Do Blockchain Tokens Have Value?

Gregory DiPrisco
10 min readJul 10, 2017
CC BY-SA 3.0 Nick Youngson

Value is an esoteric concept. When it comes to valuing networks it can become especially convoluted due to its traditionally theoretical nature. Yet blockchains must maintain a sufficient market capitalization to secure the value held within their respective networks. This is easy enough to take for granted when prices are soaring and speculation provides a massive disconnect between the security being provided and the actual need for it, but in a speculative contraction there must be a mathematical reason why a blockchain token is actually useful (as others have highlighted).

One of the more popular attempts to value networks is Metcalfe’s Law. This performs well for information networks where the value of connections increases exponentially according to a constant, but poorly in value networks (like blockchains) where connections are valued at varying degrees. For value networks, I believe Beckstrom’s Law may provide a better fit.

Beckstrom’s Law (simplified): V = B-C

If I wish to use my bank to send a wire transfer of $1,000 from Country A to Country B, it will cost $25. If I wish to send $1,000 of Maker Dai from Country A to Country B, the cost will be $.05 (the approximate cost of a value transfer on Ethereum at the time this was written). Plugging this into Beckstrom’s formula:

V = B -C

V = $25-$0.05

V = $24.95 (Benefit to User)

Benefit to Network = $.05

In this instance, value added to a single user by the network is $24.95. Multiply this by 100M transactions/year and the net benefit to users of the network is $2.95B. However, the benefit to the network itself is only $5M.

The other half of Beckstrom’s formula factors security into the cost of using the network. This has been an Achilles’ heel for almost any scenario where the internet is used to store real value.

V = B -C’-SI-L

B = benefit to network

C’ = all costs excluding SI/L

SI = Security Investments

L = Losses (due to security failures)

What is brilliant about blockchains is that, in the context of Beckstrom’s security function, the benefit to the network on each transaction is the security investment. The market capitalization of the network (in conjunction with the consensus algorithm) helps to prevent security failures — thus eliminating the problem of achieving security at scale. This does, however, introduce a new problem.

The value of the network to users is not the value of users to the network and therefore not reflected in the market capitalization.

The utility value of ether is not a direct derivative of the value of the network to users. This creates a problem, considering the value of ether will be indirectly securing the value of the network. $5M, regardless of what multiple you may attach to it or how you pass it through the system, is a rounding error when securing $2.95B.

In simpler protocols, such as Bitcoin, this is not an issue. The only value that can be stored on the protocol must be in the form of the native token. As others have highlighted, if you want to send $50B in remittances over the Bitcoin network, you (at various points in time) need to buy $50B worth of bitcoin.

This does not make Bitcoin a superior protocol, in fact it makes it drastically inferior (for anything besides being ‘digital gold’) as value abstraction is limited to bitcoin. For one, the Bitcoin protocol introduces two immediate points of friction (buying/selling) which cannot be avoided unless the entire planet wishes to transact solely in bitcoin (this doesn’t even take into account Bitcoin’s lack of ‘smart contract’ capabilities).

Regarding utility value, it has been proposed that this is a function of velocity — i.e. the dollar amount going through the network, divided by the number of times it changes hands, discounted to the present over an arbitrary time horizon. Yet the assumption of a blockchain network being at ‘utility’ is that all holders of tokens would rather hold fiat unless they absolutely need to hold tokens. This creates a ‘chicken and egg’ problem for bitcoin as the network can only grow in proportion to the acceptance of bitcoin as a store of value.

Without its ‘store of value’ function, bitcoin has no utility. This is because users of the network will not pay a greater premium to purchase bitcoin than the net benefit they receive from using the network (net benefit as highlighted in Beckstrom’s formula). When using the network strictly for its blockchain, every purchase of bitcoin is met immediately with a sale. If my net benefit of using the Bitcoin network to send money overseas is $24.50, I will not be willing to pay more than $24.49 in premium to use the network. This is a result of my immediate sale costing me the premium that I have paid. For example, if I wish to send $1000 over the Bitcoin network and the current best bid for bitcoin is $1000, I will only buy bitcoin up to $1024.49 because I will otherwise lose the net benefit of using the network in the first place — this is because when I immediately sell my bitcoin to complete the remittance the best bid will still be $1000, netting a loss of $24.49 and bringing my net benefit to $0.01. The implication of this is that Bitcoin’s success must be a self-fulfilling prophecy contingent on belief in its use as a store of value. If not enough people believe that bitcoin is a good store of value, then the market capitalization will not be great enough to remain within the net benefit threshold of users — I.e. You cannot run $50B through bitcoin at any given time unless there is $50B worth of bitcoin to buy without driving the price beyond your net benefit threshold.

This is not so in protocols such as Ethereum, where value can be abstracted away from the native token and stored within the protocol independently. In Ethereum the roles are reversed — ether is not useful because it’s a store of value, but is a store of value because it’s useful. To call on my earlier example, in order to send $50B in remittances over the Ethereum network (in the form of Maker Dai), you only need to buy $5M of ether. Therefore, I propose

Any value stored directly within Ethereum, will require an equivalent or greater amount of ether.

