In our last post, we discussed how service provider revenues are massive but have settled into a low growth rate of around 1 % per year. How does this affect CapEx?
Network operators can drive new sources of revenue because they invest in their networks. Service provider CapEx is a source of intense interest because the absolute dollars are so large, it affects the prospects of carriers and the spending patterns have such an impact on suppliers looking for their piece of that pie. The prospects for new technologies like 5G are evaluated by the CapEx plans.
Service provider footprints and services profiles can differ significantly. One way that we can compare investments is to look at capital intensity which is CapEx expressed as a percentage of revenues. Obviously, this will vary somewhat over time and from company to company, but it typically is between 15 and 20% of revenues. For example, the European Telecommunications Network Operators’ Association, did a study that showed carrier CapEx rising from 15% of revenues in 2012 to 18% in 2017.
Carriers have a predictable range of capital intensity because they are constrained. CapEx either must come out of a carrier’s cash flow or the carrier must take on debt. When revenues are increasing slowly as they are now, this limits a service provider’s ability to fund CapEx projects. Debt is an option but debt service also has to come out of cash flow so that is constrained as well. Too much debt can also be a cause for concern. If there is an economic downturn, this would hurt revenues and cash flow making debt service more of an issue for a given company. S&P summarized the scope of the problem:
“Clearly, the biggest risk is that tightening credit market conditions could lead to higher borrowing costs and elevated refinancing risk. This is particularly relevant given the substantial increase in global telecom debt, which now approximates $1.9 trillion, or almost double the amount in 2012. There is also about $180 billion in global telecom debt coming due over the next year, much of which will need to be refinanced.”
What all of this suggests is that service providers should and are rethinking business as usual. There is a lot of anecdotal evidence to that effect. One is a shift from proprietary hardware to software with NFV. For example, AT&T stated its goal of having 75% of its core network functions virtualized by 2020 and as of the end of had 65.5% of its network virtualized. Software is much more cost-effective than hardware appliances. Commodity hardware in the form of servers or white box switches is gaining traction.
Another option is to better align costs with service revenue. With software, a subscription model allows the service provider to combine both acquisition and maintenance costs into a single payment. An annual subscription does not come out of CapEx and does not need to be amortized. Moreover, a good subscription model should be closely tied to the revenue stream it enables. At Volta, we like to think that we have that. Our subscription model is inclusive, meaning support and upgrades are included. The bulk of our pricing model is based on the number of virtual routers you run. It allows you to turn a VR on quickly when required by customer demand. You can also shut it down or redeploy the license just as easily. The shift in both service provider edge architectures and revenue models will make it more important to align cost and revenue over time. The days of buying a big router with a big upfront cost and selling off capacity over time are gone.
In our final entry on TCO, we will look at two important revenue considerations: velocity and agility.