Technology Stars Sparkle Across the New European Union

The mosaic that makes up the European electronics industry is full of surprises. Innovation is concentrated in many areas but often crops up in unexpected places.

What follows are four innovative companies with differing characteristics—large and small, fabs and fabless, old and new. The common thread is innovation and results.

Europe's predominant chipmaker is at a crossroads. A new boss, a market that's reformatting and a company overhaul have given U.S. $8.76 billion STMicroelectronics pause, according to industry analysts.

Earlier this year, Carlo Bozotti took over the company when Pasquale Pistorio retired after 18 years as CEO. Pistorio helped shaped the Franco-Italian chipmaker into a flexible giant. Pistorio's strong leadership forged tight relationships with key customers for products that were less fickle commodity, more stable application specific.

Things are changing, though. Bozotti has his hands full. ST was well-positioned during the cellphone boom through a tight relationship with leader Nokia Corp., says Chris Crotty, senior analyst at iSuppli Corp.

"[ST has] tied their fortunes to some key customers instead of broadening their outlook to the market in general and are struggling a little to find their way," Crotty says.

Flash memory is another area where the company must make choices. ST produces NOR Flash and hopes to expand further into NAND Flash with Hynix Semiconductor Inc. through its U.S. $2 billion chip plant under construction in China (see memory rankings, page 49).

NAND Flash, with total expected revenues of U.S. $8 billion this year, is the turf of big DRAM makers. "ST has to decide what they will do in [NOR and NAND Flash] memory exactly," Crotty says. "These are tough commodity markets and are all about scale. A lot of their products are more application-specific."

A spinoff of the memory business could be one solution. What should help ST through these challenges is its track record of innovation in multiple areas. Among them are analog and mixed-signal expertise, which is something of a black art. In fact, ST is the world's second largest analog player with U.S. $4.2 billion in revenues, behind leader Texas Instruments Inc., according to research firm Databeans Inc. (see analog rankings, page 52).

ST has also been an innovator in micro-electromechanical systems (MEMS) devices with accelerometers and pressure sensors earning the company about U.S. $200 million in revenue in a U.S. $4.7 billion total market. But the company is increasingly targeting consumer gadgets such as PDAs and cellphones, says Benedetto Vigna, MEMS business unit director.

ST also developed a bioMEMS "lab-on-a-chip" that allows small amounts of bodily fluids to be tested in seconds. ST's bioMEMS is on silicon where other companies use polymer. Silicon bioMEMS have the potential to bring high-volume production and lower costs to a market that is currently more like a niche. Volume production is expected at the end of 2006, according to Vigna.

The overall semiconductor market, however, is what counts the most, and that's changing.

As the sixth largest chipmaker, ST and other leaders have seen slight but steadily eroding marketshare, says Malcolm Penn, CEO of Future Horizons Ltd., a U.K.-based semiconductor research firm.

Penn contends the shift of the chip customer base from traditional telecom and computer to a more consumer-oriented one has caught big players by surprise. Customer loyalties are more short term in consumer and faster time-to-market is required.

"There's an overall restructuring going on that favors smaller companies over big ones," Penn says. "ST has to reinvent themselves to today's market reality. A lot of semiconductor companies are struggling with that."

On top of all this, ST is also in the midst of an internal cost-cutting restructuring to boost profitability, reflected in an 82% decline in Q2 earnings.Repositioning Europe's premier chipmaker will ensure busy but perhaps rewarding days ahead for new CEO Bozotti.

The Czechs have sometimes been too impressive. Czech engineers developed the Silly Putty-like Semtex explosive (ushered in under government control), as well as an electronic weapons system supposedly capable of detecting Stealth bombers by tracking electronic emissions.

Little surprise then that Tescan, based in Brno, would commercialize scanning electron microscopes (SEMs) and, so far, perform impressively in the market.

SEMs use a focused beam of electrons instead of light to produce very high-resolution images of a surface. SEMs were initially lab instruments that have exploded into multiple applications such as IC manufacturing, where they measure critical dimensions of a device and are used for wafer defect inspection.

