americasnetwork.com
The Coming Bandwidth Bubble Burst
It's the pain phase for America's fiber barons. Nearly 600,000 miles of new intercity fiber is on the way. Capacity prices are dropping and major dot-com & CLEC customers are failing. And Wall Street doesn't want to help.
By Grahame Lynch
If anyone doubted there was a worldwide bandwidth explosion, the evidence came from Singapore Telecom late last year. Last October, it announced an eight terabit cable connecting Singapore to India’s Chennai and Mumbai. Nevermind that India and Singapore have just 188,000 Internet hosts between them.
Or that this one cable’s capacity dwarfs total current international broadband demand by 400%, according to one estimate. Or that India’s fixed line teledensity measures just 3% and that Singapore, with a population of just 4 million, will never have a bigger Internet base than Los Angeles County.
But while the Singapore-India cable is the largest capacity link announced to date, it is by no means an unusual case.
As multiple fiber backbones come online in North America, their ambitious owners are seeking to take their business model global.
The bandwidth growth in North America alone continues to astonish. Consultancy Adventis Corp. said in a recent report that if 60% of installed fiber was lit this year, the amount of available bandwidth will be 400 times that of 1998. Using a simple methodology, it calculated that bandwidth demand will have increased 20 times over the same period.
But the growth in inter-continental capacity is equally dramatic. According to Telegeography, the amount of international bandwidth supply is set to double from under 3 Tbps last year to nearly 7 Tbps this year. It will nearly double again the following year to almost 12 Tbps.
For the most part, it’s the same companies who are driving both North American and inter-continental capacity. Global Crossing, Level 3 and Qwest are expanding beyond their North American bases with rollouts to Asia and Europe. FLAG is adding to its Europe-Asia cable with both a trans-Atlantic and trans-Pacific link. Global Crossing and 360networks are rolling out fiber links across Latin America.
These American rollouts are being emulated by regional players such as Telefonica in Latin America and Telstra, Pacific Century Cyberworks and Singapore Telecom in Asia.
The Global Data Explosion
The fundamental basis for these rollouts is the observed exponential growth in North American bandwidth use and the expectation that it will be emulated elsewhere. By one estimate from Phillips Group, 2001 domestic data demand is likely to reach nearly 113 terabytes a month, up from 46 terabytes last year. Phillips forecasts that demand will double again next year to reach 215 terabytes a month.
Internet globalists such as Level 3’s Asia managing director Steve Liddell say that the opportunities for bandwidth suppliers are even greater in places such as Asia because those markets have been constrained by monopolies.
But the big story of the last 12 months has been Wall Street’s rejection of the bandwidth barons’ prospects. Every fiber backbone major has been hammered by the stock market.
In fact, over $100 billion of value has been wiped off the global fiber operators over the past year. One of the worst affected is FLAG, valued at just $850 million, even though it has $1.3 billion of cash on hand and a net asset position of $1 billion — effectively valuing the company at a negative worth. Global Crossing’s price has plummeted from the $60s to as low as $11 billion while Level 3’s price peak of $132 billion has fallen to as low as $25 billion.
So what went wrong?
Epoch Partners’ analyst Mark Langner traces it directly to the negative sentiment against “first generation dot-coms, business-to-business commerce enablers, application service providers and voice-based CLECs.”
Adds Philips, in a white paper, “The understandable fear is that, if the Internet model is failing, the supporting layers of that model will fail too.”
Bandwidth is not a commodity
Compounding this sentiment is the view that bandwidth is becoming commoditized and that margins are disappearing.
Langner believes these perceptions are incorrect. He points out that contrary to popular perception, bandwidth is not becoming a commodity. “It lacks some of the key traits that define it as a commodity such as liquidity, accessibility and price visibility,” he says.
Philips concurs, writing, “At present, corporate users do not find capacity easy to obtain or financially viable to deploy. Commercial Web site operators, such as eBay and Amazon, have to use multiple providers in order to attain sufficient capacity to support their enterprise.”
For Philips, the success of Napster highlights the exponential nature of bandwidth demand. “The arrival and deployment of a single MP3 distribution application overwhelms university backbones,” he observes.
Doing well?
