There is a steady increasing number of bitcoin ATMs on the map. However, many people complain that fees to use these machines are still high, although these complaints might not be that prominent as they were at the start of first bitcoin ATMs launch times. Users finally realize that running a bitcoin ATM is not that easy and there are many costs associated with it.
Bitcoin ATM fees were already analysed in our blog back in 2014. This article is supposed to look at what are the latest figures across the industry as of the end 2016.
There are two ways of collecting fee size on the site:
First is based on the manually defined values. This is normally used for bitcoin ATMs, which don’t support passing online information over API. Most prominent example is Lamassu with currently 185 installed and operational bitcoin ATMs across the world.
Another way to collect data is via online feeds. There are several bitcoin ATM providers, which support online reporting of information similar to the one used at actual machines. These are: Genesis Coin, General Bytes, and Bitaccess. In case of Genesis Coin and General Bytes machines, online price is taken and converted to fee size based on Bitcoin Average rate. This allows to make all fee size comparable. In case of Bitaccess fees are reported directly.
Since April 2015 we collect daily information about fee size. There were some algorithm adjustments to calculation and storing procedure over time, but generally fees are comparable historically.
Current worldwide average bitcoin ATM fee is 8.4% for buying bitcoins from machines, and 5.4% for selling bitcoins for cash.
Bitcoin ATM industry is growing fast. There are several new bitcoin machine installations worldwide daily. Existing bitcoin ATM operators increase number of machines in their networks, as well as new operators join the market. Even on the production side, there are new manufacturers developing their products and coming to the market with new bitcoin machine models targeting different segments and niches, e.g. recently were added BitTeller, Bitlox, BitConcept on our site.
Running a bitcoin ATM is an interesting business opportunity in the new era of digital currencies, however, sometimes the lack of information keeps people away as they don’t know where to start, or get any practical estimations of how much you can earn with a bitcoin ATM. In order to increase transparency Coin ATM Radar conducted a survey with a number of bitcoin ATM operators to find out what are current metrics of an average bitcoin ATM business.
Now back to survey. It was sent to 24 operators in total, 11 responded either in full or skipping some questions due to not wanting sharing info, although it was mentioned that none of info will be associated in any form to the company or person responding. Even limited info we received gives a good estimation of level of costs and earnings one should expect.
Lately many bitcoin ATM users complained that they feed cash into machine, but don’t receive bitcoins to scanned address, or they sent bitcoins to a machine, but can’t withdraw cash. This post is supposed to check what are the possible reasons of this and how to prevent it or solve it when it happens.
Prehistory – Bitcoin network congestion and block size limit
Since Bitcoin launch in 2009 the network has been growing steadily with respect to number of transactions. There was a 1 MB limit set on block size by Satoshi Nakamoto to prevent spam/DDOS attacks on the network in the early days and this limit was left there untouched as there was a lot of space before this limit could be reached. Meanwhile, there were many hot discussions about what to do when we reach it. Unfortunately, measures were not taken on time, and today we still have this artificial limit set in the Bitcoin Core client, which is the most widely used by nodes and miners. So we reached the point when we are literally hitting the ceiling of the network capacity.
International money transfer market – best fit for Bitcoin
Remittance market is a huge transfer of money happening every day. According to World Bank report the estimated remittances size in 2015 was forecast to be $601 billion. International transfers were one of the best-fitting envisioned applications of bitcoin, as it is state border-less by definition and cheap to send money irrespective of amounts sent, while only fees per transaction size in bytes matter. However, it is still a hard way to get Bitcoin adapted at a massive scale. Early stage of its development makes it an instrument for tech advanced people only.
With introduction of new bitcoin services we are moving closer to the case when bitcoin is used daily by many households, bitcoin ATM is one of such services. Initial idea of bitcoin machines is to exchange cash to bitcoins or other cyrptocurrencies and vice versa, bitcoin to cash. However, there are other use-cases for bitcoin ATMs evolved. One of them is international remittance.
Bitcoin ATMs are spreading quite fast since the end of 2015 and beginning of 2016. Although the most useful application of bitcoin machines could be in the developing countries where financial inclusion is low, the majority of bitcoin ATMs are still appearing in developed countries.
