Will echoes of the dot-com bubble resonate through the industry supply chain?
August 1, 2011
History tends to repeat itself, but let us hope that the optical communications industry will handle the current rise in dot-com valuations better than it did ten years ago. Valuations of Facebook, LinkedIn, and many other Internet darlings are sky high, indicating that an echo of the dot-com bubble of 1999–2000 may be approaching. Although strong recovery in demand for communications equipment and components lifted valuations across the industry in late 2010 - early 2011, few investors stay committed to this market and it remains volatile, as small fluctuations in quarterly earnings can lead to huge changes in stock valuations of companies.
Profitability across the whole supply chain of the communications industry improved substantially in 2010. However, a wide gap remains between companies like Google, which generate high profits by taking advantage of the network at almost no charge, and those composing the rest of the industry, which are struggling to stay profitable while upgrading networking infrastructure to keep up with rapidly increasing traffic.
Service providers are under pressure to maximize their revenues while offering the infrastructure to connecting content providers, who want access to the largest number of customers, with end users, who want access to the widest selection of services. Until service providers find a balance in this two-sided market, they will continue to struggle to generate enough revenue to realize a profit and invest into the infrastructure to keep up with growing Internet traffic. Underinvestment will inevitably lead to lack of network capacity, negatively impacting the whole supply chain including content providers and end users.
LightCounting analysis of the global market for optical components in 2001–2011 suggests that upgrades to network bandwidth lagged behind growth in Internet traffic for most of the decade. According to one of the U.S. service providers, excess network capacity created during the bubble years was large enough to avoid substantial investments in infrastructure for several years. However, most of this excess capacity must have been used by now or made obsolete by changing traffic patterns. Unless service providers increase investment in networking infrastructure, bottlenecks in capacity will start disrupting the network in the next few years.
The urgency in resolving this situation was recently exposed at a CEO roundtable hosted by the European Union Commissioner for the Digital Agenda, Neelie Kroes on July 13, 2011. Europe is lagging behind Asia and North America in installations of fiber to the home systems, and service providers are in no rush to increase investments until regulations ensure a return on their investment. “While it is understandable that commercial players try to maximize their own advantages, we also need to recognize that we have common interests. We all want Europe to catch the high-speed broadband train,” Neelie Kroes said after the meeting.
The problem with profitability of broadband access extends beyond installation cost of FTTx networks. End users’ demand for significantly faster access will inevitably impact traffic across the whole network, which will require significant upgrades to manage the increasing traffic. LightCounting analysis shows that current utilization of FTTx access networks is very low—less than 1%. As consumers learn how to use bandwidth available to them, the impact on internet traffic may be dramatic. Problems with limited bandwidth of wireless networks, such as those that AT&T experienced in California two years ago after introduction of the iPhone, may be just the tip of the iceberg. Bandwidth of FTTx networks is up to 100 times higher than that required by wireless devices, so their impact on the rest of the network will be much more significant.
Although China leads in installations of FTTx systems, average household income in China remains well below those of families in the United States and Europe. There is no solid statistical data on how bandwidth consumption scales with income, but it would not be surprising to find that more affluent consumers put a greater load on the network. If Europe does board the “high-speed broadband train,” there may be a lot more traffic across the core network than operators are planning for.
The current situation in the United States offers another example of the problem of sharply increasing network traffic. AT&T and Verizon report continued growth in the number of broadband customers, including FiOS and U-verse TV subscribers. Meanwhile, traffic generated by video on demand distributed by Netflix accounts for 30% of peak downstream traffic in North America. It is good that Netflix has a service contract with Layer 3 Communications, which is now merging with Global Crossing—the two companies most often credited for creating network overcapacity ten years ago. This overcapacity is being put to good use now, but how long will it last?
