Internet for all through infrastructure sharing

I delivered a TEDx talk in Budapest on May 5, while the video will take a few weeks, for now I wanted to share the written script and slides

– – – –

Thousands of years ago many different tribes lived along the Yellow River in China. Many of them depended on the river for survival and prosperity, but all of them also suffered from periodical floods and draughts. No tribe could do anything because each of them controlled just one section of the river. Eventually all these tribes came together to form the Chinese nation which got an opportunity to control the entire length of the Yellow River. This coming together of tribes allowed the Chinese nation to bring thousands of people together to build canals and dams in order to regulate the river and prevent the worst floods or draughts.

Today, in the words of my favorite historian, “we are all living alongside the same cyber river, and no single nation can regulate this river by itself”. While I fully agree with Yuval Harari that overcoming present day challenges like climate change will require global coordination, I want to emphasize that the “we” he speaks of, only includes half of us, because even now half the human population cannot afford to access the internet. How can we achieve global coordination when 4 billion people cannot get online? I don’t think that’s possible.

I want to make internet access affordable for all because I think everyone should have the ability to ask a question and get a straight answer. You might think this is a nice, utopian but very expensive idea. I think otherwise. However, before going any further into why I think my goal is achievable without any major infrastructure spending program or technological breakthrough for that matter, I want us to be clear on what the internet is….

In an age of instantaneous data transfers and magical handheld devices, it’s easy to forget that the internet is physical. The internet is made up of fiber optic cables, data centers, massive co location centers like 60 Hudson Street in New York, servers, routers, cell phone towers, portions of the electromagnetic spectrum, and virtually any physical medium that allows data to be stored and transferred from one point to another using the TCP/IP protocol.

Most of this internet infrastructure is owned or leased by private enterprises. When you run a google search on your smartphone, you are accessing this infrastructure, and at month end when you make a payment to your mobile network operator or internet service provider you are indirectly paying a toll or fee required to access this internet infrastructure. Such tolls and fees help mobile network operators and internet infrastructure owners cover their own costs and earn a financial return as well.

The earliest known internet infrastructure was paid for by the US Government to make powerful computers more accessible to American researcher. Back then, accessing the internet was free for researchers, and expansion of internet infrastructure was centrally coordinated. Then in 1995 the US Govt owned infrastructure got privatized. In the same year the GSMA got established, which stands today as the largest global trade body of mobile network operators. The GSMA helped in creating standards which enabled the launch of SMS, Data and Fax services over wireless networks.

This intersection of internet infrastructure privatization and launch of wireless services worldwide created a gold rush of sorts because selling connectivity services without laying down cables or wires seemed like an attractive idea to investors in telecom companies. This also meant that many telecom companies offering virtually identical services would end up building excess and overlapping infrastructure to win over customers.

For example, when mobile network operators started building out their respective wireless networks around the world, they made a conscious decision not to share cell towers because in the early days of network roll outs, coverage was viewed as a competitive advantage. A mobile network operator with presence in 60 cities was a more promising story than one with presence in 50 cities. Thus, mobile network operators often constructed cell towers right next to each other and this choice led to a scenario which was equivalent to a country building 4 to 5 overlapping highways that could not connect with one another.

Today, most mobile network operators no longer view coverage as a competitive advantage because a lot of the wireless infrastructure needed to provide coverage has already been built, in fact in many areas it has been over built. According to the International Telecommunication Union, 95% of the world population is touched in some way or form by a basic cellular network. Yet, we live in a situation where half the world population cannot get online largely because they cannot afford to pay the tolls and fees needed to access this information highway.

The silver lining though is that many of these overbuilt highways I speak of are under utilized, meaning capacity clearly exists to accommodate a lot more traffic. However, accommodating more traffic will require these information highways to be shared and connected more effectively so that key choke points can be released.

To facilitate sharing and remove choke points, infrastructure sharing companies have emerged. An infrastructure sharing company typically buys and leases back a piece of infrastructure like a cell tower to enable sharing. Three mobile network operators sharing one cell tower results in lower operating and maintenance costs for all compared to a situation when one mobile network company uses one tower all by itself. Lowering of operating costs makes it more feasible to serve customers who want to get online but cannot afford to do so.

