In Context: Follow the Money; Cathedral Investments in the Bazaar

When you attempt to “follow the money” being invested in
Linux startups by established technology companies, you find these
“tech giants” are motivated by all sorts of considerations when
they pour millions of dollars into high-risk ventures based on a
radically different way of making money off of software. What it is
very refreshing to note you don’t find is an attempt to
use these investments to dominate markets. Also of interest:
several Linux startups seem to have done quite nicely without any
corporate investors.”

by John Wolley, Linux
Today

In early October, TurboLinux announced it had received $30M in
funding from “tech giants” including IBM, Intel, Fujitsu, Hitachi,
and SGI, among others. The event was hardly news — none of the
stories about it posted on Linux Today drew as many reads as the
press release titled “TurboLinux Tours with IBM to Showcase Linux –
‘Mission Possible’. “

It was quite a different story two years ago. The announcement
in late September, 1998, that Intel and Netscape were making direct
investments in Red Hat was a watershed event in the emergence of
Linux into the awareness of the wider technology business world
outside of the traditional “Linux community.” In the now familiar
terminology of “The Cathedral and the Bazaar,” traditionally
proprietary, “Cathedral” companies were pouring major sums of money
into a radically new, open source, “Bazaar” company. Some of the
stories about this posted on Linux Today set records for the time
for the number of pageviews.

Although the details on the exact amount were not disclosed (as
seems to be standard for these investments), it was inferred that
millions of dollars were involved. That was enough to pique the
curiosity of financial analysts. And it sent a signal to the
mainstream technical trade press that something big was
going on with Linux.

Investments in startups by established companies are nothing
new. It’s a reflection of the way companies operate these days, not
trying to do everything from scratch within one corporate entity,
but forming ad hoc alliances with “partners” whenever and wherever
that appears to be mutually rewarding. Established companies like
Intel typically invest in startups when they see the possibility
that what the startup is trying to do will create or expand markets
for their products.

What’s different about the established companies investing in
Linux startups is the Cathedral/Bazaar divide. With these Linux
companies, the startups are trying to implement a whole different
business model from what the established companies investing in
them are used to, and this translates into a much higher level of
risk. It also translates into a much higher potential
return on investment, if it the new open source
business models do, in fact, turn out to be the “wave of the
future,” the “disruptive technology” that obsoletes the traditional
business models based on proprietary software.

Over the months following the Intel/Netscape announcement, the
floodgates seemed to open on investments by Cathedral companies in
Linux startups. It quickly provided a potent argument to counter
the FUD point about “no big companies backing Linux” — here were
all these big companies putting their money where their mouths
were, pouring millions of dollars into Linux
companies.

What’s the Motivation?
What has motivated the established technology companies — just
about everybody but Microsoft — to pour millions of dollars into
Linux startups that are widely regarded as high risk
investments?

Microsoft’s dominance in the desktop OS and applications
markets is so complete that there’s little opportunity for other
companies to make money selling software in the traditional
Cathedral model — this makes them very open to
considering the alternative Bazaar model.
Potential Microsoft competitors see a real possibility of
breaking Microsoft’s Windows/Office monopoly, which would open up
enormous markets that have been closed for more than a decade —
the fact that they might have to make their money selling services
instead of software isn’t a show-stopper.
The close working relationships that inevitably result from
investments will give the investing companies an inside track on
servicing the developing open source markets and making the
transition to open source business models, if the Linux
revolution really takes off, forcing established businesses to
adapt in order to survive.
Companies investing in a Linux startup generally receive, in
exchange, “equity” in the company — shares of the startup’s
privately held stock — which can potentially be worth many times
what the company invested if the startup has a successful IPO.
A few million dollars is “small change” to a company like Intel
or IBM — it’s even relatively small change to a companies
like Novell or SGI, who have been having trouble financially at the
same time they’ve been investing in Linux startups.

What’s the Impact?
What is the significance of the money to the Linux startups? Of
course, a few million dollars to work with is a nice thing for
anyone trying to get a business started, but what’s the
significance of this money coming from established technology
companies, instead of from “angel investors” or wherever else
companies get money from before they go public?

The association with established, financially successful “tech
giants” like Intel and IBM gave the Linux startups instant
credibility in the financial world.
The big companies that have invested in Linux startups now have
a vested interest in seeing those startups succeed — they have
every incentive to “throw a little business their way” and, if
necessary, to pump in some more cash, if that’s what it takes.
When a Linux startup is ready to go public, that backing by
“tech giants” greatly increases the interest of the financial
community — it’s highly unlikely that the Red Hat and VA Linux
IPOs would have been as spectacularly successful without such
backing and the investor confidence that it inspired, although
Cobalt Networks and Andover.net seemed to do just fine without
it.

What’s Involved if You Want to “Follow the
Money?”
If you were to “follow the money” with Microsoft and its
investments in various companies, there’s a (surprise!) very
sinister dimension to the story — it looks like MS is trying to
buy its way into control of just about everything it doesn’t
already control: cable, satellite, wireless, content providers, you
name it.