Indeed, the major backbone operators don’t seem to be performing too badly in the marketplace. 360networks predicts it will beat analyst forecasts by about 40% in 2001. Metromedia’s 3Q revenue results beat expectations. FLAG says it has already made $800 million of pre-sales on its planned trans-Atlantic cable, nearly exceeding all of the sales made by the company on its Europe-Asia cable to date.
But beneath the bullish exterior lie some warning signs, in terms of both the quality of customers and the prices that can be charged to them. For example, XO Communications (previously NextLink) recently re-negotiated its European fiber agreement with Level 3, slashing 47% — around $150 million — from the original deal. According to the International Telecommunication Union, international circuit costs are decreasing at a rate of 72% annually.
Even the largest of the new US operators in terms of fiber miles, Metromedia, is making many of its sales to startup carriers and dot-com companies. Its most recent major wins include such no-names as First Mark Communications and GiantLoop Network, as well as online providers who may fall victim to the vagaries of dot-com fashion such as Datek.
Life on the edge
Out on the network edge, immediate prospects don’t look good. The strong growth of cable modem services has been somewhat overshadowed by the disappointments of independent CLECs and ISPs. Struggling Covad Communications, North America’s largest CLEC, admits that 14 major ISP customers aren’t paying their bills and that four have entered Chapter 11. It’s freezing its rollout and reducing equipment orders.
Other CLECs are dropping like flies. Northpoint reseller Flashcom, with nearly 13,000 DSL customers, has entered Chapter 11. Another CLEC, Digital Broadband, also succumbed to financial duress, while HarvardNet has decided to drop DSL altogether.
These competitive broadband providers should have been the driving force behind fiber demand. But while DSL and cable modem connections will continue to grow at a healthy clip, the stimulus that comes from competitive providers is likely to be much diminished in coming months.
Is demand slowing?
Some observers have become openly critical of industry claims, promulgated by the likes of UUNet and Level 3, that Internet demand is doubling every three to four months.
The actual figure, according to studies of academic institutions and peering point data conducted by AT&T Research Labs, is a doubling every 12 months.
“The myth of Internet traffic that doubles every three or four months is dangerous,” argues Andrew Odlyzko of AT&T Research Labs.
“It leads to bad decisions. It sure helped to inflate the current bubble in optical networking stocks. After all, if demand is outpacing supply of transport capacity, then money making opportunities are virtually limitless.”
Telstra managing director, international carrier business, John Hibbard estimates that there is only 100 gigabits of bandwidth actually being utilized on submarine cables right now, and at present growth rates, this won’t reach 1.6-terabits until 2005.
“If you look at these and 6 and 8 terabit cables, what sort of growth will you need to get fills?”
As the Internet market matures, usage patterns are consolidating. Nielsen NetRatings has recorded a significant drop in the number of active US users, from 77 million to 70 million between January and October 2000. Other surveys suggest that the average user is spending less time online and visiting fewer sites.
While mass applications such as Napster and the Seti project have spurred demand, growth rates can only accelerate if new applications of this type emerge every few months.
Weak overseas demand?
Outside North America, where much of the current fiber investment is concentrated, demand patterns look even weaker.
FLAG says its three-year old Europe to Asia 2x5 Gbps cable is just 25% filled. The cable is still nearly 30% short of paying for its capital cost.
Demand has been higher on its Hong Kong to Japan leg, but Level 3’s recent deployment there has multiplied the level of available bandwidth on that route by a factor of some 800%.
Asia Global Crossing admits that it has sold just 3% of its new cable, although it quickly adds that this has paid back 18% of the buildout cost. Global Crossing says it is the dominant Internet backbone provider in Japan, which in turn is the second largest Internet market in the world. The size of Global Crossing’s Japan to United States IP backbone? Just 2.5 Gbps.
According to AT&T’s Odlyzko, Australia’s Telstra actually reduced its US link from 592 Mbps to 515 Mbps last year — an occurrence he attributes to greater deployment of local peering infrastructure.
America’s Network sister magazine Telecom Asia recently conducted a survey of 1,400 Asian corporate users across Asia’s 12 biggest economies to establish their IP deployment plans. On average, they buy less than 256 kbps each for their total external bandwidth requirement and plan to increase this by 50% on average next year. The 1,400 in total currently buy less than one gigabit of collective connectivity and most of that is sourced for their home markets.
