The economics of storage

I just wrote an important piece on the economics of storing things — and I mentioned MaidSAFE there (I know, it’s just called SAFE now). Here is the article:

Towards the end of the article you can see a link to this Web2 solution for micropayments, essentially we don’t need Web3 to enable it:

(there is no Qbix token yet. This is just about the economics and technology for how to achieve a sustainable and open system while compensating people for creating digital content.)

Discussion welcome! I wanted to share it here and get your feedback on how SAFE handles it. You can reply on here or on there — regardless, I just want to have the conversation, because is fundamental to long term sustainability and power dynamics in the whole SAFE ecosystem.


Is this the tl;dr ?

Clients pay to browse as opposed to pay to store.

1 Like

@GregMagarshak ,

Good to see you here again! - will read your post . .


1 Like

Yes it is. The rest is comparing the societal effects and the second link is for how to achieve it.

What are the economics of safecoin around this at the moment?

1 Like

Yeah, I’m curious as to why you jump to this conclusion:

In decentralized systems, the people who want to access the data in the moment should be the ones paying.

We are taking a different approach, where all GETs are free, but uploading data is charged for once at upload time by the Network. All data stored is perpetually available.

Could this inverted approach perhaps stem from the design being led from the starting objective of digital abundance, and universal access, as opposed to artificial scarcity and private ownership?


Can you explain that, Jim? I thought the network no longer held the tokens.

Privacy. Security. Freedom

1 Like

Currently those ideas are in flux


I’m not entirely clear on what you are asking here, but what I mean is the user pays the Network as a whole, not an individual farmer. The Network sets the price, decides the location of data chunks, and rewards farmer etc. this is quite different from other models, such as filecoin etc.


Jim, I explain in the article how we arrive at the conclusion. It’s spelled out there. have you read it?

“We are taking a different approach, where all GETs are free, but uploading data is charged for once at upload time by the Network. All data stored is perpetually available.”

I go into why this is a problem, at length. The whole piece is about this.

Could this inverted approach perhaps stem from the design being led from the starting objective of digital abundance, and universal access, as opposed to artificial scarcity and private ownership?

When you mention “this inverted approach” I am not sure if you are referring to readers paying, or publishers being the ones to pay. But basically, you cannot assume anything about the amount of popularity something will have in the future. If you want digital abundance and universal access, then you need the readers to pay when they access.

I sure have! I was quoting from it in my response to highlight this differences in our approaches.

Yeah I agree, that’s why we think perpetual storage is important.

  1. Because humans aren’t very good at predicting the value of things in the future.
  2. It is a prerequisite for individual data sovereignty

I think this is where we disagree… because what you are describing in not universal access. It’s access restricted to those with the means to pay. It’s one giant paywall. And not something that we feel should be put in front of a public knowledge commons.


I like this metaphor for pay to get.

Another image that would apply is from @happybeing in another context :


It’s only one giant paywall if you don’t already have the data. If you do, then you can always just pay “yourself” to store it. Storing it pays off if you access it later.

Look, the bottom line is that storage isn’t free or unlimited. Storing something means you can’t store something else. Any scarce resource needs a way to indicate how badly you want it, versus someone else.

Paywalls are only bad because they’re annoying. They make you pay for an entire NYTimes subscription just to read one article. Making a micropayment to read the article is appropriate. Anyone can afford a micropayment.

That’s the thing… for personal consumption, practically anyone can afford things. If you’re trying to do something in bulk, likely you’re running some business so you should get enough coins to sustainably use it.

Same as water… if it’s totally free then why can’t someone just abuse the commons and just refill a giant pool every day during a drought?

For what it’s worth, Freenet takes this approach: resources that have been accessed least recently are dropped when the network reaches storage capacity. But in Freenet, accessing doesn’t cost anything. So people can just pay someone to keep accessing resources and keep them around. It’s a mismatch, and unsustainable if people abuse it. By having market competition for accessing things, you get a correct signal for how needed something is, at every hour.