The absence of any comparable pattern in the investments by
Cathedral companies in Bazaar companies is a refreshing change.
What you don’t see is evidence of a big company trying to
use its Linux investments to dominate a market, the strategy we’ve
come to expect with Microsoft. For example, although Red
Hat is the Linux distro in which IBM appears to have made the most
direct investment, IBM has been very scrupulous about supporting
multiple distros on its products and is pre-loading Caldera on its
ThinkPad laptop.

When I first got the idea to try to “follow the money” going
into Linux startups for a story, I envisioned tracking down and
totaling up the dollar amounts — how much total from which
companies has gone into which Linux companies. I knew the exact
amounts were never specified in the press coverage, but assumed it
wouldn’t be that hard to track down in the investing companies’
quarterly reports; after all, they’re all public companies so they
can’t keep these details secret for long. I quickly learned two
things:

For whatever reason, the exact numbers are not as easy
to locate as I had imagined. The public companies doing the
investing are typically doing it through an investment subsidiary,
which tends to show up as a single line entry in the quarterly
report. I need to do some more research here: whenever information
is not easy to get, I tend to suspect someone must be
trying to hide something; but so far I can’t see what it may
be.
The number of companies investing in Linux startups is
substantial — covering all of their investment activities in any
detail would require a whole book!

So the rest of this article is not a comprehensive
summation of corporate investment in Linux startups — it’s just a
summary of the highlights of six “Linux companies:” Red Hat, Cobalt
Networks, VA Linux, Andover.net, Caldera, and TurboLinux. The
companies are detailed below in the order in which they went public
in 1999, or by the likelihood of their going public in 2000 or
2001.

Red Hat
Intel and Netscape made the first Cathedral investment in a Bazaar
company when they invested in Red Hat in September of ’98.

IBM, Compaq, Novell, and Oracle announced investments in Red Hat
during the first LinuxWorld Conference and Expo (LWCE) in March of
’99. SAP, the German ERP software house, added its financial
backing to Red Hat a few weeks later, followed by Dell in early
April.

Novell ponied up a second sum for Red Hat in July of ’99.

In August of ’99, Red Hat went public. The first day run-up in
the stock price set unrealistically high expectations not only for
Red Hat, but for all the Linux companies queued up behind it with
plans to go public in 1999 and 2000. The IPO made Red Hat a public
company — people and companies wanting to invest in Red Hat today
buy its stock on the open market.

In April of 2000 Red Hat turned around and invested some cash in
Rackspace.com. Then in May it formed Red Hat Ventures to funnel its
investments into open source startups, which “will take $500,000 to
$2 million stakes in new companies specializing in open-source
software and Internet infrastructure technology.” According to the
press release, “Red Hat has already taken stakes in a number of key
Open Source-related startups, including Sendmail Inc.,
Rackspace.com, ZoneTrader.com, techies.com and e-smith,” prior to
creating the venture fund.

Cobalt Networks
Cobalt’s pre-IPO funding didn’t produce any stories that got posted
on Linux Today. A search of CNET’s archives turned up one “short
take” in June of ’99 that described a $36 million round of funding,
all from venture capital groups — no established tech giants
participating.

In November of ’99 Cobalt became the “second Linux IPO,”
apparently without any corporate investors pre-IPO, and bested Red
Hat’s first day performance.

In March of 2000 Cobalt used some of the cash raised in its IPO
to purchase Chili!Soft, based in Bellevue, Washington. Cobalt had
been bundling Chili!Soft’s software on its server appliances to
support Microsoft’s Active Server Pages (that .asp extension you
see on web pages) on the Linux platform — this allows developers
using Visual Basic and other Microsoft tools to develop web server
applications for the Linux platform without switching (and
relearning) development tools.

In September Cobalt announced that it was being acquired by Sun
Microsystems for some $2 billion in Sun stock, a considerable
premium over the price at which Cobalt stock was then trading. What
happens with Cobalt under Sun’s ownership is still an open
question, but for Cobalt’s investors the Sun purchase was very
rewarding.

VA Linux
VA Linux was still known as VA Research when Intel invested in
early March of ’99, right before the first LWCE. The Intel cash
presumably was instrumental in allowing VA to make quickly make
three acquisitions: Electric Lichen, at the end of March; and Linux
Hardware Solutions, of Wilmington, N.C., and Enlightened Solutions,
based in Atlanta, at the end of April.

In June of ’99, right after VA Research had changed its name to
VA Linux, SGI and Intel joined in a second round of funding that
totaled $25 million.

In December of ’99, VA Linux went public with a bang — its 698%
first day gain broke all records for first day trading of a public
company’s stock up to that time.

In February of 2000, VA Linux surprised the Linux community with
its LWCE announcement of the purchase of Andover.net “in a cash and
stock deal worth roughly $813 million,” based on the value of VA’s
stock at the time. By the time the deal closed in June, stock
values had dropped significantly — terms called for “each share of
Andover.Net stock to be exchanged for 0.425 of a share of VA Linux
Systems, making the deal worth about $107 million.”

