I believe that everyone is already aware of Google's tracking records as one of the most privacy-invading company of all time, given their terrible privacy policy, business choices, and ongoing user tracking. However, Google does have certain products that are pretty good, like Google Photos app is one of its better ones. I can see why people would use it It's quick, packed with incredible features, and makes finding photos simple.
However, it is awful for privacy given Google's track history. Recently, though, I found a upon a fantastic alternative called Ente. With real privacy and it doesn't steal your data, it offers nearly all of the features of Google Photos. It is open source, features easy sharing, end-to-end encryption, and allows you to view, organize, and download your data in its original quality across all of its platforms with consistency.
Unfortunately, if you are someone who frequently takes images and has pictures of years and years worth of stuff, you may have to pay for more storage, although the costs are not too high. To minimize space while using Ente, I would consider copying a bunch of images to an external hard drive or SSD. You can also delete unnecessary old screenshots or photos that you don't want to keep saved in your memories to save up storage.
https://ente.io/
I can't answer your Ente questions but I do recommend using a self-hosting strategy (with or without Ente) if it is within your skillset. The costs are your time and commodity hardware and cloud services. For example, if you just use a cloud service for backups, then the cost of storage will be something similar to Amazon Glacier (I use a different S3-compatible host). $4/TB/month to store, around $100/TB to download. You will presumably only download all of your (hopefully encrypted) photos when you need to test or restore the backups. There are also S3-compatible services that are cheaper than this, but this is a well-known one.
Also, depending on how your photos have been organized, you may benefit from a deduping storage strategy that does incremental backups. If you used Borg or a similar piece of software, you would benefit cost-wise in two ways:
It will dedupe at the beginning of the backup process. If you have 3 copies of the exact same video for whatever reason, it will realize this and only include the data once in the backup.
Incremental backups will sometimes use a multi-file strategy that is efficient for efficient synchronization to other services. Basically, they sometimes store files or groups of files together more or less chronologically. When you then want to send the obfuscated backup data, you can use a sync program that tries to send only the backup files that have changed, which is usually a small minority of files. This is great for when you need to get that data back off the cloud to test backups. Rather than download 1 TB every time you can synchronize the data to a dedicated test volume on your own server where you always keep the last copy (so it can recognize and download only the changes).