Bitcoin ATM industry grew over 2017 year with accelerating speed. Back in 2014 we did a forecast of how many installations of machines there will be in several years. You can find the table at the end of the post on comparison of traditional bank ATM growth vs. bitcoin ATM growth. The forecast was done based on bank ATM installations back in 1970-s. After 7 years we assumed there will be 2900 ATMs installed worldwide. Today (~4.5 years since first bitcoin ATM installation) there are already more than 3000 ATMs installed worldwide. And the speed if installations is only accelerating over previous years:
With growing number of installations, diversity on the market also grows. This is reflected in new manufacturers entering the market, new operators installing machines, machines are installed in countries where there were no bitcoin ATMs previously.
In May 2018 according to information we collected there are:
- More than 3000 bitcoin ATM installed;
- in 67 countries;
- machines in operation are produced by 31 various manufacturers;
- and there are more than 400 operators running bitcoin ATMs.
The following trends were observed during 2017 and tend to continue in 2018.
Prominent growth of large operators
Initially (first year since market inception) bitcoin ATMs were run by amateur operators. Some companies starting doing it as a business, but it was still in proof-of-concept stage. This was the time when small Skyhooks became very popular, but later disappeared due to no support from producer, on one side, but also the fact that those could hardly be used by large companies due to small form-factor and limited support for regulatory requirements they could provide.
In 2017 the large operators grew very quickly to many tens or even hundreds of machines.
The following charts represent Gini index for three major countries: USA, Canada and UK. On X-axis the percentage of operators is illustrated with operators sorted from left to right with increasing number of ATMs in operation. Y-axis shows what percentage of machines belong to corresponding share of operators.
Over years there was increase in number of operators on the US market, however, at the same time the Lorenz curve sags to the right each year. This means larger operators control larger share of the market.
Canadian market had similar to US tendency: larger share of ATMs was concentrated in smaller share of operators each year. Curves in 2017 are practically the same in US and Canada. It is interesting fact, that on both markets the distribution ideally follows Pareto Principle in 2017: 80% of operators run 20% of machines, and 20% of operators run 80% of machines.
United Kingdom is still growing market (22 operators at the end of 2017) compared to Canada (49) and United States (115). This explains that the market is still not that concentrated: 80% of operators control 35% of ATMs, although the trend is pretty similar to the one in the U.S. and Canada. Year 2015 stands apart from the rest dynamics, as it had the most curved line.
Local -> National -> Cross-border Installations
Another trend is inevitable with growth of operator’s network: the geographical area of operation increases as well. In 2017 operators mostly were growing within national borders.
It is hard to measure this effect and show in figures. We calculate distance index for this purpose. For every operator, which had at least 2 bitcoin machines at particular time frame, we calculate the distance between all machines and take average of all such distances. The next step is to calculate the aggregate metrics across all operators, which are represented with two measures:
- Average distance: average across all operators (average of average distances);
- Weighted average distance: similar value as former, but weighted by a number of ATMs each operator has. So large operators’ average distance will have larger effect on the final figure.
As territory size for each country is different, it makes sense to look at each country separately. We analyze the same three countries (USA, Canada, UK) for the period since beginning of 2015 until now.
While average distance between ATMs among operators stayed the same or even fell during the period (~300km), the weighted average distance has a clear increasing trend since ~ April 2016. This can be assumed a start of active expansion of large operators to various new cities. Over roughly 2 years period weighted average distance increased almost 2x times from 500 km to 900 km.
Canada has a difference picture. In the earlier years the average distance was much larger and then reduced. This can be explained that at the beginning the number of operators was relatively low and two companies could have larger effect. In case of Canada BitSent and QuadrigaX were largest operators and both had distributed locations across all Canada, however, later reduced or closed business. And this explains the drop at the end of 2015. Other companies on the market were concentrating locations in close geographic areas (average distance ~100 km). Since June 2017 there is also seen a trend in increase of weighted average distance compared to average value, which is explained by geographic expansion of operators, which started about one year later than in USA.
