Skip to main content

Cisco Slips in ATM Switch Market

The telecom-equipment leader has failed to develop a new ATM switch, which some say will put it behind competitors.

SAN FRANCISCO -- Uncharacteristically, Cisco (CSCO) - Get Free Report doesn't seem to have the answer to the growth of a pivotal network technology.

Cisco has failed to develop a new "asynchronous transfer mode" switch, which telephone carriers are installing on their networks to direct long boxcars of Internet traffic. ATM switches are more dependable than the "Internet protocol" systems Cisco is promoting, and so far seem the best way to combine voice and data signals. Researcher

Dell'Oro Group

says ATM sales grew 49% to $3 billion last year.

Now Cisco, which is expected next week to report strong third-quarter earnings and revenue growth, finds itself lagging further behind competitors





(ASND) - Get Free Report

(set to merge with


(LU) - Get Free Report

by late June). The divergence threatens to erode Cisco's market share in the carrier ATM business, which according to Dell'Oro fell to 19% in the fourth quarter from 23% in the first quarter of 1998. In the same period Nortel's and Ascend's market shares expanded to 25% from 18% and to 23% from 18%, respectively.

So Cisco is in a bind, and its usual crisis tactics might not do their magic this time. Typically the San Jose, Calif.-based company responds to market shifts by acquiring technology or applies the marketing charm to its

Fortune 1,000

clients. Those maneuvers, coupled with superb managerial skills, explain how Cisco's sales have swelled to $10 billion over the last four quarters and its market cap has ballooned to $174 billion since the company booked its first sale in 1986.

This crisis is different. Cisco's new target clientele, carriers such as


(T) - Get Free Report




, boasts closer ties to the competition. And there are few ATM companies available for Cisco to acquire.

Cisco has canceled development of an ATM switch called the TGX 8750, widely viewed as a crucial product for competing with Nortel and Ascend. Instead, it intends to incorporate TGX features, including wide optical-fiber connections, into an older 8850 product line and re-release it early next year.

Cisco won't say when it made the decision. The company discreetly noted the product change on its Web site in early April, but didn't make executives available to discuss the move, which was reported Monday by the trade publication

Network World

. The stock has slipped mildly this week along with the


, to 106 15/16 Thursday from 114 1/16 last Friday. Cisco is down nearly 10% from a 52-week high of 118 3/4 on April 7; it now trades at 130 times trailing four-quarter earnings.

Next Tuesday Cisco is expected to announce a 23% increase in operating profits for the fiscal third quarter ended in April, and should even report growth in its business with carriers. Equity analysts are standing pat with their buy recommendations and their bullish estimates. But Cisco's failure to deliver such an important product raises unfamiliar doubts as the pros study their favorite stock.

"It shouldn't just be brushed under the carpet," says one analyst, who rates the stock a buy and did not want to be identified. His firm has no banking ties to Cisco. A year from now, "if you don't have a product and you're not there soon enough, you can't win the business."

In the past, Cisco has successfully adjusted to shifts in customer tastes. In 1993, Cisco spent $90 million stock buying


, a 60-employee start-up in Sunnyvale, Calif., branching from network routers into the emerging switch business and juicing its revenue growth for several years.

As it grew more powerful, Cisco used marketing to cover up bad moves: It promoted the need for "universal standards" to buy a year and a half of time to integrate the ill-fitting technology of

Granite Systems

into its product line. The marketing spin was "classic Cisco," says analyst Craig Johnson with the

Pita Group

consulting firm, who owns Cisco shares.

This time Cisco is in a tighter spot. For one thing, marketing is more challenging, because it is trying to court business with carriers which know Nortel and Lucent much better than they know Cisco. And while Cisco struggles with ATM, Nortel and Ascend already deliver strong ATM offerings.

So winning significant contracts remains difficult. A month ago, Cisco announced plans to furnish AT&T with so-called IP+ATM switches; however, AT&T says it still primarily uses ATM equipment from Ascend. AT&T expects to grow IP traffic in coming years, but for now still needs Ascend ATM gear.

"They're walking into places where they are not the established vendor," says one industry analyst who has worked for a Cisco rival. "They are seeing competition like they haven't seen before."

Cisco has delayed other ATM products. Several analysts say the company was six months late in furnishing Sprint with a switch from its 8800 product line for Sprint's ION network, designed to meld voice and data traffic. Sprint confirms this, although an official says it fully expected the delay and is proceeding with construction as planned. Sprint activated certain ION services early this year and has signed 12 corporations in 26 cities for the service. Cisco says that, despite certain shifts in timing, it delivered the equipment on schedule.

"The problem they have right now is credibility in the product road map," says Rosemary Cochran, principal with

Vertical Systems Group

, a market researcher in Dedham, Mass.

Cisco also faces greater challenges this time because there are fewer acquisition candidates to fill its ATM hole. Ascend gobbled ATM concern


in mid-1997. Last week,

Fore Systems


, a prime ATM supplier, agreed to merge with


of Britain for $4.5 billion in cash.

While ATM company

Newbridge Networks


of Kanata, Ontario, remains independent, it has had trouble executing. Newbridge, with revenue of $1.7 billion and a market value of $5.2 billion, is much larger than any company Cisco has ever acquired. The private start-ups




are testing routers that handle ATM, but remain focused on Internet protocol. Cisco defines itself as an IP leader.

"It would be very hard for Cisco to admit they have to buy an IP company," says the industry analyst.

The TGX snafu all started with an acquisition gone bad. Three years ago, Cisco bought


, a builder of ATM switches, for $4 billion in stock. Cisco has done little to upgrade StrataCom's BPX switches since then. The TGX was supposed to restore StrataCom's worth.

Speaking in his offices Monday -- before Cisco stopped answering


inquiries about the TGX -- a marketing vice president, Larry Lang, said he was surprised that carriers and ISPs still counted for only one-third of Cisco's revenue. He says that is because corporations, historically Cisco's strength, have continued to generate strong sales.

Given Cisco's frustrations with carriers, that route might be the company's best hope for keeping its 30% growth intact.