Tescan, a private company with 75 employees, supplies conventional SEMs based on Tungsten-heated cathodes used for research and inspection in manufacturing, medicine, semiconductor fabs and mineralogy.

For a small company, sales have been brisk. In 2004, Tescan's revenues were U.S. $5.6 million, and Jaroslav Klima, managing director, expects 100% growth this year.

In Germany, the company sells to automotive industry customers like Bosch, Volkswagen and Porsche for materials testing and R&D. In the United States, universities and forensics labs have bought Tescan's SEMs. Infineon is a semiconductor customer.

A new factory going up in Brno is expected to double output. "We are not limited by the market, but limited by our manufacturing capacities," Klima says.

The SEM market for semiconductor applications is worth about U.S. $800 million, according to Bob Johnson, VP of research at Gartner Dataquest. He estimates the total SEM market at roughly U.S. $5 billion. FEI Co. and Hitachi Ltd. are major players. Klima believes among conventional SEMs, Tescan's unit sales are comparable to FEI's.

"Software is where the greatest differentiation is," he says. Tescan engineers wrote complex software that makes its SEMs far more user-friendly than many rival models. Martin Zadrael, head of R&D, says a strong knowledge of physics was required to develop the program.

Tescan was formed in 1991 as a spinoff from the state electronics giant Tesla. During the Soviet era, Tesla supplied the entire Eastern Bloc with electron-optical devices and electronic measuring instruments.

Klima himself started as a service engineer at Tesla in 1973, then moved into R&D engineering. Tesla's legacy in electron-optical devices, he says, was the core competency, but the company has gone in its own direction, tapping the local engineering institutes in Brno for fresh talent.

Moving forward, Tescan has already developed higher end SEMs. Later this year, the company will roll out a field emission SEM with stronger resolution and better probe density than its existing models.

Klima adds that Tescan is negotiating with European companies for R&D alliances to develop SEMs for nanotechnology, though he would not give details. "Generally we would like to move further into nanotechnology," Klima says.

Lousy sound quality doesn't make it anymore. Since Apple Computer Inc.'s iPod took the world by storm, the bar for mass market audio quality has been raised quite high.

A big help in lifting that bar was Wolfson Microelectronics PLC, a U.S. $119 million fabless company that developed the audio chip inside the iPod.

In fact, Wolfson's mixed-signal audio chips are everywhere. The company has consistently won high-volume design-ins in hit products such as Sony Corp.'s PlayStation Portable and PalmOne's Treo. Wolfson's chips are also in digital cameras. A stereo sound recording chip with special filters takes the noise out of the lens focusing, as well as wind noise.

And the company's ICs are likely candidates for PlayStation 3 and Xbox 360, according to Crotty of iSuppli. "Wolfson has positioned themselves in key markets by having the right products in the right markets at the right time," Crotty says.

According to iSuppli, total semiconductor revenues just for MP3 players are forecast at U.S. $3.1 billion this year. The handheld videogame market for semiconductors is U.S. $1.3 billion, for a potential total of U.S. $4.4 billion. That's not counting mobile handsets that incorporate MP3 players.

David Milne, Wolfson's CEO, describes his company as a mixed-signal analog company targeting end markets that happen to be in the digital consumer space. Milne puts Wolfson alongside Maxim Integrated Products Inc. and Analog Devices Inc.

"They may have broader product portfolios but we are still relatively small," Milne says. "We compete very favorably in terms of getting design wins in leading brand names."

Milne believes his company's competitive advantage is integrating key features such as power management into a mixed-signal chip, something traditionally done by discrete chips. "We bring high-precision audio with low battery power consumption at a price affordable for consumer end markets," he says.

Wolfson has 228 employees, with about half engaged in R&D. This year, about 13% of annual revenues will go to R&D and Milne expects to roll out a dozen new products.

Crotty, of iSuppli, says one challenge the company will face is rolling with the turbulent consumer industry. IC companies must provide semiconductors that readily span multiple generations of consumer gadgets, or get passed by. Another issue is that Wolfson makes specific parts within a platform. "The functionality they were producing could get integrated into a larger chip," Crotty adds.