Quick look at bitcoin ATM aggregated stats shows that 63% of all bitcoin machines are in North America and 25% are in Europe. Taking into account 5% in Oceania (mostly Australia) means almost 95% of all machines are in developed countries. There are several reasons for this. Costs for bitcoin ATMs might be quite high for average businessmen in developing country with respect to risk of operating and running it. Also return on investment required can be achieved only in locations with relatively large user base and steady bitcoin usage (mostly in developed countries as of now), which results in transaction volume for such machines.
Leading country within the developed countries list with respect to bitcoin ATM adoption is the United States. As of the end of March 2016 it has roughly 280 bitcoin ATMs installed across the country, which comprises ~45% of all installations worldwide.
Bitcoin ATM industry matures and it is now more than 2 years passed after the first bitcoin ATM installation. Let’s look at what were the most prominent events and trends in bitcoin ATM sphere during 2015 year.
Although there was a slower growth compared to the previous 2014 year it was still a significant increase in number of bitcoin ATM installations. Number of bitcoin ATMs worldwide grew from 330 to 536 bitcoin machines.
The speed of bitcoin ATM installations increased at the very end of the year (jump on the chart in December 2015). Important that new installaitons number was approximately similar to 2014 year, however, in 2015 we see a closure of more machines as well as more reshuffling happens (new locations for old machines). For comparison, in 2014 there were 43 locations delisted from the site, while in 2015 the number of closed locations increased to 183. Continue reading →
It’s been two years since the first bitcoin ATM went live in Vancouver, Canada on 29th October 2013. So what has been achieved over this time and what are the current dynamics of this newly evolving market?
As it is seen on the following chart number of bitcoin ATMs is steadily increasing. However, the speed is lower compared to the first year. During the first year there were 284 bitcoin ATMs installed and in the second year only 187 new bitcoin ATM locations were added and not removed later. This is lower than we forecasted a year ago based on the growth rate of traditional bank ATMs in the beginning of 70-s. It was calculated that there will be about 562 bitcoin ATMs installed by this time, however, there are only 471 in the world.
Cottonwood Vending operator added its sixth location in New York area. Now satoshi1 one-way bitcoin machine was installed in Harlem, NY at Apollo Deli & Fruit. This ATM allows quickly buying bitcoins. Online prices for this ATM are available at given above link, machine charges on average 8-12% fee above market rate. The ATM has additional bulk note acceptor, which means it can easily handle huge packs of paper money and convert it into real money of future in seconds.
As mentioned before – we have a functionality to show online prices for Bitcoin ATMs in case there is an API provided by the operator. At the moment many ATMs already display online information from the provider (all Genesis Coin ATMs, most Robocoin machines, some others). If you are running a Bitcoin ATM and would like your online rates to be displayed – please provide us with an API link.
While several operators asked for the template or API description, we put it here available publicly. In any case the set up is done manually with transformation to our standard data structure, so the list below contains just recommended examples, but not necessarily the requirement.
We gathered a lot of information about the Bitcoin ATMs for our Bitcoin ATM Map thanks to many contributors. Now it is time to pay back to the community. For many Bitcoin machines information about the fees was stored. However, it was the information that operators usually don’t want to share. Recently we added integration of API from many providers which allowed us to show online rates for many Bitcoin ATMs. Using the data from both sources, it became possible to answer the question “How much do the Bitcoin ATM operators actually charge their customers?”
For detection of the fees for every machine we used the following algorithm (in descending priority order):
If the data about online rates for two-way machine is available – calculate the fees based on the spread (buy and sell fees are averaged in this case);
If the data about online rates of one-way machine is available – calculate the fees based on BitcoinAverage feed in corresponding currency;
Finally, if there is no online rates information, but the data about fees was collected before – we use it.
Based on this logic we collected data on 127 Bitcoin ATMs out of total 230 machines (55%). While information about the fees was collected not only from the providers, but also from independent sources like reddit and bitcointalk, it is likely that the data is not biased and may be assumed as a representative sample.