JDSU rethinks priorities as rivals watch and wait
Jozef Straus, the straight-talking CEO of optical-component maker JDS Uniphase, talks to John Devaney about physics, economics and his strategy for surviving the market slump.
Sometimes it seems as though chief executives are produced on a conveyer-belt system, with big bespoke suits, well groomed hair and shiny shoes. So Jozef Straus, with his trademark black beret, distinct European accent, frank opinions and cultural references is a pleasant change for the jaded interviewer.
As boss of one of the largest companies in the optical-communications business, Straus is undeniably an influential figure. So it is not surprising that, despite the idiosyncrasies, he is as sharp as a scythe.
Jozef Straus
Straus divides his time between Ottawa, Canada, and San Jose, California, the twin homes of JDS Uniphase (JDSU). Interrupting his schedule, FiberSystems International talked to him over a coffee in California.
A good place to start
"Let's start from the beginning," Straus said. "JDS Fitel was a Canadian company making passive optical components and Uniphase was a US company making active components. We saw a convergence from passive to active."
With a lot of customers in common and a complementary product line, the June 1999 merger promised the new company broad coverage of the optical-components market. But the plan did not end there. The newly formed giant flexed its muscles and went on a spending spree. By early 2001 it had swallowed up at least 14 specialist component makers. The strategy was a simple one - each acquisition would bring a new set of technologies, products and manufacturing facilities to the mix, with the ultimate aim of creating a powerful one-stop "supermarket" for optical devices.
Epitaxx, for example, acquired in November 1999 for $400 m, brought receivers, while Sifam, purchased in December 1999 for $96 m, brought couplers. The spending spree came to an end with a phenomenal outlay for SDL in February 2001. This deal, worth $41 bn when first mooted in July 2000, took months to receive clearance from the US Justice Department.
Then the communications bubble burst and JDSU's sales plummeted from $925 m for the quarter ending December 2000 to $286 m for the quarter ending December 2001. And the end is not yet in sight. JDSU foresees another drop in sales for the next quarter of 2002 and refuses to give guidance beyond that point.
"As they say in Germany, the market has been krank [ill] a long time," Straus said. "Some carriers are saying they will see some pick-up at the end of March or in June. But when will it filter down to us?"
Playing the economic soothsayer is complicated by a major side effect of the previous market boom - excess inventory. When times were good, carriers inserted hefty penalty clauses for systems vendors that delivered late with their contracts. As a result, systems makers ordered their components way ahead of demand and were left with lots of unneeded inventory when the downturn kicked in.
On the bright side
Straus manages to see some upsides to this boom-bust cycle. "In growth mode, the focus was to fill out more for our customers," he explained. "But the slowdown allowed us to revisit our business processes and platforms."
Restructuring program
JDSU grandly calls this revisitation its Global Realignment Program, which should cut the company's annual spending by about $900 m. The cost-cutting plan includes reducing the workforce by more than half and closing several production plants. Straus says that automated manufacturing is crucial for creating a more cost-effective and productive method of making devices.
JDSU is embracing automation through both in-house innovations and its new arrangement with specialist company Adept (FiberSystems International February 2002 p16). Straus says that the needs of automation vary from one product sector to another.
"In passive components, automation is different from in active components," he said. "We have large volumes in passive components so we need to automate manufacturing processes onto a single platform. But in semiconductor-laser manufacture automation is important because yields are still very low and you have to pick [out] chips all the time."
Few manufacturers will admit to low yields, but Straus says the problem is common and understandable in such a young industry. He makes a comparison to the electronics industry before the invention of the op-amp and the integrated circuit.
"The first major fiber-optic links were put together in 1974, just over 25 years ago, and the first [coarse] wavelength-division multiplexing was in 1981," he explained. "The industry has utilized the technology to the fullest extent while that technology is still very young."
JDSU's investment in automation should leave the company with enough capacity to respond when growth begins again, says Straus. But an accurate estimate of capacity is crucial for making automation work - too low an estimate will mean missed opportunities, while too high will mean idle machinery and wasted expenditure.
As for how quickly the market will grow when it starts to pick up, Straus admits that he is uncertain. "I don't think we're going to see 100% growth. It is more likely to be a respectable 20-30%," he postulated. "I don't want to be a Luddite as I would be the first to benefit if it was 100%, but I think people will be more careful with their ebullience."
With a nuclear-physics degree from Czech Technical University, Prague and a doctorate in low-temperature quantum physics from the University of Alberta, Canada, Straus has an invaluable depth of scientific knowledge that sets him apart from the executive pack.
"When I talk to our engineers, I can talk physics or fiber with them," he said. "But I have also learned that the rules of physics apply well to economics. That's why I encourage physicists to take economics courses and economists to take some science courses."
From a physicist's point of view, according to Straus, a laser is a laser whether it transmits Gbit/s or Mbit/s speeds and whether it is in an undersea transmitter or a transponder in a central office. He says that having many different product lines under one corporate roof means that valuable lessons and experience can be shared easily.
As an example, Straus mentions JDSU's acquisition of the IBM transceiver business (see FiberSystems International March 2002 p19). "Before, we had companies in the submarine, long-haul and metro markets," he said. "Now with IBM transceivers we have extended to the enterprise. But it's all photons. A transceiver's a transceiver."
JDSU's expansion strategy was simple: to cover horizontal slabs of the market - passives, actives and subsystems. So will JDSU start climbing the food chain to cover yet more complex equipment?
The complete solution
"We are not a systems company. We are a components company," Straus said bluntly. "Today, though, our customers prefer that we give them more complex solutions because they don't have time to gather the individual components and assemble them."
This trend is matched by a reduction of in-house optical-components divisions within systems giants such as Straus' first employer, Nortel (Bell Northern as it was then). He reasons that their very creation, about 10 years ago, was because certain components could not be bought on the open market.
As if to show the contrast between then and now, Straus reels off the components JDSU will launch this year - new amplifier designs he calls microamps, uncooled lasers and 10 Gbit/s transponders. But he seems to be most excited about a "colorless wavelength switch" for add-drop multiplexing.
Although there's no shortage of rivals pushing back the frontiers of technology, Straus is not too worried: "I'm a technology man, so I always watch that we don't get surprised by some new development."
• This article originally appeared in FiberSystems International April 2002 p31 |