Wolfson derives 88% of its sales from Asia. Last year the company moved into Japan and tripled sales. Japan now accounts for the bulk of Asian revenues. "There's not a single Japanese brand name that doesn't use one of our products," Milne says.

Milne's background is in physics and materials science, and he earned a Ph.D. from Bristol University. He ran a research institute at Edinburgh University that helped develop a range of products for a U.S. company that made U.S. $100 million off the designs. "I asked, 'Why the hell don't we do this ourselves?'"

In 1985, he spun off Wolfson as a mixed-signal IC design house doing work for specific customers. Ten years later, Milne raised money and transformed it into a product company.

Milne admits his company's path is uncommon in Europe. "Design houses in Europe don't seem to have taken the entrepreneurial step into fabless manufacturing," he says. "The secret is that the market is in the Far East and that's where you have to go."

Russia has long been a place for cherry-picking low-cost engineering talent. Startup companies have been rare. After all, entrepreneurial activity was banned for nearly 70 years.

But Research Center Module in Moscow could be just the breakthrough needed to spark Russia's technology sector. Fabless by necessity—Russia has no semiconductor infrastructure to speak of—RC Module doesn't make ICs in high volume, but develops and licenses intellectual property (IP) for high-performance microprocessor architecture. Out of 150 employees, 120 are engineers.

Penn, of Future Horizons, describes RC Module as a leading-edge microprocessor player not in the desktop space. "RC Module does advanced DSPs and microprocessors for a lot of number crunching in real time like fingerprint recognition—anywhere a lot of analog type input needs to be analyzed. It's a niche that's becoming increasingly important," Penn says.

Oleg Novikov, RC Module's managing director, says his company in 1998 was the first to come out with a 64bit DSP on the worldwide market. A unique feature of RC Module's NeuroMatrix DSP is that the architecture allows users and software developers to select the necessary operant bits—1 to 64—to meet application requirements. The flexible capability, protected by a Russian patent, helps it accelerate neural-network emulation.

Novikov says Texas Instruments and Analog Devices are now doing something similar, but they do not offer the user the full range of 1 to 64 bits. "We have a more flexible and powerful architecture in terms of vector matrix processing and this is our competitive advantage," he adds.

Novikov, whose background is in finance, founded the company in 1990 with four engineers who left collapsing Soviet defense enterprises.

A strategic partner has been Fujitsu Microelectronics Europe. Module uses Fujitsu's ASIC design center in Moscow to design ICs for outside customers, allowing the startup to avoid the heavy cost of EDA tools. In return, Fujitsu uses RC Module's technologies for ASIC, DSP and RISC development and is also a customer.

Other foreign customers include Samsung Electronics, which licensed software for HD TV applications. Based on RC Module's DSP, a German automotive company developed a car collision avoidance system and an Italian car maker built a lane-departure warning system.

Revenues in 2005 are expected to be U.S. $5 million, up about 10% from last year. IP licensing is still a small part of business. The bulk of revenues come from R&D projects commissioned by Russian customers. RC Module also develops embedded computers for industrial and space applications.

A next step is perhaps opening an overseas office for local support, something for which the company is seeking investment. Novikov would also like to grow the IP licensing business along the lines of MIPS Technologies Inc. "The main target is design of general-purpose baseband processors for broadband telecom systems and navigation systems," Novikov adds.

As a Russian fabless company, RC Module faces unique challenges. Penn of Future Horizons cites Russia's lawless image as an investment obstacle. Another is the lack of a semiconductor infrastructure, which hampers IC design development. Current design rules for the few Russian fabs are 0.8 micron.

On the plus side, Russia's significant science and technology base was generously supported by the U.S.S.R. for decades. "They were so closed all those years, they were forced to be innovative," Penn adds.

Dmitry Fomin, RC Module's marketing manager, said about six Russian fabless companies exist in the Moscow region. He believes more would emerge if chip fabs upgraded and barriers to entry were lower. Licensing EDA tools can start at $1 million, he pointed out.

Drew Wilson is a freelance writer based in Berlin, Germany.