|
Terrestrial Bandwidth 2004 Executive Summary |
|
|
|
Terrestrial Networks, the second volume of International
Bandwidth, profiles over 100 terrestrial network operators in Europe
and the Americas, and provides a comprehensive overview of long-haul
network supply, demand and pricing. Submarine Networks, the
companion volume to this report, provides similar analysis of undersea
fiber-optic systems.
The initial chapters of this report offer a basic overview of bandwidth
products, contracts, technologies, and costs to provide context for
understanding how that supply is bought, sold, and produced. Next, the
volume quantifies the supply of fiber-based capacity—both current and
potential—and explores the effect bankruptcies have had on the field of
providers. The report then examines the demand side of the equation,
charting current demand growth and the factors determining factors. A
thorough review of pricing trends concludes the analysis.
Supply & Demand
In the last ten years, new fiber-optic technologies coincided with a
frenzy of network construction to multiply the supply of lit bandwidth
between many European and North American cities by over a thousand-fold.
The anticipated demand boom that would fill these new circuits and
wavelengths, however, hardly materialized. The resulting glut of unsold
inventory depressed capacity prices to 10 percent or less of the value
envisioned in the (pre-bust) business plans of many network providers.
Facing a mountainous overhang of unsold bandwidth, sellers may hope
that demand for bandwidth catches up to supply, or at least that industry
consolidation reduces supply levels. Data suggest that neither hope will
be fulfilled within the next several years.
Scenario 1: Increase Demand. Bandwidth providers may
take comfort in some encouraging signals. Despite the market downturn,
demand for capacity has remained robust. International network usage from
London, for example, surged from 34 Gbps in 1999 to 620 Gbps in 2003 (see
Figure 1. Map of International Circuit Usage by City, 1999 & 2003).
Burgeoning Internet traffic has driven the majority of recent bandwidth
purchases. After dipping in 2002, the pace of new capacity deployments for
use in Internet backbones re-accelerated in 2003.
Despite such gains, quantities of purchased bandwidth pale in
comparison to both current (lit) and potential supply. For example, 10.2
Tbps of lit capacity passes through Frankfurt, yet the sum of
international bandwidth from Frankfurt used for Internet, voice, and other
networks totals only 357 Gbps—not even 5 percent of lit supply.
Inventories of potential capacity are even more astounding. The town of
Helena, Montana (population 26,000), for example, has access to a
potential (lit plus unlit) 113 Tbps of bandwidth. At mid-year 2003, total
Internet bandwidth to all U.S. homes totaled only 29.2 Tbps—26 percent of
the potential capacity passing through tiny Helena. Because companies can
light dark wavelengths and fiber with relative ease, providers cannot
discount latent bandwidth as merely hypothetical.
Scenario 2: Reduce Supply. Falling prices and
slower-than-anticipated demand have driven over two dozen bandwidth
providers into bankruptcy since 2000, yet major industry consolidation has
not occurred. Only a handful of bankruptcies resulted in outright
liquidation; most companies emerged from bankruptcy after a restructuring
period (see the timeline in Figure 2. Bankruptcy Timeline). With prices
for fiber and other network assets fetching so little on the open market,
most creditors have elected to resurrect bandwidth service providers
rather than liquidate their network. Even when bandwidth providers have
been forced into liquidation, their network has invariably been acquired
by a rival company, usually for pennies on the dollar.
As a result of the deliberate pace of market consolidation, quantities
of bandwidth and fiber have not changed significantly in last three years.
Most major cities in Europe and the U.S. are connected to multiple
terabits of unsold lit capacity (see Figure 3. European and U.S. Network
Supply, 2003: Providers, Fiber Count, and Bandwidth). As of early 2004,
supply overshadowed demand to such a degree that a supply-side solution to
the supply/demand imbalance would require that over 75 percent of existing
bandwidth providers cease operations entirely—and that the market
survivors refrain from purchasing the stranded network assets.
Prices & Outlook
The severe supply/demand mismatch has caused capacity prices to
collapse. In the past five years, OC-3 lease prices on major U.S. routes
have plunged by 85 percent to more than 90 percent, cumulatively. The rate
of price erosion abated somewhat in 2002, with annual decreases on most
U.S. routes averaging a relatively modest five to 15 percent. This respite
proved to be short-lived: price cuts picked up speed again in 2003, with
prices on most routes dropping by 10 percent to more than 30 percent (see
Figure 4. OC-3 Lease Prices on Major U.S. Routes, Q1 2002-Q4 2003).
As was the case in the U.S., European price declines abated slightly in
2002, only to gather speed again in 2003. After falling five to 15 percent
in 2002, European circuit prices dropped by 15 to 35 percent in 2003.
Price erosion has taken on a momentum of its own, as buyers have come
to expect steady price reductions, and sellers have become accustomed to
meeting buyers’ aggressive target prices. In early 2004, TeleGeography
conducted a survey of bandwidth providers to determine how they expected
to adjust their prices in the coming year. Responses were surprisingly
varied. Some carriers expected to keep prices unchanged, while others
projected price cuts of up to 40 percent. In practice, these differences
may be less significant than they appear. Companies expecting to make deep
price cuts were usually relatively high-priced providers, while some of
the most fiercely competitive carriers expected declines of ten percent or
less. On balance, carriers’ responses suggest that prices would fall 14
and 16 percent on terrestrial routes in the U.S. and Europe, respectively.
Whether carriers’ expectations are an accurate predictor of market
trends remains to be seen. However, given that bandwidth providers already
expect to cut prices by approximately 15 percent, it would be surprising
if prices fell less, given that aggressive buyers will certainly demand
deeper reductions.
Carriers’ views on future prices reveal an underlying tenet of market
expectations: any market must be comprised of both buyers and sellers, and
market expectations are a reflection of the views of both parties. After
more than half a decade of rapid price erosion, there is now a real danger
that market expectations are becoming a self-fulfilling prophecy.
|