One of the more curious dotcom era startups was VA Linux. In a way, it makes sense, because so much of the dotcom boom was powered by Linux, so we’d expect a Linux company to do well in those times. But VA Linux operated on a business model that doesn’t really exist anymore. Its December 9, 1999 IPO set a record for its time.

Linux in the dotcom era

Arguably, the most successful Linux startup of the dotcom era was Red Hat. Red Hat survived as an independent company for nearly 20 years, before being purchased by IBM in 2018. IBM needed to modernize its portfolio, and Red Hat was a good fit.

But during the dotcom boom, Red Hat wasn’t the strongest performing Linux startup. VA Linux held that title, with its stock price surging 698% on December 9, 1999, the day of its IPO, setting a record for the time. Its share price shot up from $30 to $239 per share. It raised $132 million, nearly double what Microsoft raised in its IPO. And its market capitalization reached $9.5 billion, 12 times what Microsoft’s market cap was after its IPO.

Not bad for a company that still hadn’t turned its first profit yet. To draw a more recent parallel, early in the dotcom bubble, VA Linux looked to investors like it might become like AMD or Nvidia during the golden age of AI, or arguably better since it offered the opportunity to get in on the ground floor.

Investors were smitten, at least at first. But a year later, it was trading at $8.49 a share. At some point, investors remembered that a company needs to make money to be worth $9.5 billion.

What did VA Linux do?

Founded in November 1993 as VA Research, VA Linux sold professional workstations and servers with Linux pre-installed. At its peak, it held about 20 percent of the Linux hardware market. It started when the company’s founder, Larry Augustin, wanted a Sun workstation but couldn’t afford one, so he spent $2,000 building himself a 486-based PC and loaded Linux on it. Then he and his friend James Vera started selling computers on campus, Michael Dell-style.

This may seem odd, because today, Linux has a reputation for running on anything. Take a random computer some business threw away, load Linux on it, and you get a computer for free. So a business model selling Linux computers for $2,000 seems odd today.

In the 1990s, Linux hardware support was much spottier than it is now. Today, it’s a little unusual for something you buy not to work with Linux. In some cases, it’s easier to get certain older hardware running on Linux than it is on Windows, even if the peripheral was originally designed for Windows. But in 1999, you couldn’t take any of that for granted. Some hardware had excellent Linux support, some didn’t work at all, and some worked, but not as well as you might like.

It was a lot easier to just buy a pre-built system from VA Linux with your choice of Linux operating system pre-installed, and then you knew everything was going to work really well. It saved you a lot of time and effort, because you didn’t have to read endless mailing lists or Usenet groups to find out what works well and what didn’t. VA Linux did all of the research ahead of time, built systems with best of breed components, and shipped it to you.

It especially appealed to business customers, because time is money. They may have had staff who was capable of doing that research, but they had better things for their staff to be doing. It was much more cost effective to just pay a couple of hundred dollar premium for a system from VA Linux that just worked. Its customers included fellow dotcom startup etoys.com.

The problem for VA Linux

The problem for VA Linux was its business model wouldn’t last forever. Eventually, the larger, more established computer makers were going to realize they were leaving money on the table by not competing with VA Linux.

There is no rule saying that a system that runs Linux really well can’t run Windows really well also. In fact, usually a system that runs Linux really well does also run Windows really well. That was true both then and now. So all a company like Dell or HP really needed to do was build a class of systems using the same hardware VA Linux was using, and then offer those systems with your choice of Windows or Linux. And then you had a system that would run just as well as the VA Linux system. But because of the larger vendor’s volume, it probably cost less.

Over time, Linux hardware support improved too. So VA Linux’s business model of selling Linux PCs was being eroded from two sides by two different factors.

What happened to VA Linux after its IPO

I’m not sure the dotcom investors realized VA Linux was operating on a business model that had a limited shelf life. I know plenty of people were excited about that stock in 1999. But I don’t know anyone who worked in the IT field and really knew a lot about hardware or Linux who invested in it. We should have been shorting the stock, frankly, but none of us knew how to go about doing that.

VA Linux abandoned the hardware business in June 2001, reinventing itself as an operator of technical websites. In effect, it became a lot more like its former customers.

By December 2002, VA Linux was trading at $1.19 a share, well below its $239 per share IPO price. It changed its name to VA Software, then to SourceForge Inc, and, later, Geeknet, to better reflect its business model. On June 2, 2015, video game retailer and future meme stock Gamestop announced it would acquire Geeknet for $20 per share, or $140 million. The deal closed July 17, 2015.

In the end, VA Linux did better than many of its dotcom-era peers. It survived as an independent company for 16 years. While it sold for less than its IPO price from 1999, it rebounded and sold for considerably higher than the $1.19 it was trading at during its low points.

Related stories by Dave Farquhar