Hang in there Tim;)
Solving the bandwidth bottleneck has become the Holy Grail for both established companies and startups.
Silicon Spice is not the newest member of the hit British supergroup, Spice Girls. Nevertheless, this Mountain View, Calif. startup is working on chip technology, which has the potential of becoming a big smash.
In the hype-driven world of technology startups, Silicon Spice is a breath of fresh air. The two-year-old company does not want any media spotlight, and refuses to divulge any details about its chip technology--except that it is working on a communications chip that will solve the bandwidth problems plaguing the Internet.
Why the mystery? "We would like our product to speak for us," says Vinod Dham, president and CEO of the company and a chip-industry veteran who joined the company in April 1998, as chief executive.
Dham, who worked on the development of the K6 chip for Advanced Micro Devices and the Pentium chip for Intel in his previous jobs, is known to bet on winning projects. And he is confident that Silicon Spice will be a big winner. (See "Dham good?")
The technology may be under a cone of silence, but the company's creation is old hat. Three entrepreneurs, Ian Eslick, Ethan Mirsky and Rob French, all engineering graduates, started Silicon Spice literally in Eslick's garage in Palo Alto, Calif. in September 1996.
Solving the bandwidth bottleneck has become the Holy Grail for both established companies and startups. While one option has been to increase the size of the pipe that carries the data, by replacing the copper cables currently in place with fiber-optic cables, many find it prohibitively expensive. The smarter way, some industry pundits contend, is using smart chips to control the way data is transmitted.
Thus far, these smart chips have been placed in networking hardware like switches and routers. MMC Networks (MMCN), Applied MicroCircuits (AMCC), PMC-Sierra (PMCS) and Vitesse (VTSS) are some of the companies that make these complex chips for the networking hardware.
While the company will not give out any details, a look back into the research conducted at MIT and Stanford by its three cofounders points to programmable communications processors. Linley Gwennap, the editor of respected chip industry newsletter, The Microprocessor Report, shares this view: "They [Silicon Spice] are working on a new signal processing architecture around a high-performance programmable communications chip."
Gwennap, normally stingy with praise, is impressed by this approach, which makes such a chip easily programmable for all and any types of Digital Subscriber Line (DSL) technologies. DSL is a modem technology that boosts the digital speed of ordinary copper telephone lines, and is already splintered into many competing formats like asymmetrical DSL (ADSL) or rate adaptive DSL (RDSL) or high bit rate DSL (HDSL).
"Our chip sits between the network and the PC and reduces the bandwidth bottleneck," says Dham. "Moore's Law has run into Moron's Law, where old monopolies like the Regional Bell Companies have not kept up with Moore's Law." Gordon Moore, Intel cofounder and billionaire, in 1965 proposed that the number of transistors crammed onto a piece of silicon every 24 months doubles, thereby increasing the power of the chip. He later changed his estimates to 18 months, and that proposal is now widely known as Moore's Law.
Silicon Spice has many believers, like Dham, and has managed to attract some major semiconductor talent. The latest to join the company is John Nikolls, who used to oversee SPARC development efforts at Sun Microsystems.
The founding trio's concept intrigued Rob Ryan, one of the four cofounders of networking giant Ascend Communications. Ryan helped seed Silicon Spice and since then has helped it get venture capital. "They seem to be well connected with the VC community, which means a lot of people think that they are doing the right thing," says Gwennap.
In March of 1997, Silicon Spice received a first-round funding of $3.3 million from an all-star VC lineup that included New Enterprise Associates (NEA) and World View Technology Partners. Kleiner, Perkins, Caufield & Beyers pumped in another $7 million in the second round in April of this year.
While the company has yet to book any sales, it is targeting Internet service providers, telcos, modem and set-top box makers for its signal processing products. Assuming that its product is as good as the founders claim, the company's target market offers good potential. Semico Research, a Phoenix, Ariz.-based research firm, predicts that the demand for the communications chips will increase 20% every year through 2001, a market which generated about $16 billion in sales in 1997.
With funding in place, Silicon Spice is hoping to unveil its product in early 1999, and get into volume production by year 2000. Perhaps by then the mystery will be solved, but in the meantime, it would be better to withhold judgment on whether or not the company will be a smash or a dud.
|
