I built these 3 fundamental valuation models for Bitcoin in Excel. Details in the comment. : CryptoCurrency


This will be part of an ongoing series where I look to value cryptocurrencies according to fundamental drivers. Valuing a blockchain network is difficult, as there aren't established tools like DCF analysis like for equities, however we can build several models and use a consensus range between these models.

First up we have the granddaddy of all crypto: Bitcoin. You will hear a lot of people claim that Bitcoin is in a bubble, but few provide any analysis on what valuation methods satisfy that conclusion. First up some comparative statistics on BTC compared to the other big cryptocurrencies:


Daily Transactions (last 24hrs)350,3761,158,641166,156
Average Transaction Value
Median Transaction Value
Average Transaction Fee

You can find the latest info at https://bitinfocharts.com

Bitcoin is the de-facto "store of value" and the pairing currency for almost all exchanges. With its high transaction prices, slow transaction times and high median transaction values it now functions as more as an interchange settlement layer and "digital gold" than a currency. Currently more than 54% of Bitcoin addresses have a smaller balance than the average fee, making it impossible to move this money out.

I will be using the following 3 models for valuation:

Unfortunately we can't do a DCF analysis as these coins don't produce free cash flow, however we can value them based on cost of acquisition. Most people get Bitcoin by buying them at an exchange like Coinbase, but you could also get some by purchasing a miner and letting it run and mine Bitcoin. This produces a set of cash outflows and in theory there should be a point for a rational actor where they are indifferent between the two acquisition methods.

This method looks at the total cost required to acquire 1BTC through mining and discounts the net value of cash flows at a rate (r), to arrive at a price where an individual would be indifferent between buying one Bitcoin and mining it themselves. The rate (r) would be discounting for opportunity cost of time, and should be much higher than the average market return.

In my model I use the assumption that one is purchasing an Antminer S9, usually considered one of the best miners and is also fairly easy to acquire. It has an average hash rate of around 13,000 Gh/S and uses 1350W, and at current mining difficulty can mine 0.59 BTC per year meaning that acquiring 1 Bitcoin should take around 2 years. I will also assume that the cost of electricity is the US average ($0.12/KwH) and that the discount rate (r) is 4 times the average long term S&P500 return of 8%, set at 32%. This is to compensate for the high risk factor of cryptocurrencies.

Hence the model uses this formula to arrive at an net present cost, which would be the indifference point:

Indifference point = (Hardware Cost)/(1+r)2 + Cost of Electricity (Yr1) + Cost of Electricity (Yr2)/(1+r)

Different assumptions on the discount rate as well as the cost of electricity/hardware will change this valuation.

Method B: The Metcalfe Law Comparitive Valuation

Metcalfe's Law was initially used for telecom networks aund stated that the value of the network is "proportional to the square of the number of connected users of the system (n2)". Its also important that the user is active within the network, as buying a fax machine but never giving out your fax number to anyone doesn't add value to the network. This can be applied to cryptocurrencies as well, as Dr. Ken Alabi from Stony Brook University showed in his paper Digital blockchain networks appear to be following Metcalfe's Law. I will derive the Metcalfe valuation ratio (P/n2) and compare it to the other "Big 3" cryptocurrencies. This is somewhat similar to P/B ratio in equity analysis in that a higher ratio implies investors expect the network to create more value from each individual user in the future than another with a smaller ratio. One of the difficult things is determining the number of users in a crypto network. The vast majority of addresses have no activity, and most have a balance that is either zero or too small to cover the average transfer fee.

I will use two different metrics, the peak 1Y active users and total daily transactions as per BitInfoCharts.com:


Daily Transactions (last 24hrs)
Active Addresses (Peak 1Yr)
Price/Metcalfe Ratio (Active Addresses)
Price/Metcalfe Ratio (Transactions)

This gives us a relative valuation ratio compared to ETH and LTC.

While the Net Present Cost method uses the mining assumptions for an individual with a single miner, most Bitcoins are mined in large scale operations with much lower mining costs. This model aims to derive the average cost to mine a Bitcoin on the global network. We can then think of the valuation of Bitcoin as some multiple of this.

The long term cost of mining depends in large part of electricity price, and this model uses the information from Blockchain.info to create a sensitivity analysis based on the price per Kwh. Now since most Bitcoins are mined in low electricity priced countries (primarily China), the lower end of the spectrum should be used.

In my model after deriving the cost of mining across various electricity prices, I applied a generous multiple of 3x (300% profit over mining cost) to derive the following two valuations:

At Chinese average electricity prices ($0.08/kwH): $5,091.54 per Bitcoin

At US average electricity prices ($0.12/kwH): $7,385.81 per Bitcoin


To sum it up:

  • The Metcalfe Law gives us a relative measure to compare against other cryptocurrencies and depending on whether we use the number of active addresses or number of transactions we get different ratios, but both show that BTC is overvalued in comparison to its nearest competitors (ETH and LTC).
  • Using the Net Present Cost method and a discount rate of 32%, we get a valuation of $7,269.
  • Using the Multiple of Average Mining Cost and applying a 3x multiple, we get between $5,091 and $7,385 for the valuation.

These models all have their limitations and the value depends on assumptions made about multiples and discount rates. In addition there is the inherent problem of correlation for mining costs and price, with miners responding to price increases and increasing difficulty. However the difficulty increases at a much lower rate than the price. Nevertheless these models give us a way to think about Bitcoin prices in more objective terms rather than simply going off emotion or sentiment.

In my personal view any price over $10K is overvalued, and I expect to see a correction downward in the following months now that we have future contracts, although Bitcoin has never behaved rationally and it could pump some more. I think its fair to say that Bitcoin today is radically different to how many early adopters imagined it would be once it gained mainstream attention the way it has this year. Back in 2013 I envisioned that bitcoin would be widely accepted among eCommerce sites by now, but this sad chart from BI shows how far away from that reality we are and how its getting worse. Currently its transactional function is not as a widely adopted digital currency but as a settlement layer between various crypto exchanges, and it underwrites the entire crypto market because of its pairing status. Its future will largely depend on how well it maintains this status.

I also plan to do this for several other cryptocurrencies, to do a fundamental quantitative analysis with one or more models. For other currencies I will also talk a lot more about the team and technology and market segments.

Disclaimer: I first purchased BTC back in 2013 and liquidated my position after it hit 15K. I do not currently hold a position in Bitcoin, although I do hold several alts that are paired to Bitcoin.