Infrastructure investment: Roads, rail, ports… cloud?

Do we need to widen the definition of infrastructure in the 21st century? Is it still just roads, ports, rail, electricity and the like, or should we be including the cloud – the virtualised and scale-able platform on which an increasing number of companies are starting to operate.

Like who? Netflix, for one, as well as Kellogg’s, Kraft, Conde Nast, Unilever, Suncorp, Expedia, Vodafone, Berkeley (the US university) Siemens, Philips, Brooks Brothers – it’s a long list, and getting longer.

It’s true that there are some who already include in infrastructure the physical server equipment used in networking, but now it is extending to include the software that enables virtualisation, which is what the cloud is really about. This virtualised cloud has become vital, but it is not yet at the stage where it has been commodotised by investors, and so priced as a utility with a low multiple. This is a good place to be.

Like rail, ports or any traditional infrastructure, the cloud works because it has massive capacity, which it is able to parcel out in any-size pieces at per-unit economics which work well enough to pay the significant capital costs as well as the tiny marginal costs.

But utility-sized scale only works if the demand side is there too.

And it is. As noted, the big end of business is getting used to the cloud, while at the other end, many start-ups and SME’s could not exist without it.

Think about it. Even a small business would once have required a rudimentary IT capability for email, spreadsheets, CRM, sales and back-up. Twenty years ago, an organisation of 10 people would see the cost of this run to over $100k. Now, it’s $20/month, pay as you go. The tools available are powerful, scale-able and oh-so-cheap.

It’s the existence of the cloud, this infrastructure, which has enabled so many start-up businesses, like Airbnb, Soundcloud and Slack to name just three. And the runway for the next ten years is long, and getting longer as disruption rolls through businesses previously thought to be moated.

Remember, the internet worked at first because the thing that is consumed – information, whether a voice call or Game of Thrones – was already digital, and so could travel around the globe in an instant. If operating on the cloud were constrained to these businesses, which are at their core already digital then the demand curve would not be as long or steep as it is.

The internet shouldn’t have worked for companies where the goods were physical – for example retailing because it involves not bytes travelling at bits but the physical exchange of atoms – goods – in much slower, people time.

But retail itself is an amalgam of physical and digital elements including selection, purchase and fulfilment. Some elements do involve physical transfer of merchandise, but big chunks do not, like browsing, payment and the logistics in the supply chain which enable on-time delivery.

If a sector such as retail, which is so rooted in physical exchange, can go digital, how about manufacturing, transport and tourism? Of course they all can and have.

This helps explain why there are so many start-ups at present – the internet takes industries which were previously considered basic building blocks of the economy and breaks them down into even more fundamental bits, then makes efficient those bits by injecting information which improves cost and efficiency. The difference between a travel agent and a travel website has been eliminated because there is now almost no asymmetry of information, which was the way the former used to make their margin. (By this I mean that the traveller has access to all the information which used to be held by the travel agent, and through which the travel agent return was earned.)

It used to be that the physical copper networks (or the railways or ports) were considered the infrastructure on which all this was possible, and it still is. But this definition is too narrow and doesn’t include the new layer of capacity that facilitates so much of this. This is what the cloud is, and why it is increasingly being classified as infrastructure.

There are really only a handful of big players in this space – because massive scale is the critical success factor – with Amazon through AWS the biggest, and Microsoft (through Azure) second, and Google(!) a distant third. The share price performance of the first two has definitely been assisted by the understanding by investors of the valuable positions their cloud services business have. There is probably more to come.

We could add to this companies like Oracle and IBM, but these are really half-way houses – for example, Oracle is creating cloud software to run in the data centres of its big enterprise customers.