At present 62% of total cell phone towers around the world are owned by infrastructure sharing companies, and while that number looks encouraging do bear in mind that cell towers are just one piece of internet infrastructure. To highlight this point further, even today most mobile network operators are not sharing the electronic equipment which gets installed on cell towers.

While electronic equipment on cell towers in not being shared at present, it is continuing to become a lot more energy efficient and smaller in size as well. Just like mainframe computers morphed into laptops and tablets, the same trend is playing out in the world of wireless communication equipment. So while the cell towers or the power capacity attached to those towers is staying the same size, the equipment which gets installed on them is becoming a lot smaller and energy efficient.

Even using the older larger telecom equipment as a proxy, I can tell you that most cell towers have the capacity to accommodate twice as much electronic equipment as they do today.

The larger implication is that by renting out a little bit of space on cell towers which is not getting used to begin with, combined with energy efficient small size electronic equipment, while utilizing the immense computing capacity available in billions of smart phones, a company or cooperative could theoretically create a connectivity service that is a lot cheaper than any alternative simply because this new service I speak of would not have to build infrastructure to get coverage in the first place and will not be held back by highly bureaucratic processes either.

A good example is FabFi, which is a wireless internet network made by MIT’s FabLab in Jalalabad Afghanistan using trash in 2010. This network remains one of the cheapest wireless networks ever built. If instead of trash, what if FabFi equipment could be installed on small little rented spaces on towers across the globe? Could such an idea help make internet access affordable and ubiquitous?

To some extent yes, but a global FabFi type wireless network will require access to something called spectrum. Spectrum is basically airwaves which are used by mobile network operators, TV stations and radio stations to transmit signals. Some spectrum like the one used by Wi-Fi routers or Bluetooth devices is free to use, most of it though, is allocated through auctions by governments. Governments have been known to charge as much as possible for licensing spectrum, and this desire to maximize spectrum price is a major factor other than the overlapping infrastructure capacity which keeps cost of connectivity high. The silver lining here is that similar to all other pieces of internet infrastructure, spectrum can also be shared.

A great example of a product using amateur spectrum on a shared basis is FireChat, this product is a messenger that works without an internet connection because FireChat creates a network of smart phones. If a quarter of people in this hall had FireChat installed and were equally spread out, I could send a message all the way to the back of the hall through FireChat without switching on my data service.

So clearly we have a situation where the technology and infrastructure capacity both exist to reduce the cost of internet access, but such technologies are not well distributed. Why?

Primarily because mobile network operators have created an inflated cost base which makes it impossible to provide internet services to less affluent communities, but also because FabFi like solutions appear quite technical in nature. Good thing is that already there are many companies consciously innovating to make wireless communication equipment a lot more modular and user friendly. Examples of such products include the Supabrick by BRCK and MeshPotato by Village Telco Project.

A combination of shared infrastructure, user friendly wireless equipment and amateur spectrum will not prove sufficient to live stream Game of Thrones in high definition, but such a service will make it feasible to run a basic google search and steam an audio podcast as well.

I envision a future when a Kalasha girl living in Northwest Pakistan can afford to own a device which she can speak into in her own language and get a straight answer. This is not feasible today considering 56% of web content is English, she can barely afford a smartphone and speech to text technologies have not quite matured. However, I stand convinced that a day will arrive when she will be able to speak into a device, ask a question about how best to treat her daughter’s fever, and get a straight answer. When that day comes, I want to ensure that this girl right here can afford to access the internet.

At present I don’t have a business model in my mind that will help me sustain this quest for answers, but I have roughly 3 years worth of savings to push this conversation as far as possible. I believe the best way to predict the future is to create it and that is why I decided to quit my career from the world of private equity investments and corporate board rooms.

I learned recently that there are 43 quintillion possible combinations of a Rubik’s cube, but in any position there are 20 moves that solves the cube. However, you have to get the first few moves right. I am convinced that making internet access affordable for all is an optimal first move. If we get this first move right I believe overcoming other global grand challenges we face today will become a lot easier as well, and we will finally become able to regulate this cyber river which we’ve all grown to depend on for our survival and prosperity.