July 2, 2004 – Three years ago, optical networking was a highly anticipated end use for optical MEMS. Just as the buzz reached a fever pitch, the market collapsed and all opportunity evaporated.

Now that the market appears to be coming back to life, a critical question must be asked: Is telecom finally ready to embrace MEMS-based solutions? And if so, then is this market still worth pursuing? The answer appears to be yes on both counts.

The Optical Fiber Communication Conference and Exposition (OFC) held in Los Angeles in late February was telling in several ways. There was cautious optimism that the market is now finally beyond bottom. Considerable cost-cutting efforts have vastly improved the health of carriers, which have slowly begun increasing their expenditures for capital equipment, or are at least now seriously considering it.

This creates a domino effect in the industry, which is important for those offering MEMS-based solutions; since these companies are essentially situated at the opposite end of the supply chain, they are not the first to benefit from a recovering market.

Whispers of increased customer interest in MEMS-base products were notable at OFC 2003. This appears to have come to fruition as a number of MEMS suppliers at the 2004 show indicated that they had seen a definitive upswing in revenues in the last quarter of 2003, which continued into the first quarter of 2004.

Some saw just modest increases, while others experienced a tenfold rise in sales.

While this is encouraging, it doesn’t necessarily mean that the telecom market as a whole is improving — or is at least well on the road to recovery.

I have my own personal barometer:  While it sounds insanely silly, trade show giveaways can be a telling, albeit quirky, means of gauging industry health. Compared to previous years, OFC 2004 was pretty quiet (meaning there was little hoopla). It was definitely all business.

On the surface, few giveaways were to be found beyond the occasional pen or bag. But there was indeed some more creative stuff out there — if one just kept a sharp eye out. It would appear that companies are now feeling good enough to spend a little bit on trinkets, but not enough to give them away en masse as in previous years, so they were kept under wraps for real customers — who were apparently there.

There were some serious goings-on behind the scenes, as I was told that quite a number of deals closed at the show — some of which pleasantly surprised a few of the exhibitors, who apparently did not expect this to happen.

This begs the question of who’s left to supply MEMS-based solutions? OFC was very telling in this respect as well, as the market has evolved considerably.

Given the fact that, in the past 12 months, 10 companies have refocused elsewhere (either with non-MEMS technologies, or MEMS for markets other than telecom), and six companies have gone dark, what is most surprising is that there are still more than two dozen companies left that are offering or developing MEMS for optical networking.

Three companies in the product development stage have recently emerged from stealth mode, and four are, in fact, newcomers. Even better, 19 companies are actually shipping product for revenue.

What’s different is product focus. Gone are the days of mass pursuit of super-sized cross-connects (the mega switches). These days, of the 20 companies pursuing MEMS switches, only three are offering anything larger than 16×16: Calient Networks, Glimmerglass Networks and Movaz Networks.

Most of the remaining companies have concentrated on 1×2 and 2×2 arrays. The focus has shifted to Variable Optical Attenuators (VOAs), with more than a dozen companies in this space (and most of them also offering switches). In fact, this is where much of the revenue generation is occurring right now.

Eight companies are now developing tunable and/or dynamic components, such as tunable lasers, tunable filters, Dynamic Gain Equalizers (DGE), Dynamic Channel Equalizers (DCE) and related products.

All of this does seem to support the fact that telecom is indeed embracing MEMS-based solutions, and that this market is still worth pursuing. Venture capitalists sure seem to think so, given that investing is again on the rise.

More companies are talking about acquisitions (although I think the days of the multibillion-dollar valuation a la Xros are history). Surprisingly enough, that includes startups looking to expand their product focus in a cost-effective way.

In addition, whereas MEMS had, at one time, a slightly negative connotation (due to misunderstanding about reliability), some companies are now embracing the term because it’s seen as a competitive advantage. What a turnaround.

But let’s take time for a reality check. Telecom, and optical networking in particular, is still in recovery mode. An increase in customer inquiries does not necessarily translate into increased sales.

MEMS suppliers, particularly those who are relatively new, must still demonstrate that their products are the most cost-effective solution available to meet their customers’ needs. In addition, the evaluation and qualification cycle is pretty lengthy. So, even though sales are likely to increase strongly over the course of the next 12 months, we’re still looking at 2005 as the real turning point.