Alas, it's still cold outside. Deals that pumped $120 million into Caspian Networks and $53 million into Pluris, based in Silicon Valley, as well as $45 million into Montreal's Hyperchip and $80 million into Chiaro Networks of Richardson, Tex., say as much about VCs' lack of alternatives as their confidence. All four companies are chasing the same $2.3 billion North American market for core routers, computers that read packets of data and pick the fastest path to get them to their destination. And almost all of that market now belongs to Cisco Systems Inc. (CSCO
) and Juniper Networks Inc. (JNPR
Sound crowded? It is. But core routing is a rare area of telecom capital investment that's still expected to grow. Market researcher RHK Inc. says telecom capital expenditure won't get back to 2001 levels until 2006. But it figures core routers will reach about $6 billion in sales by 2006--enough to sustain one newcomer. In the land of telecom investors, blinded by the industry's depression, that makes core routers the new, if one-eyed, king.
What the newcomers bring to the party is the prospect of much faster handling of network data. Since 1996, fiber-optics has revolutionized how much traffic can move in a straight line between cities, but routers haven't improved nearly as fast. Juniper's top-end machines can process 160 billion bits of information a second, much less than the 6.4 trillion bits telephone lines can dump into a network hub. Hyperchip says its machine will have 8,000 times the capacity of Juniper's biggest router.
Underscoring just how tough things are: In several cases, the new deals wiped out all or most of the value of earlier investments. Investors in Pluris Inc. lost nearly all the $167 million they had put in earlier. Caspian Networks Inc. concedes that its value is sharply lower than when it last sold stock in 2000. Only Hyperchip's stated value, some $400 million, is the same as in 2000, but rivals carp that it boosted its numbers by getting most of its new money from a government agency trying to keep the company in Canada.
Nonetheless, venture capitalists don't have a better option if they want to invest in telecom. And VC firms are still sitting on the bulk of large pools of capital they raised in 2000 and 2001, which they need to put to work. Peter Morris, a Menlo Park (Calif.)-based venture capitalist at New Enterprise Associates who backs Caspian, said firms have little choice but to forget the past and compare these deals with current prospects. "They recognize the past investment is pretty much gone," Pluris marketing Vice-President Gary Law says.
Venture capitalists aren't dumb: The new deals may actually be less risky than the alternatives. The four upstarts have each spent more than $100 million, mostly on research. All have equipment undergoing laboratory tests with carriers, and all are aiming to ship products late this year or in 2003. So the technology appears to work. And the high price of entry ensures that new rivals won't emerge soon.
But there's plenty of risk. Gary Morgenthaler, a venture capitalist backing Caspian, says VC firms always back several players in new markets--but never has so much money been involved (or written off). These firms don't dare predict they will soon go public. Privately, VCs say even the most successful may be takeover bait for Juniper or Cisco, who aren't exactly shaking in their boots. "They're trying to impress with slideware and ideas," says Juniper marketing chief Carl Showalter.
The lesson: Don't believe what you see. Telecom's share of VC investment in 2002 has reached 34% including the four deals, higher than the 20% to 25% it claimed during the fiber-optic boom of 2000. But it's a groundhog's shadow, not a sign of spring. Mullaney covers venture capital.