|
While manufacturers of broadband equipment race to differentiate themselves in an increasingly standardized world, suppliers of under-the-hood silicon are even busier, ruled by Gordon Moore's law to double performance every 18 months while keeping the cost-line steady, or, preferably, lower.
Broadband's big chipmakers-Broadcom Corp., Texas Instruments, and Conexant Systems Corp.-sometimes are overlooked in the go-go pace of telecommunications and entertainment equipment sales. But in reality, it is the silicon leaders that are paving the path of the future.
Across equipment lines-in cable and DSL modems, set-top boxes, residential gateways and other residential Internet Protocol-based gear-the name of the game for chip makers perennially is to add more features, usually months or even years before the end devices that use them are put into the field.
Take, for instance, home networking. There are at least four in-home distribution possibilities: Phone wires, nested in walls; cable wires, located near TVs; power wires, also inside walls, and wireless. Chipmakers are faced with two choices: Wait for intra- and inter-industry momentum toward one of them, or build as many types as possible right onto the silica.
Most are doing the latter, blending techniques such as Home PNA 1.0/2.0, 802.11-B, Bluetooth, and others onto chip sets. The reasoning: Predicting the winner is risky, so assume all will win.
Simultaneously, the name of the game in silicon is chip consolidation. Cramming more capability into fewer chips is what eases the cost line for manufacturers. In May, Pace Micro Technologies, the British manufacturer of digital set-tops making inroads with Time Warner Cable and other cable operators, was the first supplier to use a single-chip design made by Broadcom. In cable modem chip sets, most silicon makers are down to a single chip, too.
For chipmakers, it's a nearly illogical mandate: Squeeze more stuff into less space, while driving costs down, and in a brutally competitive market. That partly explains a spending spree by several silicon leaders in 2000. Broadcom, which started out as a maker of demodulation chips for cable's digital set-top boxes, released an aggregate 45.7 million shares of its stock for a dozen acquisitions in 2000, spending $7.5 billion of equity value (based on closing prices at deal times). It's reasoning: Time-to-market pressure made it quicker to buy talent, than to develop it in-house.
The question for Broadcom, Conexant and other heavy buyers in 2000 is whether or not they can digest the 2000 buying binge while they're at a flat-out sprint to stay ahead of their customers'-the manufacturers-needs.
Broadcom's competitors and detractors cite its spot as the market-share leader in cable modem and digital set-top silicon as the sliding-off point, saying the silicon powerhouse can only go in one direction: Down. TI submits that its work in furnishing the inside of cable modems made by Toshiba and others is already making a dent, claiming 600% growth from year-end 1999 to an estimated 30% percent foothold in worldwide cable modem chipset sales in 2000.
But executives familiar with Broadcom's leadership submit that the company's leaders, and in particular its CEO, Henry Nicholas, "has the stamina of 50 men." Nicholas himself says that each of Broadcom's 17 acquisitions since inception were carefully orchestrated to be self-integrating. That means Broadcom partners with its potential children before it buys them-try before buy-which sets up an atmosphere where chip integration with acquirees usually is in progress at the time of purchase.
Broadcom's corporate goal makes even Microsoft chairman Bill Gates' lofty vision of "a PC on every desktop" seem myopic. That's because Broadcom's intent is to plunge its chips into every possible point in the broadband chain:
* In managed Internet backbone waypoints, with optical networking chips;
* In metropolitan, wide area, and local area network equipment;
* In cable, DSL and wireless broadband access networks,
* And in every end-device, including cable and VDSL set-tops, cable and DSL modems, and home-networked devices.
Nicholas tells investors that every single one of its silicon competitors could merge, and the resultant company still wouldn't have the firepower to induce broadband convergence. That's what makes the company one to watch.
In 2001 silicon, also watch for heightened activity among chip vendors to plow more stuff into fewer chips. Set-top silicon will sidecar cable modem and home networking, as well as compression techniques that enable video-conferencing, with the TV as the display device. Cable modem and DSL chips will tack on home networking, tuners, and voice-over-IP circuitry.
And the battle cry for all silicon vendors is to blend voice, video and data, onto chips sitting under the hood of broadband gateways, consumer electronics devices, and at every depot stop along the Internet.
|
