Let's talk economics

As a long time follower of the project, the recent months have been absolutely fantastic with the release of the SDK and the constant stream of new versions and features, but I feel like one important part has been missing from the discussion, namely economics.

I know that dfinity sees itself more as a decentralised cloud provider than a blockchain, but a core characteristic of the network will be that the payment for compute power on this network will happen through some token. Anyone who is going to build a serieus application on dfinity will want to know the cost structures involved. But at this moment there is almost no information around this aspect of dfinity. Here are some questions that come to mind:

  • With regards to storage, will you have to pay some sort of rent or will it be a one time fee. I almost can’t imagine the second option giving ethereums growing state problem and if it is the first options what happens to your data if you forget to pay the “rent”.

  • What will be different costs of write and read operations on the network. What would be the cost if I would want to host a completely static website.

  • Do people have to pay gas to make a transaction on the network like with ethereum?

  • Will a future release of motoko contain functions for moving around and programming with DFN? So far we have seen nothing around programming with tokens, which might be one of the most interesting usecase in the beginning as something which you could only do with dfinity as opposed to AWS.

I also think there are other important questions which are still unanswered like are there any regulations dfinity will be considering in the structure of the network like CCPA or GDPR or what if somebody does something illegal on the network?

I’m loving the releases of technical content that seem to be ramping up lately but the technical side isn’t the only important factor in making dfinity a succes. Most code is written in service of a business and a serious business will want to have a clear picture of the costs and legal implications involved when choosing dfinity as their cloud provider. I’m very curious what will be the answers to these questions in te coming period.


Very good questions, @Fulco. I´m looking forward to reading the answers :slightly_smiling_face:


Nice thoughts, these are some places to start start with for anyone interested:

1 Like

yeah Yulin Liu’s medium blog gives quite a bit of insight… all of his posts have the disclaimer that his writings do not express the interest of DFINITY, but I would imagine a good starting point.
As for specific costs to run a website… that information has not yet been released but yes I agree would be awesome info to have.
One point I remember is that Dominic had
said in a podcast that the UI will be developed in such a way that the user doesn’t even realize or have to figure out gas fees. I guess it will be directly
Into the system with 0 headaches and having to calculate gas costs from the developer.
As a final thought, I remember talk about developing PHI stable coin which would act as a central bank and pay all validator rewards. Interested in hearing further on the progression of that and whether that is still being developed given the regulatory hurdles in the decentralized stablecoin markets.

Great conversations to have and great topic

1 Like

Hi, do you have an ecosystem grants program for now or the future?

1 Like

Hi kako, we do! Take a look at the Beacon Fund and other ecosystem support options here: https://dfinity.org/ecosystem

1 Like

Hi all,

the thread is quite old already, but I guess it is the right plance for my thoughts.

I watched the whole Sodium launch event, but still there are questions remaining concerning the token economics.

My central question is:
What i the “exchange rate” for Nodes of burning Cycles in computation, an in turn beeing rewarded ICP Tokens

Here is how I got to this question.
What I understand is the following (please correct me, if I got it wrong):

The supply of ICP is not limited. “Initially ~500 Million ICP are created”

To run canisters and use the IC, Cycles are used and burned in the process.
Cycles are kind of a stablecoin for CHF “The price of a trillion Cycles is always the ICP value of one CHF”

The usage of the IC will accordingly get cheaper over time for two reasons:

  1. CHF is subject to inflation, thus the “value” used to create 1 TCycles decreases.
  2. As Dominic said, computation will use less Cycles when the hardware in the Nodes get stronger, thus more computation for less Cycles/CHF.

Supply of ICP Tokens is increased by two means:

  1. ICP Tokens locked in Neurons are rewarded with more ICP Tokens (“limited to a maximum amount of ICP Tokens per Month, which is devided over the Neurons”)
  2. Nodes are rewarded ICP Tokens in exchange for computation (Cycles get burned, ICP Tokes are produced)

Supply of ICP Tokens is decreased by transforming ICP Tokens to Cycles.