United Kingdom market stands apart as both average and weighted average distances were reducing over all years. That means operators on opposite were concentrating locations in narrow geographic territory. UK’s bitcoin ATM market is still very concetraed geographically with most of bitcoin machines located in London. This explains the downward trend in distances, as operators installed new machines in London in proximity to previous locations.
Another trend which have not yet started in any significant manner, but we assume could begin in 2018-2019 — expansion of activity cross border and installation of machines in other countries compared to the origin country of operator. Some operators already started to check the market in other countries, e.g. large US operator Athena Bitcoin has installed machines in Colombia and Mexico. Others might follow the lead. But US market is still leading the world bitcoin market, and it depends whether location outside will be profitable enough due to lower demand in general.
Self-manufacturing of ATMs
With impetuous growth of operators they review business model. Previously the standard approach was to purchase hardware from main suppliers, e.g. one of the largest bitcoin ATM producers whose ATMs are widely used in USA is Genesis Coin. Economy of scale allows to start manufacturing own machines by operators.
First example is Coinsource, largest operator in the world. The company has switched from using Genesis Coin machines in 2017 to own developed software used in combination with Genmega Universal kiosks (same ones used by Genesis Coin).
Another large player is Coinme, who developed not only own software, but also launched production of bitcoin machines from scratch. The company has been on the market since mid 2014. Initially running several Robocoin machines, which later were switched to operate based on own software. In 2017 the company has conducted ICO of UpToken, which allowed to attract funds and finance the rapid growth of bitcoin ATMs over the country.
Another example of own production is bitcoin machine produced by existing operator DigitalMint. The machine’s software and hardware doesn’t remind any of existing machines on the market before.
Polish operator WBTCB developed own bitcoin ATM and pursued with larger scale installations compared to previous business size. Previously the company was running Robocoin machines, which were switched to General Bytes software. After developing own software, the company also switched them to in-house developed solution.
There are several other examples, but the trend is clear, operators start to enter on the manufacturing market. Reasons for this are the following:
- Over several years operators built expertise in the area of requirements for software, especially in the area of obeying regulatory requirements. This allows to streamline the transaction process flow for machines in the way they find best fit.
- Cost reduction for large operators. It might be cheaper to produce own machines. The main costs are R&D and later maintenance of software, but the costs are fixed and don’t depend on transaction volume, which allows to better scale business. While it might be not profitable to cover costs for just several machines, it becomes more profitable when you have 100 machines. Running own machines also eliminates producer’s ongoing fees, which is usually %% from the transaction volume.
- Better control over the user experience and maintenance of machines. Running own bitcoin ATMs eliminates one more intermediary in the process (manufacturer/provider). When amendments need to be done to software to improve process, operator is not dependent on third party, and can allocate own resources to fix the problem faster.
However, there are definitely cons of the process as well:
- Initial R&D costs are high, and this mostly fit for large scale and long term business model.
- Additional staff to maintain and further develop own solution.
- Handle more risks on own side, especially security-wise, while control over software stays within company.
- Logistics and capacity with ordering or producing of hardware.
We expect that this trend continues and new operators might switch to own solutions.
Bitcoin ATM industry has matured a lot over last year. The growing price during 2017 brought a lot of new users in cryptocurrency sphere, which was a stimulus for growth of service providers to obtain first bitcoins or other altcoins. Buying it for cash with help of bitcoin ATM is one of the easiest ways. Industry saw rapid increase in installations from top operators, which allowed them to significantly increase networks of machines further.
Another trend mentioned is geo expansion of operators activity, mostly moving from local business from the area where the company originates to nation-wide business represented in many major cities. The next expected step is expansion cross border to other countries.
Finally, some large operators started to develop and switch to in-house developed solutions, which allows to save costs and control the process better, which again allows further smooth expansion.