What would abusing the commons of free access look like though? The analogy of breathing air seems more appropriate, what would the world be like if we had to pay to breathe?

It may well be unlimited, we don’t know.

They are bad as access to data is needed for humanity to progress. Think of Ramanajan (however you spell that) and what he did with access to a few books?

Marketing, they want you to subscribe to them. Micropayments allow you to shop around. Companies want stickie approaches to customer retention. What better than be their only subscription (or one of few).

Micropayments can be bad for companies customer retention (another bad thing, but a reality right now).

This is not true, folk are dying for lack of food etc. I believe lack of information only harms folk.

Come to Scotland and see :slight_smile:

They do not have perpetual data and that is a significant difference to Safe.

Good to see lots of opinions and sometimes to look 180 degrees at problems and solutions, so thanks for that.


Amazing chap, thanks:


It doesn’t really matter how much it costs to access data, it is still without question, a barrier to access. And by extension it deepens inequality.

A paywall that takes pocket change to open for someone in wealthy country may be no more than an inconvenience but that same pocket change, in another country, suddenly becomes a couple of hours work, or a days wage, or a week’s salary.

That information may be worth paying for. It might make life just a little bit easier. For the wealthy individual that pocket change suddenly made working a little bit more efficient and thus made their money go further, and their wealth increase.

For the poor individual that money needs to go on housing, or shoes, or baby food. The information behind that paywall seems like it could be important, or helpful, but it’s still out of reach.

So the gap between the these two individuals, in knowledge and in wealth, just expanded.

This is why information must to be free. That should be our starting point. And together, collectively, we should find solutions to achieve that aim.


Jim: Perhaps I’m not understanding, but how do you avoid people abusing the free resource? Even undervalued resources like Polygon’s blockchain have been easily overrun (just last month using the Sunflower game for instance). Blockchains have such limited resources, of course, that the fees skyrocket as soon as any single app becomes even mildly popular. But with SAFE, you can’t just assume that storage is limitless. Freenet doesn’t. You need a mechanism to manage scarce resources - and that is the market.

In your system, all data will be stored perpetuity? Why would someone store that? Because they are getting compensated for it (farming) like with BTT or Filecoin? Proof of spacetime? So who compensates them?

@dirvine - I will take you up on your invitation to come to Scotland and see :)) Provided you can show me around! I have never been and love the landscape and the people.

1 Like

The uploader pays them.

1 Like

If a node stops storing data it no longer earns from accepting new data because it will be kicked out. So a node has to store data in perpetuity if it wants to earn. Any nodes which go offline, will quickly be replaced by a new node that will store the same data (replicated from sibling nodes).

1 Like

I don’t think this is correct Jim, because with pay once for eternal storage we are asking farmers to predict the value for all time!! I don’t think this is a game-stopper though … it’s a tough problem to estimate, but this is where markets shine - as they collectively figure these things out.

@GregMagarshak There is a ‘market’ here and while it’s still harder to estimate for all time as opposed to how much does it cost for, say a week, a month, a year etc.; it’s still doable. Safe Network has many safeguards to prevent people from breaking the network and prices can and will go up and down over time as the market figures out a good estimate of the the costs of perpetual data … importantly though if the price of data storage for some reason goes super high in the future, then the cost of storing data on the network will also go much higher. This doesn’t break things in the same way as people who bought groceries 50 years ago got a much better deal than those who are buying them today (because of inflation).

If the network effect of the Safe Network grows as we believe it will, then the value of the network token will also grow enormously as well - more so than any other crypto probably as network effect is everything. And as the cost of storage on the network is paid with the token, for those utilizing the network and building on it, the costs should stay quite reasonable indefinitely.

Freenet, Tor hidden services, ZeroNet, & others have other considerations that prevent them from becoming strong contenders - IMO, mainly speed. Hence making a direct comparison isn’t helpful. Additionally, while there are costs involved in storing and sharing data on those networks, the costs are not public and data itself is hosted and stored directly by the users - hence there is no real market for storage cost on those networks.