After several reorganizations of its Semiconductor Products Sector, Motorola Inc. is convinced that its chip arm now has what it takes to come on strong when market conditions improve.
It appears that SPS is already making some progress. In the second half of 2002, Motorola's chip business posted an operating profit, following eight quarters of operating losses. SPS reduced its annual revenue break-even point to $4.9 billion from $7.2 billion by executing its "asset light" program, which involved severely paring in-house manufacturing and streamlining its product portfolio to target three markets: wireless, networking, and automotive.
However, the transformation under the direction of Fred Shlapak, who took over as president of Motorola SPS in 2000, hasn't been without challenges. The chip business suffered a tremendous loss of manpower during the "asset light" implementation and still faces financial difficulties.
Motorola expects SPS to report an operating loss in the first quarter of 2003, citing seasonality in key markets such as wireless. Despite some minor bumps, however, the pieces are now in place for building a business that performs consistently, according to Shlapak.
Regardless of chip demand, SPS will continue to heavily invest in developing new generations of processors, a tactic that will put the company in a sound strategic position, he said.
"When the market does turn up, we will enjoy the benefits," Shlapak told EBN at SPS' Smart Network Developer Forum held here last week.
"Has it been difficult?" Shlapak said. "Yes. It has been difficult for everybody: suppliers, customers, us. But the business model of this industry has changed, and will continue to change."
Two fundamental changes in the chip industry to which Shlapak believes SPS has responded well are the escalating cost of fabs and adjustments to changes in the semiconductor market's growth rate.
While the semiconductor market has traditionally grown at an average rate of about 17%, Shlapak said he believes growth in the next 15 years will be closer to an average of 10% to12%.
The cost of building new state-of-the-art fabs, estimated at $3 billion or more, combined with the emergence of the large foundry industry, has changed the manufacturing dynamics of even the largest semiconductor suppliers, he said.
"We do not intend to ever build our own 12in. fab," he noted. "If it becomes necessary from a capacity standpoint, we intend to share the cost with someone, or expand the partnership we have with TSMC."
SPS' "asset light" program has enabled it over the past few years to reduce its number of wafer fabs from 27 to 8, and assembly and test facilities from 12 to two. The company also entered into a five-year development partnership with STMicroelectronics N.V. and Royal Philips Electronics, which includes operating a 300mm-wafer R&D fab in Crolles, France.
Ray Burgess, director of strategy, marketing, and communications at SPS, said that, leaving aside the Crolles facility, SPS has enough capacity to produce annual revenue of $7.5 billion, based on utilization rates at its fabs and its relationships with foundry partners, which manufacture 20% of its parts. By installing additional equipment at its existing fabs, SPS could expand capacity to bring annual revenue up to $14 billion, the company said.
"For the past two years, we've spent an enormous amount of our resources and time cleaning up our portfolio, reducing our asset base, eliminating processing, and transferring customers," Burgess said. "And we probably gave up a point and a half in overall semiconductor market share by getting out of products like DRAMs, SRAMs, and DSL chips, which are areas in which we didn't think we could command a No. 1 or No. 2 position."
Burgess estimates that in the past 18 months, SPS dedicated 50% or more of its design resources to moving products to new fab lines and requalifying customers and processes, tasks that are now completed.
SPS has also increased its efforts to protect patents and collect related royalties, as well as license technology, he said. Royalty and licensing revenue has increased from around $70 million in 2000 to about $200 million last year, according to Burgess.
SPS made a key move nearly 18 months ago to demonstrate that it is more than just a supplier to its parent company when it began offering its wireless baseband chips to the merchant market. Since then, the chip division claims it has collected more than a dozen cellular handset design wins.
"They're not playing a closed game anymore," said Tom Starnes, an analyst at Gartner Dataquest in Austin, Texas. "There is no doubt that SPS is also there to support Motorola's internal operations, but for the most part I think they have demonstrated the two things can be fairly synergistic."
Starnes said that SPS' recent announcements of new products for the wireless infrastructure market in the form of a reconfigurable compute fabric (RCF), sampling of its new PowerQuicc III family, extensions to its existing PowerQuicc families, and a renewed commitment to MCUs shows that "when you get down to it, in processors they've got it all, and can tie it together with the appropriate software."
However, given the imminent merger of Mitsubishi Electric Corp. and Hitachi Ltd., Motorola SPS could lose its longstanding position as the leading supplier of microcontrollers, even if it maintains or even increases its current market share, Starnes said.
Will Strauss, an analyst at Forward Concepts Co., Tempe, Ariz., said new products like the RCF "are a real eye-opener that could give SPS a leg up."
But whether SPS has put all the pieces together won't be known until the communications market fully recovers, he said. "They're clearly No. 1 in the communications processor market, but it doesn't mean a whole lot if there's no one there to buy what they've got."