DSL Prime Newsletter: December 25, 2000

DSL Prime Newsletter

By Dave Burstein
DSL Prime

February 27, 2001

Alcatel's own Voice over DSL
Gateway on a 7300 DSLAM card
With a little help from friend TDSoft, Alcatel later this year will deliver a voice over DSL gateway. according to an presentation at a European conference. First targeted at Europe, the goal is to deliver distributed VoDSL right in the telco CO, feeding a local switch. (In the US, Alcatel will continue to work with CopperCom for now, and is in advanced trials at telcos.) Alcatel last year was considering buying a gateway vendor (CopperCom or Jetstream) but the price was too high, so they chose to develop their own product.

The 7300 is a watershed for the industry, a true "next-generation DSLAM" now becoming the standard unit in the major telcos, where Alcatel has two thirds of the world market. A video grade unit with gigabits of backplane speed, processors and control systems capable of IP quality of service and multicasting, optional test capabilities, and interfaces soon to reach OC-12 and beyond.

What the heck is going on at Covad?
"We are in no danger of Chapter 11 at all"
Covad's Chuck McMinn and team built an exciting company, but this second financial crisis implies a tough road ahead. Covad's writing off at least another $50M - and probably much more - when they finish closing their books this quarter. Part of the writeoff is an accounting rule (if a modem is provided as part of a three year contract, the modem income has to be spread over three years.) Another part is Marketing funds advanced to ISPs that have not paid off, and a large part is additional ISP debt, most of which we thought had been dealt with months ago.

They are closing 300 of 1900 COs, extending layoffs, and returning trucks to the lessor. They are essentially writing off much of the Bluestar deal. A few days earlier, Covad played hardball with ISPs, shutting off thousand of customers of Internet Express and DSL Networks, creating permanent fear among customers, terrible PR, and convincing us, before the financial announcement, that the problems were much worse than reported.

If Bob Knowling hadn't pushed through the last $500M financing just before he left, the situation would be dire; with the cash on hand, Covad has many months to turn things around, and we hope they do. Fortunately, AT&T, XO, SBC, and others able to pay their bills are bringing in most of the new customers, and Covad reached a businesslike settlement with Flashcom in court.

NorthPoint for sale
Bids coming in for late March
Verizon suit advancing
There are many rumors - but little solid information - about the bidders for NorthPoint, in a process leading up to a court date in late March, the hoped for (but not guaranteed) final resolution. One probably out of the running is NTT/Verio, which is pulling away from DSL. (Will NTT have to take a write-off on their $5.5B purchase of Verio, now clearly worth less than half that?)

Meanwhile, they've won a ruling their $B plus law suit against Verizon can proceed in California, and recent events have only improved their case. NorthPoint should win the suit if the "material adverse events" were caused by "facts, events, changes or effects that are generally applicable to (A) the data industry, (B) the United States economy or (C) the United States securities markets generally or the Nasdaq Technology Index in particular". Clearly, problems in the data industry played a major role.

Lucent, Cisco, and Nortel have also suffered shortfalls that NorthPoint's lawyers will attribute to "problems in the data industry" - an argument whose merits may well remain unclear until years from now in a jury trial. Meanwhile, we're watching Verizon's financial statements. They presumably will have a lawyer's opinion that minimizes their material risk, but the lawyers we've consulted are not sure at all, and a reserve (or a settlement) would be prudent.

Nortel cutbacks - we got the story right by accident
"CLEC strategy" failed, as Nortel's layoffs demonstrate
We were just referring to DSLAM future sales, but several misread our comment suggesting deeper Nortel trouble, which we were not predicting. We have enough trouble tracking DSL. Turns out, the same reasoning applied to the larger company as well, but we hadn't predicted it.

John Roth of Nortel is a smart and honest man, who has earned the industry's respect. We believed him when he told five hundred wall streeters that Nortel would have an exceptional 2001, growing 30%. Last Thursday, he had to accept that he was wrong, and laid off 10,000 employees. The fall speech itself contained enough to anticipate the problems, as we now see in hindsight. He emphasized sales to CLECs, but told us Nortel would no longer finance them. Since CLECs worldwide are short of capital, he was being unrealistic expecting major growth. Apparently also unrealistic were his growth goals for optical networking, the one area the bubble hadn't burst in. Nortel is the last of the telecom giants to accept the problems, which originate in the capital expense cutbacks at telcos around the world.