In March of 2000, VA Linux fleshed out its hardware offerings by
acquiring TruSolutions, a manufacturer of rack-mount servers, and
NetAttach, which makes high availability storage devices. At the
current VA stock price, the cost to VA in stock and cash was $194
million for TruSolutions and $39.2 million for NetAttach. By
September of 2000, VA had turned the NetAttach acquisition into a
NAS (network attached storage) product that was ready to
launch.

In June of 2000, VA Linux invested in the MySQL open source
database project and agreed to host the project on SourceForge.net.
In September and October, VA announced plans to expand its
marketing efforts in Japan and Europe.

Andover.net
Andover.net doesn’t fit the mold of the other Linux startups that
have received funding from established companies. Although it
succeeded in going public, it apparently did so with the least
corporate investment of all the Linux IPOs in 1999 — none. I
checked the archives of ZDNet and CNET to make sure that there
wasn’t some major investment that somehow never got posted to Linux
Today.

Without an exhaustive search on Andover across the ‘net, I still
regard this as a tentative conclusion, but it does make sense:
Andover was strictly a news operation whose revenues came from
advertising; it didn’t have any products or services that an
established technology company could partner their products with,
and its success didn’t promise to open up new markets which those
companies could go after

In June of ’99, Andover.net seemed to come out of nowhere, with
cash from who knows where, when it purchased Slashdot. This was
followed shortly in August with the purchase of a second Linux
site, Freshmeat.net.

In December of ’99, just a few days after the spectacularly
successful VA Linux IPO, Andover.net’s IPO took place. Andover’s
first day run-up of 350% was outstanding, although overshadowed by
the 698% racked up by VA Linux a few days earlier. And it was
apparently pulled off without any major investment from the
established tech giants.

In January of 2000, Andover followed the pattern of other Linux
companies that had IPOs by using some of the cash raised on an
acquisition: it purchased QuestionExchange, a website which “lets
people post questions and others submit bids for answers.” This was
about the only way in which Andover’s story was similar to the
other Linux IPOs of 1999.

In February, during the third LWCE, Andover surprised the Linux
community with the announcement that it was being bought by VA
Linux “in a cash and stock deal worth roughly $813 million,” based
on the value of VA’s stock at the time. In June the sale to VA
Linux was completed and Andover stock ceased trading, almost
exactly six months after the IPO.

Caldera
Caldera got its initial financing from a venture capital group in
February of ’99. In January of 2000 Sun, Citrix, Novell, and SCO
joined with two venture capital groups to invest $30 million in
Caldera.

In March of 2000 Caldera went public in an IPO that was
headlined variously as “Caldera’s IPO Makes Strong Debut” and
“Caldera IPO Marks First Linux Disappointment” — its first day
run-up was “only” 110%.

In August of 2000 Caldera announced that it was purchasing the
Unix and professional services units of SCO.

In September Caldera invested $3 million in EBIZ Enterprises “to
develop PartnerAxis, a Web-based B2B entity providing knowledge
exchange, Linux product sales, advertising, membership and channel
development.”

TurboLinux
TurboLinux started out as Pacific HiTech, headquartered in Asia. It
opened US offices in October of ’98 and changed its name to
TurboLinux, after the distro, in May of ’99.

In October of ’99, Intel invested in TurboLinux, along with
several venture capital groups. The total amount was not disclosed,
but one VC group put up $5.5 million.

In January of 2000 TurboLinux received more than $50 million in
another round of funding from more than a dozen companies,
including: Dell, Compaq, BEA Systems, Seagate, SCO, Novell,
Toshiba, and NEC.

In October of 2000 TurboLinux received additional funding
totaling $32 million, from: Fujitsu, Hitachi, IBM, SGI, Dell, and
Intel. By latest accounts, this funding puts TurboLinux in position
to move on to an IPO, as soon as the market warms up to IPOs
again.

Linuxcare
Linuxcare got its initial financing from venture capitalists in May
of ’99. In December of ’99 Linuxcare received $32.5 million in
investments from Dell, Motorola, Oracle, and Sun, plus several
venture capital groups.

In August of 2000 Linuxcare received an additional $30 million
from venture capital groups.

Conclusions
Established tech giants have invested heavily in some Linux
startups, while others have had to raise all their funds through
venture capital sources. The corporate funding has undoubtedly
helped those Linux startups which have received it, but several
have done surprisingly well without it. Aside from Sun’s outright
purchase of Cobalt Networks, corporate investors in Linux startups
have generally not used their investments to try to
control any of the markets in which the startups are involved.

Related Stories — Linuxcare:

CNET News.com: Linuxcare looks to future with new financing(Aug
02, 2000)
E-Commerce
Times: Linuxcare Lines Up Major League Funding(Dec 15,
1999)
CNET:
Kleiner Perkins invests in Linuxcare(May 03, 1999)

In Context: Follow the Money; Cathedral Investments in the Bazaar

Web Webster

Company

Categories