Indeed, history is repeating itself. Just as the semiconductor industry faced these questions decades ago, so too is the fiberoptic component industry today. Lessons from the chip business should give us a pretty good idea of what things will look like in fiber-optic components in 10 years.

Prediction No. 1: Over the next two to three years, large vertically integrated optical-component manufacturers will enjoy an enviable position in the market. Currently, the design and manufacturing of optical components requires more research than development. It is possible to create intermediate-term competitive advantages in manufacturing that will allow companies to grab large profits in this capacity-constrained market.

Prediction No. 2: A fiber-optic component capital equipment market will develop. Though a company like Applied Materials might be expected to capture this opportunity, it is entirely possible that a large fiber-optic component manufacturer will eventually spin out an equipment division. Just as in semiconductor equipment, one should expect companies to sell both machines and processing recipes. This is likely to happen in about five years.

Prediction No. 3: The fiber-optic component equivalent of an ASIC industry will develop. With recent strides in planar lightwave-circuit technology, there are nearly enough building blocks to make optical ICs for communications. Optical couplers, attenuators, thermo-optic switches and a variety of other building blocks can already be manufactured on the substrate. Again, we may see one of the traditional ASIC vendors capture this opportunity, but more likely, this business will be spun out of large, vertically integrated manufacturers. The beginnings of this business are already forming at companies like Agilent and IBM.

Similarly, the advent of a fabless optical-component industry association, and of overseas foundry and packaging houses, will come as no surprise. Finally, the simplest and least expensive technologies (read: planar lightwave circuits) will likely win. There will always be a place for MEMS and other exotic technologies at the high end, but planar circuits will command the lion's share of the market.

Jeremey Donovan is a Principal Analyst for Communications at Gartner Dataquest.