Open Source Weblog

I of course encourage you to take a look at the feature mentioned above, not least because it includes some insight from Robin Vasan, managing director at venture capital firm Mayfield and Bernard Dalle, general partner with Index Ventures, as well as the CEOs of EnterpriseDB, Palamida, OpenLogic, and MySQL.

In the meantime, here’s some more detail on the numbers:

As you can see from the headline figures, the growth from 2005 to 2006 is less than we originally stated. We’ve got to put our hands up on that one. The issue was an underestimation of the number of deals done in 2005.

That has now been put right thanks to our research, as well as the figures collected by Matt Asay and Robin Vasan a year ago, which we have also double-checked and verified.

These are figures from 2000 to date, so the latest rounds for the likes of GroundWork and GreenPlum are included, although the very early rounds for the likes of Red Hat, VA Linux (VA Software) and LinuxCare (Levanta) are not.

While we have some figures for the period between 1997 and 2000 we cannot in any way claim these to be conclusive due to the finer details being hard to track down. What we can say that is that at least $2.02bn has been invested in open source vendors since 1997.

In addition to VC funding, we also took a look at the amounts raised by open source vendors via initial public offerings. These figures are not included in the totals of $1.89bn or $2.02bn, but for those that are interested, between them Red Hat, VA, Caldera, Mandriva, Turbolinux and Trolltech raised $319.9m via their IPOs.

The inclusion of Caldera (now SCO Group) raises the question of what constitutes an open source vendor. Clearly any money raised by Caldera counts, as it was money invested in Linux, whereas funding for SCO Group was excluded.

A more difficult decision was whether to include vendors that have open sourced some proprietary code but have for the most part traditional software business models. The choice we made was to include them, as the funding they have raised is at least helping to support open source code.

Another decision regarded vendors such as EnterpriseDB, which has offers closed source extensions based on open source code. Again, these were included due to the fact that they are reliant on open source code, but are also contributing back (in EnterpriseDB’s case to PostrgreSQL) so the funding they have raised is again at least helping to support open source code.

It should also be mentioned that we included hardware vendors where their business is totally reliant on Linux, as investment there is a vote on confidence in the ability of Linux and open source to support their business models.

You might disagree with these decisions or want to slice and dice the figures your own way. We are happy to oblige. Given that much of the data we used was originally collected and published by Matt Asay and Robin Vasan, we’ll be publishing our data for 2000-2006 in the same format they did, look out for a follow-up post with that information later today.