Nexsi Corp., a Silicon Valley startup that was working on a data center switch, has officially shut down.

The company has filed for Chapter 7 bankruptcy, the Website has been dismantled, and the company’s phone system is not routing calls. On Monday, April 15th, the San Jose, Calif., company, shut its doors for good. Like many other startups, Nexsi simply ran out of money, says Dave Passmore, research director for the Burton Group.

Chapter 7 bankruptcy is more serious than a Chapter 11 filing. In Chapter 11, a company is filing to have a court distribute assets to creditors so that it can reorganize. Chapter 7 companies have no plans of pulling back from the brink.

The Nexsi 8000, introduced in September 2001, has been described by some analysts as a data center switch on steroids. Many likened it to a souped-up Alteon or Arrowpoint Web switch, both of which enabled networks to more efficiently route application data and speed up Internet connections. Alteon WebSystems was sold to Nortel Networks Corp. (NYSE/Toronto: NT) in July 2000, and Arrowpoint was sold to Cisco Systems Inc. (Nasdaq: CSCO) back in May 2000 (see Nortel Buys Alteon for Big Bucks and Cisco Finds A Soft Spot for ArrowPoint). In fact, Mark Bryers, vice president and chief scientist for Nexsi, was the founding CEO and CTO of Alteon.

Nexsi had raised a total of $90 million, $75 million of which was raised in September of 2000 (see Nexsi Nets $75 Million). Big-name backers like Sequoia Capital, Raza Foundries, Amerindo Investment Advisors Inc., and Presidio Venture Partners were on board for both rounds of funding. But none of the investors was willing to continue pouring money into the company. The VCs did not return calls by press time.

Although the product seemed promising, the company had one fatal flaw, according to Passmore. The architecture called for a large number of microprocessors assembled on an ASIC to achieve multigigabit throughput. The technology was cutting edge and an exciting breakthrough in switching, but it was very expensive to package and manufacture.

“The company was really pushing the state of the art in terms of ASIC design,” says Passmore. “But it’s not easy to pack that many gates and processors onto a chip, and they were burning through incredible amounts of money.”

Nexsi’s expensively designed product coupled with the economic downturn spelled disaster. But even eight months ago, the company seemed to be optimistic about its prospects, at least on the surface. Light Reading noted in an article about the networking job market that Nexsi was recruiting aggressively for new employees (see Job Market Isn't Dead).

The void left by Nexsi will soon be filled by two other startups also building super-performing data center switches. Inkra Networks, which has been around since 2000, has raised $36.5 million so far. There are rumblings in the venture community that the company is about to close another round of funding. Nauticus Networks Inc., which has raised $33 million and has recently come out of stealth mode, is also another competitor, although it is focused more on the large enterprise market (see Nauticus Emerges From Stealth Mode).

Even though these companies have raised far less cash than Nexsi had, Passmore feels they have a better chance of success.

“Inkra and Nauticus have completely different architectures from Nexsi,” he says. “They haven’t revealed their secret sauce, but I can tell you that they aren’t taking the same brute-force architectural approach that Nexsi took.”

— Marguerite Reardon, Senior Editor, Light Reading
http://www.lightreading.com