Corporate problems, unfortunately, overshadow the remarkable fourth quarter of 2000, and Nortel has now shipped over 600,000 lines of the Promatory DSLAM. Korea Telecom has become a major customer, and they are also shipping to @Link, First Mark, and PLDT. Tokyo Metallic gives them an enormous upside in Japan. Brett Sheppard believes they had 10% of the world market last quarter, and will hold or increase that percentage this year.

What Wall Street knows about telecom equipment
For the last six months, Wall Street has seen projections from telco after telco that capital spending will be relatively flat or down for 2001 and 2002. Qwest, Verizon, SBC (and of course all the competitors) geared up for data and wireless, and are unwilling to up the rate of investment. The result, of course, is equipment sales, especially in the US, also cannot rise rapidly. While individual companies may buck the trend, it was impossible for most major companies to increase sales 20-30%, as many hoped. One by one, sales projections have come down, from Lucent, Cisco, ADC, and now Nortel. Most were inevitable - you can't sell what they won't buy.

Covad: making line-sharing work
In nearly 100% of 1600 surviving offices
Linesharing cuts monthly line costs from $10-20 to about $3, and is absolutely essential if there is ever to be any competition for consumer DSL. The FCC ordered it for last June, and after delays it is working well in much of the country, with lines delivered to competitors in less than 5 days in many cases. Because it uses existing lines the problem rate is as much as 70% lower, dramatically driving down the operational cost.

Covad hoped to be this far along in December, but some areas dragged to February. As they train ISPs and restructure agreements, they are also moving to standard customer install for nearly all consumer sales this spring. These kinds of day-to-day improvements mean the industry's long-term future is bright - despite the painful stories we have to report.

Copper Mountain's cash gives them time
Revenues shrink dramatically, and future unpredictable
When times got bad for the industry, a common comment was "The good companies will do fine." This was clearly mistaken. Copper Mountain has delivered some of the most reliable DSLAMs in the world, and given customers (and reporters) excellent support. They've been fiscally prudent, and earned a profit. Their partner program was so successful, the rest of the industry copied it.

3Com just sold several million dollars worth of Copper Mountain DSLAMs in an OEM deal, and mPower expanded their purchasing. But the problems of customers like Rhythms, NorthPoint, and JATO cost them much of the quarter's revenue, which McLeod and basement DSLAM orders couldn't compensate for. NorthPoint was acknowledged as the source of the $8M write-off, and inventory/work-in-progress in the tens of millions must be written off. They now need to build international sales, where the market is growing twice as fast as in the US. Paradyne, facing similar problems, has already cut back staff and spending, and is emphasizing international sales - Copper Mountain's conservative financing will give them breathing room to do likewise, and bring a nextgen unit to market.

TdSoft $20M, Arrival $17M, MegaPath $10M
There's still money out there
TdSoft is the early leader in delivering European 5.2 VoDSL, with enormous potential and a deal with Alcatel. Because Alcatel ships more than half the DSLAMs in the world, that deal is transforming the VoDSL market. Founder and CTO Eytan Radian now is building the company in the US - he's one of the most colorful characters in the industry.

Arrival is building a network in California, emphasizing business connections in the areas not well served by Pac Bell. Many feel the industry survivors will be led by companies like Arrival, @Link, New Edge, and IP Communications, if they emerge as the strongest active competitors to the telcos in selected COs.

MegaPath is a business-oriented ISP, where Harry Taxin has emphasized high-quality service from the first, and MegaPath ratings on DSL Reports have always been high.

ConnectSouth: another one bites the dust
Staff gone, operations expected to close
"Another one bites the dust" is our headline, to make the point that industry problems are overwhelming many companies with decent management and hard-working staff. We know the pain of the people whose dreams were denied is very real; some will lose their homes and health insurance. Good companies, not just weak ones, fail in this market, and everyone looks for a scapegoat. But Bob Knowling of Covad, for example, was not a brilliant manager in April and a rube in October. Rather, he was a smart man brought down by industry changes. He's happy to talk, with hindsight, about mistakes he made - but we've yet to meet anyone perfect.

