Pirelli: On the Trail of the Holy Grail

Can the cable maker build the first optical chip?

Pirelli (PIREF ) Chief Executive Marco Tronchetti Provera made a killing on his first investment in optical electronics. The Italian executive bet roughly $100 million on new technology for fiber networks and components to speed phone calls and Internet traffic. Then, as Net euphoria peaked in late 1999, he sold his optical components businesses to Cisco Systems Inc. (CSCO ) and Corning Inc. (GLW ) for $5.8 billion.

Now Tronchetti is hoping lightning will strike twice. On Jan. 14, Pirelli, a Milan-based manufacturer of cable and tires with $6.5 billion in annual revenues, sealed a partnership with Massachusetts Institute of Technology to develop optical chips--essentially miniaturized versions of the optical components that Pirelli's labs turned out. By creating chips that can be mass-produced at dramatically lower costs than today's largely hand-built components, the partners hope to make it possible for telecom operators to deliver enormous broadband capacity to consumers at one-tenth of today's cost. Under the terms of their collaboration, which calls for a $130 million investment over five years, MIT will own the patents to any new technology while Pirelli will have exclusive rights to market it worldwide. "I'm sure this is the technology of the future," says Tronchetti, 54. "We have a chance to be first."

This time, however, Tronchetti is wagering on an industry mired in gloom. Broadband companies such as Global Crossing Ltd. are buckling under a glut of network capacity. These players have plowed billions into long-haul fiber-optic networks, which permit data, voice, and video signals to travel at near the speed of light. Yet demand hasn't materialized as quickly as planned.

A breakthrough by Pirelli and MIT could help the telcos and their equipment makers realize the return on investment they have been looking for. Because of current technology restraints, the average consumer can't take full advantage of the superfast transmission speeds fiber-optic networks offer. There are two problems. First, the "last mile" of cable connecting these networks to people's homes is still mostly old-fashioned copper wire. To travel down the copper wire, light signals must be converted into electronic signals, creating huge bottlenecks of data traffic.

The second problem concerns the optical components currently used for the conversion of light signals into electronic ones, which is what computers and TVs respond to. Producing these components is a timely, labor-intensive process that drives up the cost of the whole network operation, both for carriers and customers. If telcos wanted to offer 1-gigabit-per-second service using today's technology, they would need to deploy a set of components costing roughly $1,000 per household connection. So telcos haven't offered the service.

Pirelli and MIT hope to bring that price tag down to $100 within two years. The solution: shrinking components to the size of a molecule and putting them on a chip. "The problem is cost. If a low-cost solution exists, operators will deploy it," says Giorgio Grasso, who heads Pirelli Labs in Milan. Meanwhile, Grasso hopes demand for the chips will be bolstered by the gradual shift from copper wire to fiber. New housing developments in U.S. cities and surburbs, for example, now often come pre-wired with fiber.

While Grasso and his MIT colleagues aim to bring single-component optical chips to market in 2004, their ultimate challenge is to cram different components together on one chip--a breakthrough that would reduce costs even further. Installed at first in phone lines and later perhaps in digital appliances themselves, the new-generation chips would help direct traffic on fiber-optic networks, zapping data and video directly into home computers and TV sets at the speed of 1 gigabit per second--more than 10,000 times the 56 kilobits per second that is now standard. That kind of speed could turn desktops at home and in the office into smoothly operating interactive media centers.

The chips would also likely spur a new generation of Internet products and services, including many that have been heralded for years but have so far failed to materialize, such as video on demand or software on tap. "It's the next breakthrough everyone is looking for," says Karen Liu, director of optoelectronic modules at market researcher RHK Inc. in South San Francisco, Calif. And MIT figures Pirelli, with its proven track record in optical components, is its ticket to speedy commercial success. "Pirelli will take our innovative work and put it in the real world," says Lionel C. Kimerling, director of the MIT Microphotonics Center in Cambridge, Mass.

If Pirelli and MIT succeed in shrinking components to the size of a molecule and cramming gobs of them onto one uniform base, they could unleash a technological revolution similar to the one that occurred when semiconductors replaced transistors. As with microprocessors, manufacturing costs would plummet as production volume was ramped up.

Others are also chasing after the holy grail of integrated optical chips. Lucent Technologies Inc. (LU ) spin-off Agere Systems Inc. and Nortel Networks Corp. (NT ), as well as some 200 startups, have entered the race. NanoOpto Corp., a Somerset (N.J.)-based startup with $16 million in venture capital, plans to deliver its first three miniaturized components in March. But nobody has managed to affix many molecule-size components to a single chip.

Experts say it's still too early to predict who will win this race. Tronchetti appears to have an edge. He has billions in cash to fund his ambitions and Pirelli has escaped largely unscathed from the combined meltdowns in the telecom and dot-com sectors.

Still, formidable technological hurdles remain. Besides finding a way to reduce components to the size of mere molecules, engineers must come up with an optimum material to combine different optical elements on a chip so they can function together--as silicon does for semiconductor chips. Getting the world's telecom companies to embrace the technology will be the next big challenge. "They are way underestimating the commercial time lag," says Lawrence D. Gasman, president of market researcher Communications Industry Researchers Inc. in Charlottesville, Va.

But Tronchetti is not fretting. The money he is pouring into the project amounts to just 2% of his company's annual revenues. "The risk is investing too much too early," he says. In this game, as in many others, timing is everything. And Pirelli's chief has already proven that he knows just when to walk away from the table.

By Gail Edmondson in Milan, with Andy Reinhardt in Paris

Before it's here, it's on the Bloomberg Terminal.