The demise of Megisto Systems earlier this month brought the amount of VC cash invested in failed wireless router startups to a whopping $191 million, prompting industry speculation as to the cause of the fallout.

At the turn of the millennium, a number of young hopefuls – Cambia Networks, Megisto Systems, Starent Networks Corp., Tahoe Networks, and WaterCove Networks – burst onto the wireless router market, all attempting to lobby carriers into bypassing the usual roll call of established vendors, such as Cisco Systems Inc. (Nasdaq: CSCO), Juniper Networks Inc. (Nasdaq: JNPR), and Nortel Networks Ltd. (NYSE/Toronto: NT).

Wireless routers, which link next-generation GPRS and CDMA2000 radio networks to the IP backbone, are called gateway GPRS service nodes (GGSNs) in GSM-derived networks and packet data serving nodes (PDSNs) in CDMA networks.

Although these startups all had slightly different propositions, they were all focused on allowing operators to develop and provision new data services. Early forecasts for market growth predicted that this sector was set to become a major revenue spinner, to be worth $2.5 billion by 2005 and $4 billion by 2006 (see What's the Wireless Router Market Worth?).

The VCs certainly agreed, as a bevy of startups all scored serious cash: Cambia notched up $11 million; Megisto secured approximately $60 million; Tahoe raised $50 million; and WaterCove racked up $70 million (see Having a Flutter on the GGSNs).

One by one, however, these companies fell by the wayside. Cambia was the first to go. Then, in November 2003, Nokia Corp.'s (NYSE: NOK) bought Tahoe's assets, after the startup had failed to register a single carrier win (see Nokia Sweeps Up Tahoe). In January 2004, Alcatel (NYSE: ALA; Paris: CGEP:PA) acquired WaterCove at a rumored value of "virtually zero," following its solitary win at Orange UK (London: OGE) (see Orange Juices Watercove and Alcatel Swallows WaterCove). Earlier this month, Megisto sold its assets to Interactive Technology Holdings Ltd. in a deal reported to be worth around $1 million, following its earlier attempts to break out of the wireless router space (see Megisto Breaks Silence and Megisto Sells Up).

Now, only Starent remains as the lone successful independent player in the market.

So why did the majority of startups in this space struggle for carrier attention? According to analysts, these innovative vendors may simply have been just too far ahead of their time.

"As the 90s internet boom waxed full, entrepreneurs experienced at delivering high-performance IP forwarding boxes looked to mobile devices as the logical next great thing," notes Current Analysis's Ken Rehbehn. "Yet mobile vendors and operators had inadvertently deployed deeply flawed GSM and CDMA IP data architectures. GSM GPRS packet data sessions limped along at data rates barely reaching 28.8 kbit/s. Simultaneous voice and data sessions were not implemented. Always-on IP network access was omitted. The consequence of these crippled early network and handset implementations was that highly scaled, richly functional gateway boxes from innovative entrepreneurs were really not needed – giving incumbent vendors plenty of time to evolve their GGSN/PDSN product offers through acquisition or innovation."

And although the market needs envisioned by early entrepreneurs are finally coming to fruition via high-speed UMTS and CDMA 2000 1xEV-DO technologies, Rehbehn states that incumbent vendors are now prepared with products that meet carrier demands.

"In short, the market just took longer to evolve than expected," adds Current Analysis's Peter Jarich. "Look at the offerings from the big guys – it's taken folks like Alcatel and Nokia until now to release their next-gen GGSNs... It's taken this long to really see the need for advanced solutions develop."

Starent, meanwhile, remains an isolated ray of light. It has banked $80 million in VC investment, but unlike its failed rivals, it has also racked up a number of high-profile wins, including a $30 million deal at Verizon Wireless and a rumored OEM partnership with Nortel that would involve a potential deployment at Sprint Corp. (NYSE: FON) (see Starent Shines at Verizon, Starent May Have Nortel Win, and VCs Shine on Starent).

Such success is largely attributable to Starent's background: A number of its execs once worked at 3Com Corp.'s (Nasdaq: COMS) CommWorks division, the PDSN products of which were sold to UTStarcom Inc. (Nasdaq: UTSI) for $100 million in March 2003 (see UTStarcom Cops CommWorks).

Also, an early OEM deal with Samsung Corp. in the booming Korean CDMA data market gave Starent a strong foundation for future business, and the startup appears to have made a wise decision in initially targeting CDMA carriers, rather than the larger GSM-based market that its competitors were keen to grab (see Starent Teams With Samsung). Indeed, nine of Starent's ten announced wins are for the supply of PDSN, rather than GGSN, equipment.

Starent's VP of products management and marketing, Gennady Sirota, explains the company's choice to focus on CDMA: "The throughput with CDMA was significantly higher than with GPRS... As we became more and more successful in CDMA it allowed us to offer solutions in UMTS, such as the win we have at SK Telecom" (see Starent Shines On).

"It's clear we were focused on the right market with the right technology at the right time and with a very driven team," adds director of marketing Andy Capener.

— Justin Springham, Senior Editor, Europe, Unstrung