Morgan Stanley funded CLECs in different regions, hoping to duplicate the remarkable early success of Rhythms, Covad, NorthPoint and DSL.net. eAccess in Japan is now growing rapidly, and we hope proves a prudent investment. But nearly a half dozen groups had a similar idea in the US south and southwest, and the result is that all are in trouble. Before the general CLEC collapse, Bluestar's IPO failed and their VCs were temporarily bailed out by Covad (another portfolio company.) Vectris and Darwin followed, and now ConnectSouth, when the next round of funding fell through.

New Edge: we're cutting back
55 jobs lost
Dan Moffat's company is pegged as one of the survivors, but that means he has to make tough decisions along the way. With no more capital in sight, he had to drop 55 more workers (second layoff) as well as temporarily shut down some locations in Florida and Georgia, reducing COs to 500. He's also had to turn down very attractive offers to take over others, including Vitts and ConnectSouth; a very smart banker should find a way to provide him the funding to take advantage of opportunities like that.

In the works is an editorial "Seven realistic steps to increase competition, bring down prices and move toward universal service" One of them will be to bring the price of fiber backbone connections down in rural areas, the main reason Covad, New Edge, and others are shutting down small offices. The marginal cost of a fiber-optic lambda is dramatically less than the fees charged; if there is no competition, then creative unbundling or regulation should be used to reduce monopoly-like prices. Congressmen Tauzin, can you deliver that to your small town constituents?

Siemens takes Efficient for $1.5B
Also investing 100M for new DSLAM plant
Mark Floyd gave us the good news "We plan no layoffs, and I as President of Siemens ICN Access solutions expect to continue to lead and grow the company." Besides a major SBC contract, he's expanding in China and Brazil, where Siemens international experience will help. Floyd added "urban" China is apartment buildings, allowing rapid growth as we've seen in Korea. Telefonica in Brazil looks to become a major customer." With Deutsche Telecom aiming at millions of DSL lines, Siemens is preparing to serve - and grow worldwide. Efficient is the leader in modems, and their Flowpoint division was critical in early tests of voice over DSL. The bid is 70% above recent trading, but only a small fraction of last year's value.

Briefs:
* Where will the future demand come from? The mantra remains "we have more orders than we can handle", and the installation delays would seem to confirm that. SBC, Verizon and most CLECs still take over a month for most installs, compared to a telco (and cable industry) standard of 3 days, suggesting a backlog. But we believe that overall installation capacity has caught up with demand, with SBC, as we will report next issue, with the capacity to handle 500,000 more orders than they desire in 2001. Similarly, looking closely at order backlogs at CLECs Covad and Rhythms, install backlogs are stable to down. The easy early adapters have been reached in the major cities; we're now in the "post-DSLAM" era, needing DLCs and loop extension to expand the market, and service quality that will inspire customers to want DSL.
* Andrew Odlyzko of AT&T has convinced us we were wrong in our forecasts of bandwidth demand, a key determinant of the equipment you need to purchase today. He's found substantial data that suggest the widely quoted "doubling every 120 days" growth of the internet is overstated by half or more. That's what we've been hearing reflected in network demand within DSL as well. Even doubling over a year will make typical DSLAMs installed in 2000 obsolete in 4-5 years, but that's better than the 2-3 years we have been projecting.
* 540,000 DSL customers was the first highlight of Verizon quarterly financials, 30% below their spring goal but ahead of the 500K they expected in the fall.
* Motorola is splitting its broadband division (the old General Instrument) into two parts, entertainment and networking. Their Streamaster set top is already in use in the Enron/Verizon trials, and they intend to offer DSL versions of the gateways and intelligent home controllers they are selling to cable
* Fiber in the sewer, anyone? Citynet's robots are now at work in Albuquerque, deploying cable.

