One of the discussions I am currently having with my international colleagues is about the global trend towards urbanisation and the resulting shift of political, economic and financial powers from centralised states and federal structures to mega-city or mega-urban region centres. Some of my American colleagues expressed the fear that this would further marginalise rural communication.

My comment was, first of all, that this urbanisation trend is happening anyway, so in general people seem to prefer to live in cities; and I also think that in general this mega-city movement is a positive development as more political, economic and financial power would move towards the cities away from the often impotent central system. This fits in a far better way with developments in the digital/sharing/interconnected economy, smart cities, micro-grids, renewables and so on, as it will bring power closer to the people. In general local politics are far less polarised and there is a far greater chance of national interest developments (public transport, fibre, e-health, etc) being developed in a more efficient and effective way from a grassroots level upwards.

At the same time, I indicated that there is no need to worry too much about the negative consequences for rural communities because the changeover to a more locally-empowered structure will take many decades and within such a timeframe there is plenty of time to ensure that the rural population is included as well.

We already see more and more farmers wanting to be connected to fibre networks as they increasingly need this for many aspects of running their businesses. We can also link these hubs into the urbanised super structures.

My comments were followed up by my Dutch colleague, Hendrik Rood, published below in full —

* * *

We are observing in the Netherlands that farmers now create co-ops willing to pony up a substantive investment of several thousand Euros for getting connected.

The keyword is Precision Farming. Some farmers today fly with drones above their field photographing the growing plants with 1 centimeter resolution. 10 minutes of drone flying provides ca. 4 GB of data (they photograph in multiple colors, blue, green, red and two shades of infrared). This is done primarily on cloudy days and to complement satellite remote sensing pictures. Satellite remote sensing now photographs with hundreds of colors and that data is collected by specialised IT firms.

So what happens is that farmers upload their drone data to the cloud applications with satellite remote sensing data for joint processing. To get this running, you do need fiber. A big wireless microwave link might suffice, but it tends to be more costly to deploy.

In 2014 I interviewed Jacob van den Borne who, with his brother, manages 1359 acres of potato-farmland—partially owned by his family, other areas leased or serviced as an agricultural contractor. Discovery Channel with Vodafone made a documentary about him.

Based on his data volumes per acre, I did a calculation on the actual data volumes that would have to be uploaded on a monthly basis when the entire Dutch agriculture sector engages in deploying these yield-improving technologies.

It turns out that the total video/photography volume per month surpasses the entire raw video output of the production side of the Dutch Television industry if all arable land farmers deploy. That total volume is a bit larger than the production of AV-material in raw formats (HDTV raw at 3 Gbit/s and studio quality processed files at 120 Mbit/s etc) which are used and exchanged in the AV-processing and editing efforts before actual broadcasting of the compressed files from playout systems.

So we have a sector of industry that is moving towards modern IT-intensive data collection methods and will surpass the industry considered to be the most data-volume-intensive IT consumer.

The economic driving force is that when some farmers start to improve their yields (expectations in the short term are yield improvements of 15%, longer-term improvements could reach 50% more yield per acre) the other farmers will have to adopt the technologies, or they will go out of business in a few years.

Dutch agriculture is intensive both in land use as well as in capital use. In terms of international exports, the Dutch agriculture and its accompanying food industry is the second exporter in the world, behind the USA, but before France. It is responsible for ca. 60% of the Dutch export surplus, which as a percentage of GDP is higher than Germany’s export surplus and China’s export surplus.

The key point when we presented the interview result to our Ministry of Economic Affairs (which recently has merged with the Ministry of Agriculture) was that we got a surprised response. They thought that flying drones above fields was ‘something for America’. I had to make the remark that, as the second agricultural produce exporter in the world, these farmers compete with American farmers on efficiency, yield and costs, before the point started to sink that this is the future of modern farming.

One Scoutbox releases up to 300 pictures of insects a day in a 10 hectare greenhouse resulting in several GB of data on a daily basis. They are sent to central cloud-based IT platforms for processing and analysis.Now you might think this is just one part of agriculture. But we recently bumped into the next sector: greenhouses. They now have the Scoutbox.

This is an insect catcher they deploy inside a greenhouse. On a daily basis they automatically catch insects and make high resolution pictures to determine potential illnesses and negative impacts. One Scoutbox releases up to 300 pictures a day in a 10 hectare greenhouse. So it produces several GB of data on a daily basis. They too are sent to central cloud-based IT platforms for processing and analyzing.

As a rule of thumb, you don’t want a steady dataflow of 120 MB per hour to drip through a very small data uplink. A steady flow of slightly less than 300 kbit/s. There is too much risk of dataflow breakdowns. For that volume you need something like a 10 Mbit/s upstream bandwidth to be able to go around with the Scoutbox and press ‘send’ at the end of the day. A 100 Mbit/s upstream requires ca. 10 seconds to send a 120 MB dataset and 4 minutes for filing the results of an entire day in the greenhouse.

