Strategy & Operations » Leadership & Management » Engage: launching a data centre business from scratch

A DATA centre can earn over £200 per square foot while the warehouse or “shed” itself costs less than £10.

On paper such a development project looks a great business, especially as planning permission is relatively easy to come by.

In reality, however, a typical data centre will cost up to £1,000 per square foot and anything under 20,000 square feet is too small to give a good return on investment. A potential developer therefore needs up to £20m of funding and, whether debt or equity funded, to raise this kind of money there must be a customer in place from the start who’s willing to commit to a long term contract worth around £20m to £30m.

Funding dilemmas

Bearing in mind at this stage of the game all there usually is to see is a warehouse shell, the customer also has to display a huge amount of imagination and trust, not to mention patience, considering build-out timescales of up to 12 months.

At the same time, any potential funder will need convincing as their money will be secured against the customer contract. They must be confident both of the covenant of the customer over a long period and the developer’s ability to deliver. In today’s world it’s surprising how few companies pass the test, successfully closing the loop between suitable customer, their requirement for a credible project plan and realistic funding arrangements.

To compound the problem further, data centre funding doesn’t fit exactly into any boxes which is why many banks are unsure how to approach debt funding for data centre development projects. It looks like a property deal as there is a long term contract with a tenant and money is required to fit out the building. But, a data centre is a complicated structure and there are significant operational risks with contractual penalties for failure to perform. In extreme circumstances the customer could terminate their contract if the service provided is really bad. Therefore, with this level of operational risk it looks much less like a property deal.

It could also fit into asset finance as it is a capital intensive business using large high value pieces of equipment like generators and transformers which keep their value better than computers and servers. But those assets are not stand-alone, they are woven together with hundreds of metres of HV Cable into a system worth more than the sum of the parts. While, in a default scenario this system could be unpicked to recover the assets it is not a simple task. As a result it is difficult to treat the deal as pure asset finance.

Receivables financing is the other option where the bank will have a receive revenue from the rent from a (hopefully) blue chip tenant. But these are not existing receivables which are being funded, rather future receivables, which will be generated once the data centre has been successfully built. As such, the bank is exposed to build risk which is not something usually associated with receivables financing.

Making It Happen

In 2008 and through to opening for business in late 2009, facing all of these challenges, we commenced the NGD Europe data centre development project in Newport, South Wales.

Worse still, we were trying to launch throughout the banking crisis. Fortunately, we had already identified a high quality fit-for-purpose building which we subsequently acquired from the Welsh Assembly Government; a former semiconductor factory originally built for LG that was of a high standard and already had the essential features in place including fibre networks from three carriers and an on-site high capacity power sub-station.

In short, it was a very convincing site for a data centre, especially compared to a pure Greenfield one featuring a muddy field and not much else! From day one we invested equity to upgrade our building by adding security fences, digging in cable conduits, drawing up architectural plans and acquiring planning permission.

Visitors to the site could see the level of activity and that the project had momentum. We also had a unique product. If the product had been identical to others, why would any customer take a risk with a new company?

Our product was deliberately positioned as an “out-of-town” alternative which was between 25 and 50% cheaper than the London market. Our product was also designed by experts and used only the best products.

This helped convince our initial customer – BT – that we knew what we were doing from the outset. Last, but by no means least, we found the right part of a bank to fund the data centre. Lombard, which initially put up £4m, approached the funding from the starting point of asset finance but had the flexibility to understand the risks of the business and structure a deal that satisfied their credit committee.

Bright Future

Once established as a successful data centre operator, the future looks very bright as we are in an industry where the raw material (data) is doubling in volume every two years. While the barriers to entry are relatively high, the margins are good and holding up.

There are few industries where EBITDA margins of 50% are common and with a 400,000 ft2 building, the economies of scale available are significant. Our gamble that due to tumbling fibre costs the data centre industry would begin to migrate away from congested and expensive urban centres seems to be paying off.

The emergence of cloud computing, which by definition can be hosted almost anywhere, can only add to the demand for safer, cheaper and more remote data centres.

This year we have secured significant long term multimillion pound tenancy agreements from a growing number of large European and North American organisations requiring custom built data halls within our South Wales facility.

These are being financed from a combination of our growing revenue stream, advance payments from new customers, further asset funding from outside, including Lombard, as well as private shareholders.

Throughout, it has been our priority to maximise shareholder value and therefore avoid taking any public offering or VC route too prematurely; until such time we have reached critical mass and top market value for the business. With our significant forecast growth for the coming three to five years, and bearing in mind the capital intensive nature of our type of business, such a time may not be too far away.

Rob Edwards is finance director of data centre operator Next Generation Data Limited