The growth rate of Internet traffic in North America has slowed significantly over the past year. At face value, this data point suggests negative implications for telecom investment, since the excess capacity in backbone networks will take longer to absorb. In actuality, however, the slowdown in traffic growth should have little impact on carriers' investment plans. At some level, the slowdown has been anticipated: As the absolute traffic volume grows ever larger, it is unreasonable to expect continued triple-digit growth rates. Furthermore, traffic never was and will not be a primary driver of capital investment decisions. RHK believes capital expenditures (capex) by the wireline service providers will remain stable even as traffic growth slows.
Network traffic is a critical input to any analysis of the telecommunications industry, but its role is often not well understood. The link between traffic growth and capex tends to be exaggerated, while traffic tends to be overlooked as an indicator of demand. In fact, the role of traffic is almost the exact opposite of that: It is one of the more fundamental indicators of telecom demand but not a strong determinant of capex. For the purpose of gauging demand, RHK performs its regular Internet traffic analysis. RHK has closely followed Internet traffic growth for more than three years, using actual traffic from several of the larger network operators as the basis for analysis.
Internet traffic has enjoyed rapid growth since the late 1990s, due to a succession of drivers: starting with the rise of the World Wide Web, followed by business adoption of Internet applications, the deployment of broadband, and most recently, the use of peer-to-peer software for file sharing. Peer-to-peer usage has grown so rapidly that it now accounts for about 30% of all Internet traffic in North America and a similar percentage of traffic in Europe and Asia as well.
North American Internet traffic amounts to about 315 petabytes per month, dwarfing every other type of network traffic by a large margin. Traffic growth last year was 66%, down from 85% in 2002 and just over 100% in 2001 (see Figure). For 2004, RHK predicts traffic will grow by about 50%, continuing the gradual slowdown.A slowdown in the growth rate is expected naturally as the volume of traffic gets larger. The impact of successive growth drivers becomes somewhat muted by the large base of existing traffic. However, an additional dynamic has emerged recently: Traffic has also become more volatile. Last year, traffic volume actually declined briefly over the summer—it was the first time a decline was observed in a month other than December. That pattern seems set to repeat this year, with the decline possibly persisting over several months. This increased volatility is partially due to the popularity of peer-to-peer usage, which has been subject to legal challenges.
As a gauge of fundamental demand, the traffic analysis still gives positive indications. Although growth has slowed, a 50% annual increase in traffic is more than respectable for a mature industry. The lack of prosperity in the telecom industry must therefore be due to other factors—namely, an unfavorable market structure and deficient business models.
On the other side of the equation, what is the likely impact of slowing traffic growth on carriers' capex plans? As indicated earlier, RHK believes there is only a weak link between actual traffic growth and network investment. Because it represents demand, traffic is a necessary condition for network investment—without demand, there will be no profits and thus no funds to invest. However, traffic (or, more technically, network utilization) is not the key criterion for investment decisions and actually never has been. Rather, network investments have historically outpaced actual traffic by a large margin. That is evident in the degree of excess capacity that exists in networks today.
Among the large Internet backbone networks, average network utilization at the IP layer currently runs at 5% to 20%, depending on the specific carrier and route. At 50% traffic growth per year, it could be three years or more before additional capacity is required at the IP layer. At the transport layer, existing capacity could last even longer without new investment. Fortunately, carriers will not wait that long to make significant network investments.
Network capex will instead be driven by other considerations. In the past, it was a perception of future demand that never materialized. Thankfully, carriers are more savvy and rational now. Network investments will be driven by the need to roll out new services and capabilities and to reduce operational expenditures as revenues decline. As a byproduct of these investments, capacity will also be increased to accommodate traffic growth. Two analogies may make this clearer. On the one hand, consider a semiconductor manufacturer such as Intel. The company builds fabs to take advantage of new production technologies that enable more powerful chips. On the other hand, a commodity producer of aluminum such as Alcoa typically invests for the sole purpose of expanding capacity to meet demand. Telecom carriers are more like Intel and less like Alcoa.
As a result, RHK expects North American wireline capex to remain roughly stable over the next five years even as the rate of traffic growth slows. That is good news for suppliers of network equipment that have contended with double-digit declines in carrier spending since the telecom bubble burst three years ago. Total wireline capex was $29 billion last year, one-third of which was spent on equipment. With the increased emphasis on new services and new capabilities and less emphasis on absolute capacity (which requires spending on fiber and cable plant), RHK believes the share of capex spent on equipment will notch upward, further benefiting network equipment vendors.
Shing Yin is program director of telecom economics research at RHK (South San Francisco).