Liquid Computing & the Curse of a Computer Hardware Startup

It’s hard out there for a systems vendor — that is, if the death of Liquid Computing, an Ottawa, Ontario-based startup that until last week was building a unified computing box to help manage the virtualization of the data center, is any indication. Last week, a round of funding failed to materialize at the last minute and the company was shut down, confirmed Liquid’s now former CEO, Vikram Desai, who subsequently found himself out of a job.
Liquid was one of many startups trying to build a box to improve performance inside data centers. In Liquid’s case it was trying to address some of the complexity introduced in networking and storage once servers were virtualized. But like many other startups targeting corporate data centers, Liquid had a hard time selling its wares.
Desai was hired in December 2008 to help Liquid move from selling proprietary boxes to selling its software optimized for generic hardware made by Intel (s intc) and NetApp. “It’s very risky for large enterprises to adopt the hardware platforms of young companies as their primary solution, and moreover, with the release of advanced hardware platforms from companies like Intel, there isn’t a need for them,” Desai said. “The most successful companies in the data center will be software-centric.”
Desai didn’t get into details about Liquid’s demise but the Ottawa Citizen reports that the 7-year-old startup laid off most of its Ottawa workers last week, and that it’s in the process of shutting down because it was unable to find additional funding (hat tip insideHPC). It was building a server integrated with networking capability much like Cisco’s  (s csco) Unified Computing Systems or HP’s (s hpq) efforts to add networking to its servers.
As a startup, the company was pitting itself against giants, and had raised more than $50 million in order to outrun them. But despite that fact that its products had advantages that Cisco’s and HP’s products don’t, and the company was winning customers, Liquid’s backers weren’t willing to front any more cash. A bankruptcy isn’t in the cards, but the company may sell its intellectual property.
The inability to raise more funding signals the challenges that even large systems startups still have when it comes to raising capital. SiCortex, Quadrics, Verari are other examples, either because they were unable find money of their own or because their investors were skeptical of the returns they might get and didn’t want to keep throwing good money after bad. In fact, after Verari shut down, I wondered if Liquid and BLADE Network were next.
SeaMicro and Smooth-Stone, both of which are building server systems for the data center, may also be having have troubles. Sources tell me that Smooth-Stone, a startup manufacturing ARM-based servers in Austin, is only building out a system to show off its software, a practice that can require a lot of up-front capital. Desai said that Liquid found itself forced to do the same thing in order to sell its software.
Liquid spent 2009 shifting from building hardware to selling software, but the venture firms were unable or unwilling to give it more time. However, Desai is still optimistic about the prospects for unified computing, which he believes is one of the cornerstones of the next-generation data center. “We are still at the early stage of unified computing, but that and virtualization are the key to truly deliver an automated data center.”
Related GigaOM Pro content (sub req’d):
If We Compute in the Cloud, We Need a Network Fabric
Image courtesy of Flickr user br1dotcom