Vaporcable
This lack of real demand has led some to predict that not all planned systems will be built. Asia Global Crossing CEO John Legere says, “If you add up all of these announcements, it’s said to be a glut, but they have to get built first, and a lot of these people aren’t going to put down their checks.”
“It’s true that a few of these cables won’t get off the drawing board,” agrees Neil Lambert, chief marketing officer with the Australia-Japan Cable, a Bermuda-based company affiliated with the new Telstra-Pacific Cyberworks backbone venture.
“But it won’t be the market that trips them up,” he predicts. “Most of them will trip themselves up with funding and getting organized. They’ve got to be wary about how they spend their money.”
It’s clear that financiers have turned against the sector. The most high-profile casualty was Project Oxygen, an ambitious system that originally planned to provide multi-gigabit connections to virtually every country on Earth. Headed by FLAG founder Neil Tagare, the 1997 announcement of Oxygen promised a then-novel distance-independent pricing scheme.
But the project struggled to obtain financing and was repeatedly scaled back. Early last year, it fizzled out without even an announcement. Even by the hyperbolic standards of the global submarine cable sector, its ambitions were excessively idealistic, with a business case built around the idea that bandwidth would liberate developing economies from their poverty. It even promised to refund early buyers with its profits!
To its detriment, much of the backbone sector seems driven by the urge to come up with grandiose capacity and rollout targets. Despite Wall Street’s concerns about a bandwidth glut, operators continue to tout their exponential capacities and reliance on high-risk revenue models.
Williams Communications, for example, has much to be proud about, claiming 177 carrier customers for a 31,000 mile fiber network that it says is a year ahead of schedule.
So why then did it choose to publicise just one specific carrier customer win in the past few months — an OTC-listed carrier named Hop-On.Com, which describes itself as a “diversified B2B, SOHO and B2C Internet Convergence operator that offers [users] free DSL unlimited Internet access when they use Hop-On as their long distance telephone carrier?”
In claiming a $34 million, two-year deal with Hop-On.Com, which without any sense of irony uses a kangaroo as a logo, Williams chose to highlight the very sort of customer that Wall Street fears.
Another operator that doesn’t see the reality of current sentiment is 360networks. In an announcement about its 360Americas cable connecting North and Latin America, it made the rather pointless comparison that its 1.28-terabit cable could transmit 1.3 million digital photographs per second — a demand point unlikely to ever be tested in the forseeable future.
Survivors of the shakeout
So who will survive the inevitable backbone operator shakeout? Global Crossing seems the most strongly placed of the new backbone operators. Legere says the company is differentiated because it provides IP backhaul in each city it serves with its undersea cable. “Besides, we get cash upfront, so we also have no bad debt and no collection cycle,” he adds.
Credit Suisse First Boston, in a note to clients on December 13, rates Global Crossing and 360networks as the most competitive in the sector, based on their network reach, provisioning capabilities, quality of support and ability to compete on price.
It places Broadwing, Level 3, Qwest International and Williams in the next tier on the basis of their relatively smaller reach. “We believe that the sweet spot for these carriers is more apt to be the emerging xSPs as well as CLECs and cable companies,” says CSFB.
The most vulnerable operators, according to CSFB, are regional operators with relatively limited reach such as Viatel, Hermes and FLAG Telecom. Viatel’s stock valuation has been knocked down 95% in the past 12 months to the $100 million range, even though it is building a 64-city network. Global Telesystems has experienced a similar tumble, as its Esprit Telecom subsidiary defaults on bond coupons.
Of course, this scenario could be reversed. The larger systems, with their greater exposure to risk, may be more vulnerable to softer demand in Asian and Latin American markets. Those that focus on fewer higher-yielding markets may obtain quicker returns and win favor with the market as a result.
Level 3’s International CEO Colin Williams sums up the feelings on many in the industry: “The risks are absolutely huge. Financial markets will not get a lot of the billions back. We do not have a way to understand how demand is growing. We will have significant overcapacity.”
Williams also expects vendors to take financial losses. “Basically, technology is exceding the rate of growth. Our two submarine cable vendors [Alcatel and Tyco] will lose a lot of money.”
— with reporting by John C. Tanner and Tim Marshall |
