form cannot handle transactions on-chain that is sufficient to support mass adoption. We can use second layer but I doubt we won't increase the onchain capacity through some optimization. Currently, the blockchain size growth is roughly ~100GB per year. It's decent but you have to also take into account that most people won't buy an extra HDD for Bitcoin and for the near future, most consumer computers won't include that much storage because there isn't a huge market for it.

It might not seem like a lot but for the average Joe, having something like this hog up their HDD is a far greater problem for them.

10 years ago 1TB SSD were insanely expensive. Today they are under $100 USD
Bandwidth price is falling in most places. 5G is allowing very high speeds for downloads even when you do not have the ability to connect to a local provider.
RAM & CPU pricing is falling too.

So there it's no big deal.
If you want to run a node on the PC that you just pulled out of the trash that is one thing, but if you are going to spend some money then it will be fine.

Same as now, you can run a node on a slow internet connection on a RPi3 with a slow drive attached to it.

Or you can get a 4th gen core i5 with 16gb ram and a 512GB SSD for under $250 and not worry about it.
You can probably save $40 going for a similar generation AMD based PC.

10 or 20 years from now the systems will be different but theory will be the samw.

-Dave

I've changed my two statements of ROM to HDD now.

In 20 years size of hard drives will be multiple times bigger and size of bitcoin blockchain will be trivial, and download speed will probably be much faster so people won't need to wait for months to download full node, or they can just download pruned node.
There is also possibility of some code updates and changes that would make it easier for everyone to download and run full Bitcoin nodes.

Current bandwidth is more than sufficient to download the entire blockchain within a few hours at max. Problem is that there exist a ton of bottlenecks which can't be eliminated that easily. I don't think there is a particularly huge issue with an increase in the rate of transactions in the future; assuming transactions can take place off-chain and we can sustain an increase in the efficiency of the various components. I think Moore's law isn't that relevant anymore. We're already reaching 7nm nodes.

There are more concerns within the system other than your bandwidth and your HDD.
Reducing block size is counterintuitive. We need bigger block size to support LN, even if we were to scale to our closest payment system competitor. It's quite evident that on-chain capacity isn't ideal right now, outright blocksize increase cannot be the only solution.

The size of the Bitcoin blockchain is increasing by linear progression (about 100GB every two years), while technological advancements in the IT industry continue to grow exponentially. For example, a decade ago, most personal computers had disks drives with a capacity of 320GB or 500GB. Nowadays, most new PCs come with drives that hold 2TB or even 4TB. In fact, storage capacity is doubling every few years and data transfer speeds are increasing even more rapidly.


https://www.statista.com/statistics/647523/worldwide-bitcoin-blockchain-size/

So, unless the rate at which new blocks are generated dramatically changes, there is no concern that blockchain data will surpass the limits of technological advancement in the foreseeable future.

This is spot on, and I believe this will actually be an issue in the future. There's often comparisons between storage options today, and storage options 20 years ago. However, that's because old computers were limited even for the general consumer. We are now probably reaching a point where bigger storage is possible, and could potentially become cheaper over time. However, for it to get cheaper you would need mass production of it. If you general consumer doesn't need 14TB+, then the production will only be for those that can actually use it, most likely commercial, and a few hobbyists. So, we could actually see things start to stagnate  for general consumption, simply because we won't need the higher storage values. In fact, we've probably already seen this happening, with the introduction of chromebooks, which are primarily used for browsing the internet, and with commonly used software now taking a online cloud storage point of view, i.e microsoft office there might be no need to install programs yourself in the future, reducing the need for local storage, at least for the general consumer.

In the future, if you were to run a node you might actually need equipment, which gets rarer, and rarer as time progresses, simply because the industry might cater to the general consumer, which could take a more commercially hosted application approach.

Most users don't require a txindex. Chainstate is about 7-8GB right now(?), certainly less than 10GB. Block data grows way faster than chainstate, the UTXO set is constantly having entries added and removed but the block data grows linearly.