When we last examined Facebook’s capex, we observed that year-over-year the company’s spending on property and equipment has been declining. We concluded that the reduction in network infrastructure spending reflected a maturation of the company’s infrastructure and platform. Facebook’s somewhat modest FY 2013 capex bears out this thesis. However, Facebook’s WhatsApp acquisition raises questions about capex going forward.
Facebook estimated that capex in 2013 would grow at only 12%, totalling an estimated $1.8B (later revised down to $1.6B). As we noted previously, this would reflect a slowing capex growth rate. In fact, a combination of efficiency investments on the hardware and the software side have further driven down realized capital spending in the year. According to its 10-K, as a result of these investments, total capex was a more modest $1.36B.
By tracking the historical cost of Facebook’s property and equipment we can see that the majority of capital spending is directed to growing or maintaining the company’s network infrastructure. This includes the purchase or lease of servers and other network equipment, and building costs for Facebook’s growing list of data centers. This also explains the bulk of the decline in capex over the year, with only a small increase in nominal cost in Q4 2013.
Although the decline in network infrastructure capital spending is primarily a function of Facebook’s maturing platform, the company noted that some of the reduction in spending over the past three years resulted from investments including the Open Compute project and proprietary software developments. The company said these investments have lessened its hardware requirements, suggesting the company has been forced to purchase fewer servers for its nearly 1.3B users.1
Facebook anticipates 2014 capex will increase to between $2B and $2.5B. This represents between 47% and 83% increase in capex over FY 2013. Facebook explained that its estimate was driven primarily by expansion of its Menlo Park headquarters, infrastructure-related expenses including costs related to its Iowa data center, and by other initiatives. Given its user base growth rate, it seems likely that the bulk of Facebook’s capex increase is related to its campus expansion.
However, the WhatsApp acquisition materially changes Facebook’s infrastructure composition and its user base. We should, therefore, also expect an effect on capex. Although we do not have sufficient information to trend WhatsApp usage, the company recently provided some color on the matter: 64B messages sent and received in 24 hours.2 High usage is matched by a growing active user base: 450M at the time of the company’s acquisition.
The extent to which WhatsApp will affect Facebook’s capex is unknown. Facebook’s hands-off approach to its acquisitions suggests that network infrastructure spending will not increase substantially in the short-run. However, as WhatsApp continues to grow its user base and expand its infrastructure in emerging markets, those costs should be reflected in increased capex for network equipment.