Deals:
* NHC won a large order for equipment from BellSouth. Like most telco deals, this was a long time in the works, but the potential volume is significant. With 5 North American telcos representing about 80% of the market for DSL equipment, each decision is crucial.
* BroadJump signed SBC, and is close to dominating the market for broadband installation software, with several cable providers on board as well.
* Alcatel will OEM basement DSLAM equipment from RC Networks, which has produced equipment that works well in the field and is priced far below the market. They have $3,000 units that make ideal building starters, while the more exciting products, just coming to market, include voice.
* Virtual Access has won orders from BT and BellSouth for routers, and is optimistic in other telcos. A simple to use test mode, with a graphical user interface, helps end-users and providers determine where the problem is - local machine, local connection, DSLAM, etc. Pete Castleton of Verizon told us last month that the ability to test is proving critical to network reliability, and is a key factor in their equipment decisions.

Products:
* DVTel announced a new Video over DSL system, using chips from Infineon.

Chips:
* TI has been especially aggressive in pricing, competitors claim, but while DSL is strategic to them, the dollar volume is probably not material to their earnings. DSLAM manufacturers (except Alcatel and Siemens) have all dropped their forecasts for 2001 - that's lower growth, not an absolute downturn.
* LSI Logic introduced an 8 port analog front end chip, with good numbers for noise and power.
* Tripath cut a deal for Alcatel's support of their low-powered line driver, one of the first of a new generation that will raise DSLAM port density.
* Legerity (the AMD spinoff) "opened the kimono" at ISSSC with technical details on their high voltage chips that eliminate splitters and combine voice and data.
* Analog Devices, on the other hand, is working to eliminate the line driver entirely, combining it with the analog front end.International
* "There is no consumer choice for DSL in the UK", Kingston said as they deferred their service till at least next year, blaming BT. All the CLECs are looking at business. Thus, even with the backing of Scottish Power, has had to abandon DSL plans.
* KPNQwest's Jerry Yohananov e-mailed that DSL "is not a big part of our future strategy." They've cut back from 1600 to 300 planned COs.

Competition:
* Half the companies at Cannes 3GSM conference will be out of business in a year, predicted VC Olav Ostin. Highspeed wireless handsets and customers will be very rare until much later than predicted.

People:
* Maida Chicon now directs Verizon's multicultural marketing organization, the kind of outreach lesser firms will need to develop as the surge of "early enthusiasm" passes.
* Harold Furchgott-Roth, the FCC commissioner without a television set, has resigned, because "every free-market advocate in government must fulfill his dream of returning to the private sector." Bush now gets to name his replacement, as well as Bill Kennard's.
* Kelvin Ko is the new Asia director for Videotele.com, the Tektronix video server spinoff, which has promised dramatic breakthrus in price/performance for video headends. DSL is barely visible in China, but Videotele's three offices presumably are calling on many prospects. Video growth is slower than expected, because the big telcos are scared of the investment required, after massive losses six years ago. But as costs keep coming down, (and DSL Internet traffic pays much of the infrastructure cost), the economic argument keeps getting better. We think the time is now, although many disagree; we are certain the economics will become overwhelming within a few years, which means every DSLAM or DLC purchased today without sufficient growth capability will prove a mistake over time. There's less than a 7% difference in cost between video-capable DSLAMs and obsolete ones; in the case of the Alcatel 7300, the new model is cheaper, per port, than the old one, but has 10 times the potential capacity.

Stock Market:
* Netopia's takeover by Proxim illustrated the downside of having cash - cash makes you very attractive to a buyer offering stock. But "fully funded until breakeven" is the mantra of today's successful net company. That's one reason Wall Street is abandoning Amazon; their cash runs out late this year, with no profit in sight.
* Lucent's downgrade to near junk bond status is a frightening reminder of how fragile the Internet and telecom boom really was. The DSL division continues to do well, with growth expected in 2001. The SEC investigation is surprising, but is another warning audited statements are not always accurate, especially as things change. Management hoped that DSL, like optical, would be spared layoffs, but the corporate problems are now so severe nothing can be guaranteed.
* Several analysts asked us about the impact of rapidly falling DSL chip prices on stock prices, which we are reluctant to forecast. Virata seems the least effected, as they emphasize features beyond the standard, but no one else is likely to be immune from an earnings hit, especially in Q2. (Since many chips are sold under contract, spot prices don't always have an immediate effect.) But ADI, Aware, and others have already seen severe stock drops, so don't read this as a sell recommendation, please. (In fact, since we wrote this several DSL chip companies have seen their stock rise.)