For the use of the IC the price of ICP Tokens is irrelevant, as computation is basically paid for in CHF, with a detour, one CHF worth of ICP -> TCycles.

The incentive to run Nodes is the reward of ICP Tokens

If the inflationary forces (Neurons and Nodes) are stronger than deflationary force (ICP to Cycles transformation) the ICP TOken would be subject to inflation.

This in turn would lower the value of ICP in the long run and thus reduce the incentive to run nodes, or lock ICP in neurons. Eventually endangering the whole system.

Coming thus back to the central question:
What i the “exchange rate” for Nodes of burning Cycles in computation, an in turn beeing rewarded ICP Tokens.

Thanks for your thoughts on this topic!

1 Like

Wait, is this true? My understanding is that nodes burn cycles in exchange for fiat (CHF), which is their preferred form of payment.

1 Like

If yout look at “A Technical Overview of the Internet Computer [version with slides]” on youtube Minute 2:30, Dominic says:
“The NNS produces new ICP tokens, to reward Nodes that are run by data-centers…”

1 Like

Ah, I see, this wasn’t covered in the latest Token Economics post. I’m also curious what the answer is.


Yulin replied on twitter:


Computation/cycle is a parameter determined by NNS which is a conflict of interest because ICP will rise if it increases franc/cycle, cycle/compute, or NNS/data_center reward ratio. There is no decentralized stable value here. We just have to trust the NNS which has no mechanism for enabling cycle users to govern how much they have to pay for cycles. ICP holders are not going to sell out unless they have extracted all the value they can by raising these values. There needs to be a check and balance system. I’ve worked on this problem for years. Meter started using 2-coin like this to determine stable value in 2018. I declined to work on that project because of this. The best solution for stable value is for fees-paid for byte computation & transmission to be the governing mechanism of the stable value. I do not know how difficult that will be to implement.

Assuming total quantity of ICP is now fixed, Dfinity should target & publicly state the reward ICP holders will get as a fraction of what they are paying out to data centers. This will make it transparent if ICP hodlers start acting like Facebook & increasingly & unfairly start taxing the system by voting to push up the claimed off-chain ICP/franc ratio or increase franc/cycle, cycles_spent/compute_expense, and/or ICP_reward/data_center_reward ratio. By advertising & committing to a ratio it allows buyers of ICP to be truly investing in the internet computer’s bandwidth & not just pushing up dev, VC, & founder hodlings without a concurrent increase in the bandwidth.

If the francs are changed to the cost of avg world electrical joules (9% deflation after 50 years in U.S. verses the franc’s 285% inflation although they target 2%/yr now if the SNB is trusted) and change “cycles/operation” to joules/operation (basically ask the ICP governance system to keep track of Moore’s law instead of having fixed values for gas/operation as in ETH) then the cycles coin (joules) would be the best possible stable value without fiat. The bytes/second of the network computing power would be equal the joules (which is not just electrical cost, but the CAPEX in terms of “joules” of the equipment) used by the network times the velocity of the currency. Bytes/second is a power in Watts via for both computation & communication which matches up with MV = PQ since M is in joules and V = 1/seconds.

A 2-coin system is only useful for reimbursing VCs & founders & confusing ourselves as to what’s actually going on. Instead of cycle users getting locked into the system with cheap cycles at the beginning, only to see the “tax” rise, the cycles spent should govern the ratios and choose a fixed NNS/datacenter reward ratio to give the VCs a fair reward. Nicely Dominic said VCs were not his choice, but a SEC requirement. Instead of users having to trust VC’s, founders, and other hodlers, it should be VCs and founders trusting users.

The SEC was wise to block Dfinity’s faucet out of fear it could be a freeware trap. If Dfinity governance was decentralized, they could have ignored the SEC and done it anyway & the SEC is fine with that because decentralized ICO’s do not result in a freeware trap like Facebook (and Dfninity) that has stock hodlers governing and wanting to maximize their profit at the expense of users.

This problem is why I am buying up Dfinity. I want to be in the 1% controlling how much everyone pays for their internet computer bytes.

1 Like