Who knows what’s going to happen to corporate IT spending in the next year? Plenty of people are worried, for good reasons, about an economic downturn. Still, there are companies that are clearly placing their bets on the next generation of new technology and positioning themselves for a rebound.
Last week, the CFOs from two tech leaders, Intel and Juniper Networks, told CFO.com why they think the investments in new technology are important.
Intel’s announcement that it will spend $7.5 billion in capital expenses in 2001 is largely being driven by the company’s shift to a new generation of manufacturing processes and its transition to its next generation microprocessor, the Pentium 4, according to CFO Andy Bryant.
One piece of the equation is the move to plants that will come on line this year and make chips with a 0.13 micron width between transistors. The current generation of microprocessors has a width of 0.14 microns.
With tighter spaces between transistors, each microprocessor can process more data while consuming less power. At the same time, the company is planning to bring on line a new series of plants that will make silicon wafers 300 millimeters, or 12 inches, in diameter. Currently, Intel’s fabs turn out wafers 200 millimeters, or 8 inches, in diameter.
With larger wafers, Intel can make more chips from each wafer and thus lower its manufacturing cost per chip.
Bryant says that Intel typically tries to stagger the introduction of smaller widths between transistors and the move to larger wafers. Doing both in the same year is too much for the company to absorb smoothly.
“You don’t want to change the wafer size when you change the geometry,” he says. “One time we tried to do them together, and there were too many variables.”
Finally, Intel is moving the bulk of its manufacturing from the current Pentium III to the next generation, the Pentium 4.
Bryant says Intel has always viewed its engineering as a differentiating factor and the company needs to spend that much to stay ahead of the technology curve.
At the same time, the company is also faced with a sharp downturn in its revenue. Revenue in the first quarter of 2001 is expected to fall 15 percent from the $8.7 billion in sales the company had during the fourth quarter of 2000.
Bryant says Intel normally experiences a revenue fall off of 5 percent or more between the fourth quarter and the first, but this year, the seasonal trend is being aggravated by the slowing economy. Intel could have decided to delay some of its capital outlays, but Bryant says that by going forward with its plans, the company will at a minimum stay even with the technology curve.
Juniper Networks is also investing heavily to stay ahead of its rivals. In less than three years, the company has quickly earned a status as the chief rival to Cisco Systems in the market for network routers. The company spent $87.8 million on research and development in 2000, or roughly 13 percent of its total revenue of $673.5 million.
Marcel Gani, the company’s CFO, says Juniper would prefer to see its R&D budget be in a range between 15 percent and 20 percent of its total sales.
“We are below what we believe is a sustainable rate for those expenses,” Gani says.
In Juniper’s case, its routers must read Internet addresses in packets of data and send them to their right destinations. But router technology is getting more complicated all the time, and Gani says Internet service providers now need routers that do much more than just read packet addresses.
For example, routers need to prioritize packets that have a greater urgency, use the addresses to bill the right clients and filter out messages that might be coming from a hacker trying to crash a network.
This is where Juniper is headed with its next generation of products, and Gani says one of the tricks is keeping the company’s R&D labs always working one generation ahead of the company’s current product line. But that doesn’t come cheap.