Building data centres in Africa

After two days of data centre conferencing in Monaco what emerges is consensus about a few things.   The imminent impact of the Internet of Things, outsourcing, virtualization, Colocation and the ongoing migration to the Cloud,  but also the awareness that there exists a massive shortage of data centre capacity in Africa and the urgent need to address it.   There is more data centre capacity within the M25 of London than on the whole continent of Africa.

However, where there is some disagreement is exactly how to go about it.  Where to build?  Who to build?  How to build?  Who to finance?   Doing business in Africa is not easy.  But we are inspired by JFK–– “We do these things not because they are easy, but because they are hard".     Africa is hardly the moon but to some people it might as well be!

Power, that ever present bogeyman, seemingly presents one of our biggest obstacles.   How can you build data centres without power?   In my view the lack of power should never be an impediment to the building of a data centre whether in Africa or anywhere else.    Twenty years ago when we started building mobile networks in Africa we didn’t hold back because there was no power.  That was a challenge to be overcome and to this day many mobile networks are still running on generators.  Those operators have pretty decent margins too.  Clearly this must change and it is.       

The provision of power in Africa should be pursued with alacrity.   New power sources should be located where they are required.  The age of large, monolithic, centralised power plants, whether coal fired or nuclear powered, is behind us.  New solar, wind and even clean gas powered plants should be built to feed not only the new data centres but also the new factories, offices and homes that need to be built in Africa.   This will all happen in time.

The building of a data centre should not be dependent upon the availability of grid power.  Generators should be used in the short term until the grid or other local supply kicks in as ACS are doing in Angola for their brand new data centre built by Flex.

But what are the challenges that lie ahead for aspirant data centre operators in Africa?  I posit that these challenges can be classified into two broad categories.   Hard and soft challenges.    Similar to the terms used in geo-politics – hard power and soft power -  where hard power means military power while soft power means diplomatic power.

In the world of the data centre the ‘hard’ issues are technical – land, power, logistics and engineering.   The soft issues are a blend of the local dynamics, the market, the partners, and the politics.  After all there are 54 countries and many more languages and cultures.   Is a local partner required?  How to engage with Government?  Who are the customers?  Is there rule of law?  What regulations apply?   Is a licence needed to operate?  What are the local employment laws?  And so on.  The soft issues are in fact the ‘hardest’ to resolve.  To find someone to build a data centre is much easier than having to navigate the rapids of local politics, partners and the local rules and regulations.

A critical issue that spans these hard and soft categories is that of finance.  How does one finance the first phase of a data centre that by any measure is small, i.e. 50 racks?    There is a notable lack of enthusiasm on the part of the usual suspects – the VC’s, angels, funds, banks and the DFI’s.  

Data centres constitute ‘critical’ infrastructure – in the same way that roads, bridges, harbours, railway lines, power plants and hospitals all constitute critical infrastructure.   According to McKinsey and Frost & Sullivan, Africa needs to invest $90bn per annum for the next 10 years to give it a hope of getting anywhere close to the levels of development that we take for granted in the developed economies.   Data centres are as important to the welfare of any African country as is any of the ‘traditional’ infrastructure.

As Africa continues to leapfrog its way into the digital age with the rapid take up of smart phones (about 500m new smartphones will come on line by 2020) the data which is either generated by users or content providers will need to be hosted, processed and delivered locally.  The reasons are manifold and beyond dispute and are covered in numerous articles elsewhere and include issues such as data sovereignty. 

So surely the DFI’s, who have a ‘development’ mandate, should be rewriting their rule-books and actively seeking ways of funding the new data centres in Africa.  They could provide funding which ranges from pure equity to loan guarantees.   But on a scale that is demanded by the market, not what is written in their handbook, being large cheques of $50m plus. 

That leaves the entrepreneur having to scramble around looking for seed capital.   Now if this was Silicon Valley and at issue was a geek trying to fund the development of the next Uber App or photo- sharing site then yes, it is understandable that the DFI’s would be wary.  But when it comes to critical infrastructure that will ensure that some of these economies survive and prosper then the DFI’s should be getting their hands dirty.   After all would one expect a railway line or pipeline be built by an entrepreneur?