A greenhouse farmer has a capital intensity that would allow them to afford machines like the Scoutbox (sold at ca. €5000 a piece). The real issue is the cost of getting the high bandwidth link into the greenhouse/farm.

The Dutch greenhouse sector is the largest in terms of agricultural revenues (before potatoes and tulip bulbs). Ten years ago 25% of the world’s greenhouse area was situated in the Netherlands. In recent years it has declined a bit, due to the fact that cultivation of roses has moved massively toward East-Africa (Kenya, Tanzania, Uganda and Ethiopia). Daily shipments are flown from Entebbe Airport to Amsterdam—Aalsmeer Auction where they are redistributed over the entire world.

About a year ago KKR bought part of the shares of a Dutch roses cultivator, AfriFlor, in Ethiopia, which had set up shop there only a few years before for the rather surprising amount of €100 mln for a slight majority of the shares. It turned out this company had €80 mln in revenues and €17 mln in annual profits. AfriFlor was the largest of many Dutch farmers who went to Africa. But these farmers are not ‘small’ in any sense.

Many of them relocated due to buyout pressure in the 1990s to expand Dutch towns and cities, as well as nature reserves and parks for human-residential reasons. A lot of younger Dutch farmers were bought out and relocated to Canada, the USA, South-America and Australia, Eastern Europe and Africa. The buyouts and the decline of available arable land in the Netherlands drove up the land prices.

So many of these emigrating Dutch farmers moved with a purse containing a few million euros/dollars from these ‘high intensity’ areas to much larger areas. They brought their herds with them but as frozen Artificial Insemination specimens, or seeds and advanced machinery and know how.

As far as I understand it, similar ‘moves’ came from farmers from Denmark, France, Britain etc. Initially in the 1990s a lot of these farmers went into Eastern Europe, but many of them then moved to other continents and brought techniques with them that frequently startled the agricultural communities that were formed in the 19th century and the emigration waves after the second World War.

A lot of West-European farmers have been constrained by land shortages as well as environmental regulations and went into ‘intensive agriculture’ first, learning these advanced techniques. Released from constraints, but with all the knowledge-intensive techniques, their farms tend to explode in size in the new areas outside Europe compared to what one sees in Europe, where intensive land-use farming dominates.

These agricultural businesses are ultimately driven by arable field or greenhouse yields. And improving yields with 10%-50% by deploying some IT that costs a few thousand per device (whether it is a drone or a scoutbox) to monitor a set of parcels or a few greenhouses isn’t a burden. Harvesters etc. are far more expensive machines.

My estimated guess is that as soon as you find a way to get the ‘per farm’ cost of fiber down below €10k CAPEX per farm, you may start to have a business case for many farmers to ramp up data-intensive technologies. It is why we see these successful co-ops now appearing in rural areas.

The issue was they couldn’t relate the expense to actual business profits/improvements before. As soon as one starts to explain the yield impacts, and show the real examples of competing farmers deploying the technologies, it starts to make sense for them to consider the investment.

Additional notes

We recently did deployment cost calculations in various Dutch rural areas. It turns out that in the more expensive areas, average cost per farm (farmers live or have their home next to their farm here) is somewhere in the range of €6000 – €10000.

The areas where costs are on average €3000 – €4000 per address see massive take-up, often in a combination of a one-time install fee (ca. €2500) for the link, or a monthly fixed premium of €10 – €15 per home/farm connected above broadband prices within town limits (where cable-TV, vectoring VDSL or FttH-service is available). Take-up rates go beyond 60% in these areas.

There is still a bit of a hurdle when you have to move beyond €5000 CAPEX per farm, but a few years ago, a one-off fee of €2000 was seen as an insurmountable hurdle for getting it sold—until several co-ops showed that that was no longer the case—but one had to conclude the projects were sold in a wrong way. Way too much talk about e-Health, e-Education, smart this and smart that etc., but geared toward what policy wonks and governments have on their agenda. Not with respect to the genuine demands of people working and living in rural areas in their actual business.

I don’t know what is to be considered the actual cost structure of FttH in rural USA, but when I look at farm density in states like Iowa, it isn’t much different from what I see in the more sparsely inhabited rural agriculture polders in the Netherlands (dairy farming tends to happen on smaller area farms). One also has to grasp that in the Netherlands cables are dug into the soil and aren’t hung on utility poles, as they were removed in the 1970s. So CAPEX levels tend to be a bit higher than in countries who hang communication cables on poles.

Cost-saving techniques like cable-plowing (well known to farmers who are familiar with putting drainage tubes into their land) can bring trenching cost for fiber optic cables down to ca. €7 pro meter. However, not everywhere in the Netherlands it is easy to deploy cable-plows (too many trees surrounding the roads in some areas, in other areas the road borders are too small or weak), but I do think those techniques are not too difficult to deploy in quite a few rural US areas when poles aren’t available.