In the race to connect the world to the Internet, a handful of companies dominate. Now come a handful of startups to challenge the cozy status quo.
The rise of the Internet has put phone companies in a bind. Carriers, both old and young, have invested in dependable electronic circuit networks, and they want to eke out the most use out from them. That's fine with suppliers like
who, while trumpeting Internet plans, are still quietly nursing sales of their age-old voice switches to lumbering carriers like
. On the other side, computer networker
is pushing phone carriers to transfer voice calls onto the Internet. Cisco sees the telephone infrastructure as nothing more than "old world" technology.
Eventually carriers intend to combine their phone systems with the Internet. How will they make the transition?
Several new startups propose a solution that could shake up the established powers: hybrid switches that not only cut costs but work snugly with telephone infrastructure, rather than replacing it, to combine advanced new data and voice services. That's an "absolutely logical product" for Cisco or Lucent to build, notes Barry Fidelman, partner with the venture capitalist
But Cisco engineers are still learning voice technology, and Lucent has its work cut out trying to finish acquiring
. That might explain why Fidelman has planted money in Westford, Mass.-based
, which already has carriers testing its switch. Castle is not alone -- its peers
are preparing similar products.
All five startups, many with almost zero revenue to date, are trying to make capital out of the confusion of carriers that want to branch into Internet services but are reluctant to junk their current systems. The startups are expected to ship some of their products commercially as early as this summer. One of their suggested features will allow a customer's car, home and office to be united with a single phone number.
The most promising of the group is Castle, founded in 1997 by veterans of networkers such as
, now an arm of Ascend. With plenty of influential backers, industry consultant Dan Taylor of
Taylor & Associates
says, the rest of group should keep large rivals "on their toes."
Taqua, based in Centerville, Mass., debuts its test switch Monday, while Castle showed off its technology to potential investors at a conference earlier last month.
"Voice is the dominant technology. It is today and will continue to be," says Castle vice president Michael Welts. Circuit infrastructure that cost billions of dollars is unlikely to be replaced anytime soon. "Sorry it's not sexy," echoes David Michaud, founder of Taqua. Michaud doesn't see phone companies ripping out their current technology any time soon. Taqua systems operate on a different part of the network and would complement offerings from Castle.
Some carriers are showing interest.
, a Rochester, N.Y.-based telephone company with $2.6 billion in annual revenue, has already held talks with Castle, Sonus (also based in Westford, Mass.) and Transmedia (based in San Jose, Calif.) about testing their switches. Frontier is also interested in investing in them, according to Jonathan Heiliger, senior vice president of Frontier's new venture-capital group. Currently, Nortel switches run Frontier's voice traffic, and Cisco routers handle data. Frontier wants to merge some of its voice and data traffic later this year, and might employ hybrid switches from one of the startups. Heiliger says no decision has been made about working with the startups.
To be sure, telephone carriers might use these startups as "pawns" to extract price concessions from their old suppliers, warns Hilary Mine, vice president of
This all adds up to a competitive threat to Cisco, which has the money and the tendency to acquire its problems.
"There's no question
that they should buy one of these companies" later this year, says analyst Paul Johnson with
BancBoston Robertson Stephens
. Cisco engineers lack expertise in voice technology, he says.
Cisco declined to comment on potential rivals.
Deb Mielke, principal with
Treillage Network Strategies
, an adviser to telecom companies, notes that Cisco hasn't produced a strong transition switch yet and that Lucent is busy digesting Ascend (Mielke declined to name her clients). Dan Simpkins, CEO and founder of Gaithersburg, Md.-based Salix, says his larger rivals are "machines that have a huge amount of inertia" and might be caught off guard by an upstart.
If these companies manage to go public, then they'll be able to attract financing to fund more research. Executives at the firms don't rule out one option or another. Salix, for example, fashions itself after companies like Cascade, now part of Ascend, and
, which Lucent acquired last spring. But right now Salix's founder Dan Simpkins says those questions can wait -- he is too focused on showing up his large rivals by unveiling a superior product.
But Lucent and Cisco will play hardball to maintain their position. Lucent says it has cut product development time by 30% in the last year to keep up with the fast-moving trends of the Internet. Jim Bodycomb, Lucent vice president of product management, intends to keep upgrading Lucent's flagship switch, which is installed in networks worldwide, for data services. He declined to speak about particular competitors.
Nortel just unveiled a "Succession" network that adapts its switches for Internet Protocol.
Another pressure for the startups is that Wall Street has little tolerance for short-lived technology. Two years ago, upstart
beat the big suppliers with a new optical fiber system. For a while Ciena was perceived as a major threat. Kathy Szelag, Lucent's vice president of marketing in optical networking, notes that Ciena's "creative energies" prompted Lucent to accelerate its optical development.
Ciena's leadership only lasted a short while bigger rivals rebounded with price cuts and broader offerings.
Last summer Ciena lost prospective business with
to Lucent, encouraging
to cancel its merger with Ciena. Ciena profits fell 94% in a year, and despite a sharp bounce in recent months, the stock is down 70% from last summer's highs. Another example: In 1996
introduced a new "gateway" that links corporate phones to the Internet, only to see big guys like Cisco absorb the market by building gateways into their existing product lines.
The big guns can use a tactic often employed by
-- they promise customers they'll have a product long before it's available. If they don't make the deadline, they can just buy the startup.
Given the track record of Lucent and Cisco, such an option may not be out of the question.