The echoes of the dot-com bubble are certainly impacting wireless segment of the global communication industry. Ericsson – the leader in wireless networking equipment – reported solid performance in Q2 2011 and it is on track for 20% growth in revenues for the year. In comparison, Huawei – a dominant provider of optical networking gear – is anticipating only 10% growth in 2011 revenues, which is significantly lower than 30% achieved in 2010. Alcatel-Lucent reported modest single digit increases in sales of both wireless and optical networking equipment in Q2 2011 compared to Q2 2010, but the optical business revenues inched lower compared to Q1 2011. Tellabs business declined significantly in early 2011 and the company is laying off 10% of its workforce. Cisco’s revenues are steady in 2011, but the company recently announced plan to lay off 6,500 workers, as part of its business restructuring.
Cisco is at a turning point: the company is under siege from all sides, but hopes to turn its financial fortunes around by leveraging the opportunity to converge the storage and switching infrastructures in datacom networks. Cisco's strategy is to leverage Fiber Channel over Ethernet (FCoE) and to merge the SAN and LAN markets. FCoE may well result (eventually) in replacement of the switches in the data center, with Cisco, the largest purveyor of switches, benefitting the most. However, the company is under fire from several opposing fractions, including server original equipment manufacturers (OEMs) offering a full line of not only servers but storage and switching as well, low-cost products from China, and new types of switches that enable fewer overall layers of switches such as Juniper's QFabric. To drive acceptance of FCoE, Cisco needed to either introduce its own servers that adhered to and exceled at implementing the new standard or partner with a server OEM to do so. Cisco chose the former. The company began work on the server (UCS) and switch (Nexus) combination in 2006, and introduced the systems in early 2009. Since then, former partners such as HP and Cisco have entered into direct competition, as HP acquired 3Com. The acquisition of Blade Networks by IBM last year and a recently announced sale of Force10 to Dell offer another confirmation of this shift in the industry. All three server OEMs also offer their own storage products: HP is the overall leader in total storage sold each year; IBM offers the high end XIV storage platform and many other midrange and lower end storage products, and Dell is the leader in iSCSI storage. On the Chinese front, Huawei is mounting a major offensive in the United States through its joint venture with Symantec. The few remaining pure play router and switch vendors, such as F5 Networks and Juniper, are reporting significant increases in sales and fending off potential acquirers. With all the new competitive forces targeting Cisco's business, the company's restructuring and renewed focus can't come soon enough.
The communication semiconductor market segment was swept by a wave of mergers and acquisitions in the past three months with all major players grabbing high-data-rate chip expertise, as deployments on 40G and 100G equipment gain momentum (http://www.lightcounting.com/100Gbps.cfm). Intel’s recent acquisition of Fulcrum Microsystems offers an example of this trend. Other examples include the acquisitions of Sarance by Xilinx, SolarFlare by Marvell, and Teranetics by PLX. New venture funding in this segment is increasing as well, with new rounds going to Aquantia, Calopia, ClariPhy, IPTronics, LightWire, OneChip, Tabula, and others.
Optical component and module suppliers continue to consolidate, with a recent merger of Accelink and WTD creating the largest, well-diversified Chinese vendor in this segment, comparable in size to the leading Western manufacturers. Finisar and Oclaro took advantage of spikes in the financial markets to finance their acquisitions in 2010-2011. Cabling companies have also been on an acquisition binge with Tyco buying Zarlink, Molex buying Luxtera, and most recently Samtec buying AplenIO. More details on changing landscape of the industry supply chain are available in LightCounting’s report titled “The State of the Optical Communications Industry: On a Path to Sustainable Profitability.”
This report also outlines the major technological transitions impacting evolution of the networks and the whole industry, including the increasing popularity of cloud computing and over the top (OTT) services and the emergence of super datacenters and flat networking architecture, which drive demand for higher data rate optics. All these trends combined with the relentless growth in internet traffic point to solid demand for optical networking in the future. The industry is also much better positioned now to take advantage of the market growth and not get carried away by setting the expectations too high. Depending on “dispersion of the supply chain” the echoes of the dot-com bubble may stretch into many years of sustainable growth, which is the best scenario for any industry.