Optical/IP Networks

Shedding Darwin on Light

Let's face it: The last four or five months have been brutal for the optical networking industry. The problems began with concerns about the slowdown on carriers' capital spending, and were followed by a Nasdaq annihilation that saw the share price of many of the public optical players fall by 50 percent or more -- a lot more, in the case of companies like Corvis Corp. (Nasdaq: CORV), Foundry Networks Inc. (Nasdaq: FDRY), Lucent Technologies Inc. (NYSE: LU), Redback Networks Inc. (Nasdaq: RBAK, and Sycamore Networks Inc. (Nasdaq: SCMR).

Unfortunately, the optical industry's problems aren't over yet. Like Biblical plagues, there are more woes on the way.

The next curse to strike the optical networking industry will be an almighty winnowing amongst the startup community.

Pruning is inevitable given the "if it moves, fund it" attitude to optical investing taken by the venture capital community last year. Light Reading maintains a global directory of every company (public and private) involved in the optical networking industry -- from fiber infrastructure service providers, to systems vendors, to components manufacturers. There are now over six hundred companies on the list (605, to be exact). We also maintain a separate registry of companies that are still in "stealth mode," which means they aren't at the point where they are ready to start talking publicly about their products. That list now totals 168.

That's too many, obviously. In fact, the shakeout has already claimed one of its first IP system vendor victims: Ironbridge, which shut its doors at the end of last month (see IronBridge Has Fallen Down). IPHighway Inc. and Point Reyes Networks Inc. look set to follow Ironbridge into systems vendor Gehenna (see Shutdowns Send Dark Message).

They won't be the last. Many optical startups are already having trouble finding further investment, and some are having to settle for unpublicized down rounds, where a company takes money at a lower valuation than at its previous round -- the thin end of the shakeout wedge.

How bad could things get? Vinod Khosla, the legendary optical networking investor, said recently that he thought that 95 percent of startups would ultimately wither on the vine (see Khosla: Optical Market's Still Huge).

There are several problems with Vinny's apocalyptic prophecy. First, it's a gigantic over-exaggeration. The real number is a lot lower: 50 percent, maybe. And most of those companies won't crumble into dust -- they'll be acquired. (For a more realistic taste of what's to come, consider antecedents such as the LAN switch or high-speed router markets).

Further, this kind of doom and gloom doesn't actually do much to help service providers and investors work out where they should be putting their money. The fact is that the optical shakeout will affect the three different categories of startups (service providers, component plays, and systems houses) to different extents, and at different times.

Here's a brief Light Reading prognosis for each of those areas:

Service providers were the first to exhibit marks of the Beast. Aduronet, Digital Broadband, NorthPoint Communications, and GST Telecommunications have already gone bankrupt. Global TeleSystems Inc. (GTS) (NYSE/Frankfurt: GTS) and ICG Communications Inc. (Nasdaq/Neuer Markt: ICGX) are both looking a bit leprous. E.spire (Nasdaq: ESPI) and Urban Media Inc. are also said to be having problems. And lots of the pan-European backbone operators are teetering on the brink of financial ruination.

In the long run, consolidation is likely to be the order of the day. The big service providers (especially the IXCs) will get bigger. Most of the smaller players (especially data-only players without voice revenues from an established customer base) will be acquired, or simply fade into the æther.

Of course, problems in the service provider community will inevitably affect the systems vendors, who are next in line to feel the pain. But, again, not all companies in this market are as susceptible to the shakeout miasma. Clearly, the metro market is overcrowded, with startups like Alidian Networks Inc., Astral Point Communications Inc., and Mayan Networks Inc. still struggling to establish a toehold against larger public companies (many of which have themselves bought startups in order to compete effectively in this realm).

Another subset of the systems market that is sounding alarm bells is the all-optical switching tribe -- as typified by Calient Networks Inc., Ilotron Ltd., Luxcore Networks Inc., Nayna Networks Inc.

These players have captured the imagination of the investment community. (Hey, "all optical" must be good, right?) But the reality is that the market for these products is still developing. And even when it does arrive (in a couple of years' time) it will still represent a tiny proportion of the overall market for optical equipment. Expect all-optical casualties.

Components companies have so far been relatively immune to the malaise. That's about to change. I strongly believe that the attrition rate in this area will eventually surpass that of service providers and systems startups combined.

The reason is that there are currently no standards -- or even a consensus of opinion -- over which type of component technology to use in developing the next-gen optical kit. For example, there are more than half a dozen approaches to building all optical switch fabrics alone -- including arrays of tiny tilting mirrors, liquid crystals, bubbles, holograms, and thermo- and acousto-optics (see Optical Switching Fabric).

This lack of standards makes life hard for component startups. They don't have the bulk to force their proprietary approach on systems vendors. And in the current climate, the systems vendors would rather buy from larger component manufacturers in any event, since there's less chance of them going under. A vicious circle.

The irony is that life for these companies will be even worse when standards do arrive. If the history of networking tells us anything, it's that so many diverse approaches will not survive, and that eventually one technology will win out. (The networking industry loves a consensus, even if it takes it a long time to reach it -- just look at Ethernet or IP.) At that point, components startups that invested in other than the chosen approach will be denied admission to the Promised Land.

If it all sounds a bit depressing, that's because it is. However, it's also important to remember that there's also an upside to any industry shakeout. In the long term this natural wastage will have a positive affect on the optical networking industry.

Think Darwin, with a Nietzsche chaser.

The weak will perish in the desert. The strong will enter the Empyrean -- and prosper to an outrageous degree. The one thing that isn't in doubt is the size of the optical market opportunity, which boggles the imagination. All in all, a good thing for both service providers and investors -- at least, those that back the right players and the right technologies.

This raises another obvious question: Which areas are still hot? Automated subsystem manufacturing is one example. It's a "happening scene" for two reasons: First, it enables systems vendors to shorten the time to market and reduce development costs. Second, there aren't a lot of startups in this area. Axsun Technologies is one of them; Cenix Inc., another.

2001 is also set to be the year of Jubilee for optical storage area networks (SANs). Why? On the one hand, a wave of demand for SAN equipment is coming from corporate users -- folk who desperately need to consolidate their storage requirements and are used to spending big bucks on communications.

On the other hand, this is a technology that uses whole wavelengths -- which is a big attraction for carriers, because it translates into generating revenues quickly. It means they can just plunk in some DWDM gear and away they go. They're not faced with having to install edge switches and the like to support and bill for lower bandwidth, lower-value services.

How long will these Pearly Gates stay open? Until VCs start clogging them later this year. A shakeout follows in late 2002.

-- Stephen Saunders, US editor, Light Reading http://www.lightreading.com

Sign In