In other words, value cannot be abstracted from nothing, and the utility value of ether is not equivalent to the value of the network but the value in the network.

So how can the value in the network differ from the value of the network?

A tale of two DApps:


Both Maker’s Dai and Digix’s DGX attempt to take value from the ‘real world’ and put it on the blockchain. To highlight the difference — The value in DGX, while represented in the value of the network, is not represented as value in the network. The value of DGX is in fact held in a vault in Singapore, the only value being stored on the blockchain is information about the gold. For the moment, suppose that the utility value of ether in regards to information about value outside of the network (like DGX) is equivalent to the dollar value of the gas fees associated with it. For example, $100M in DGX transactions represents $5M in utility value to the network (ether). This may seem low to secure such valuable information, but as Vitalik so succinctly put it — “A successful attack may cost $50 million, but the process of cleaning up the consequences will not be that much more onerous than the geth/parity consensus failure of 2016.11.25.”

To contrast, the value in Dai is both represented in the value of the network and as value in the network. This is because the value of Dai is held completely on the blockchain and therefore requires ether to abstract. If I want to create $1 in Dai, I will need at least (under current proposed system rules) $1.50 in ether. The utility value of ether in regards to Dai is the dollar value of the ether needed to abstract it plus the dollar value of the gas fees associated with storing the information on the network.

With this current logic, I propose (temporarily) the following to estimate the utility value of ether:

Uv = (Va + Vi) / T

Uv = Utility value of ether

Va = Dollar value abstracted onto the blockchain

Vi = Dollar value of the information on the blockchain (gas fees)

T = Number of tokens in the network

If Dai and DGX both had $1B in assets and generated $5M in gas (and other) fees per year, the utility value of these two DApps to ether would be:

Uv = ($1B + $10M) / 92,000,000

Uv = $10.97

The above assumptions included gas as contributing to the utility value of ether. While this seems to make sense, it is a flawed analysis. Mainly, that at ‘utility’, the supply curve of ether will cancel out the demand coming from the gas fees (information storage costs). Additionally, going forward, miners may choose to be paid in other tokens. While ether is a commodity in the sense of its use, it’s more like a currency in terms of its valuation, i.e., it is simply passed through the system rather than produced/destroyed (gold has a similar profile). At utility, the value of ether will be entirely dependent on the demand curve as the supply will be equal to the gas fees plus the block reward. When ether is in a normal state, you’ll see miners/validators keeping some of their gas fees in order to increase the amount of ETH they hold — This is a reasonable assumption and will probably hold true most of the time, but this does not define ‘utility’ and therefore cannot be assumed. So what drives ether demand? Well, really anything that ‘lives’ on chain. I used abstracted value as an example before, but this was mainly so I could now highlight why the creation of trustless stable-coins will not eliminate the value proposition of ether. If you’re making an Augur bet, registering a name on the ENS, or validating the blockchain with Casper, you’re probably going to need ether — It does not matter if you pay in ETH or Dai, you are locking or abstracting value into the blockchain. However, it does matter if you pay in DGX — because this value is technically stored off-chain and will not shift the demand curve of ether beyond the associated gas fees. If DGX were the preeminent way of storing and transferring value on Ethereum, ether would not have a utility value (this doesn’t mean it is worthless, it would still have velocity, but this is a completely different subject). The one caveat to this thesis is that should another token come to exist on the Ethereum network that has its own on-chain store of value (e.g. bitcoin on Ethereum), this could supplant ether as the more useful token for deposits and value abstraction. I think this is highly unlikely, but stranger things have happened!


Many users will use the blockchain strictly for information purposes (off-chain value, on-chain information). This will increase the value of the network, but not increase the value in the network, and therefore not directly contribute to the utility value of ether. That being said, the larger the value of the network and the greater the marginal utility of the network to the average user, applications that do require ether will grow and become more prevalent. The underlying belief behind this is that on-chain is ultimately better than off-chain. Anything off-chain requires trust, and I am willing to bet that while this is acceptable in the short-term, the long-term will trend towards trustless.


  • Ether is not very valuable if the vast majority of capital flows into tokens that store value off-chain (e.g. DGX). The value of ether in this situation would be the velocity of the gas fees.
  • Ether is not valuable at all if another completely on-chain token becomes the main value abstraction/deposit mechanism (e.g. bitcoin on Ethereum). The value of ether in this situation would be roughly 0 as even gas fees would most likely be paid in this new token, giving it no velocity.
  • Ether is valuable if the vast majority of capital flows into tokens that store value on-chain (e.g. Dai) and ether is used as the primary method of abstracting value/making deposits. The value of ether in this situation would be a function of the amount of ether locked away to abstract value/make deposits and its velocity for both this and the gas fees.
  • My opinion: It will be a healthy mix of the above scenarios, trending towards the latter.

A Final Note: I am posting this to foster conversation. I’m well aware that I may be wrong about any of the assumptions that I made above and I look forward to thoughtful critiques. This post is simply a collection of my personal views. Thank you to everyone who reviewed this and let me bounce these ideas off of them. Special thanks to Matt Rabinowitz for helping me explore the core thesis.

Disclosure: The author is the CEO/Co-Founder of EtherIndex LLC. The views expressed in this post are his alone.