Back in 1998, four of the biggest names in enterprise computing agreed on something wild. IBM, the Santa Cruz Operation (SCO), Sequent, and Intel would merge their separate UNIX systems into one OS for Intel’s upcoming 64-bit chips. They called it Project Monterey.
On paper, it should have worked. It had money, market share, and top engineers, all trying to patch up UNIX’s fragmented family tree before Windows NT ate the server market whole. What actually happened is strange. A free hobbyist kernel nobody important took seriously killed it first. Then its own corpse got turned into one of the most expensive lawsuits in tech history.
Four companies tried to build one UNIX to beat them all
A quick gut check on how big this deal actually was
Credit: MaIII Themd/Shutterstock
Signed in October 1998, the Joint Development Agreement divided the work among the companies cleanly. IBM brought AIX, its UNIX for POWER and PowerPC hardware. SCO brought UnixWare and OpenServer, the dominant UNIX on x86 boxes. Sequent added DYNIX/ptx and real expertise in multiprocessing and NUMA memory. Intel funded it all, since the finished OS would run on its new Itanium (IA-64) chips. At first glance, things did look perfect.
IDC recognized SCO as the single biggest UNIX server vendor in 1998, and rightfully so. It held 40.6% of all UNIX shipments worldwide. UnixWare was growing 58.5% year over year, too. This wasn’t an underdog coalition. In fact, the actual market leader was steering the ship.
Related
Linux inherited Unix’s superpower, and developers can’t let it go
Turns out freedom and flexibility are great for coding.
IBM built the coffin with its own billion dollars
The same company, two very different bets
Intel Core Series 3 mobile CPU.Credit: Intel
While Monterey inched through committee meetings, a scrappier effort called Project Trillian quietly ported Linux to IA-64 itself. It shipped working code in February 2000, beating the “real” Monterey OS to market. IBM noticed this incident. On December 12, 2000, CEO Louis Gerstner made an announcement in New York. He said that IBM would spend $1 billion on Linux in 2001. It was the same week it revealed a Linux supercomputer deal with Shell Oil.
Then Itanium fell apart. Intel had promised the chip for 1999. However, it didn’t ship until mid-2001. Roughly two years late. It landed to a collective shrug. Monterey’s whole reason to exist was tied to a chip nobody wanted yet. IBM reportedly sold only around 32 Monterey licenses that year.
Here’s the part that gets me every time. IBM didn’t lose interest in Monterey by accident. It helped found the project, then quietly built the thing that would replace it. It was like watching Linux’s three-decade climb from hobby project to industry backbone from the inside. I won’t call that betrayal. It’s just IBM reading the market faster than its own joint venture could move.
When the product died, SCO discovered litigation paid better
Revenue collapsed, then a very different business plan
Credit: Viktory Viktor / Shutterstock.com
Caldera, a small Linux distributor, bought SCO’s old UNIX business and renamed itself The SCO Group in 2002. Unfortunately, the timing was rather rough. Revenue had fallen from $170 million in 1999 to about $70 million in 2001. SCO was bleeding roughly $5 million a quarter. In June 2002, Darl McBride took over as CEO, replacing Caldera co-founder Ransom Love. Love says he warned him directly. McBride did it anyway.
In January 2003, SCO launched SCOsource. It was a licensing program aimed squarely at Linux users. Two months later, it sued IBM, demanding at least $1 billion. Later, that figure rose to $5 billion. Any company running Linux could pay SCO $699 per server to avoid getting sued next.
SCO’s technical case was thin, too. Its method for finding “stolen” code amounted to searching the Linux kernel that actually powers all this for the strings “JFS,” “NUMA,” and “RCU,” then claiming every matching file as evidence. Linus Torvalds picked the list apart himself, writing he could “totally demolish the SCO claim that these files were somehow copied.” Somewhere here, the story stops being about UNIX and starts being about desperation.
Related
5 pioneering Linux distros that quietly faded into history
Explore the rise and fall of these groundbreaking Linux systems, and how their legacies live on.
SCO wasn’t entirely making it up
The one claim that actually had legs
To be fair, one part of SCO’s case held together better than the rest. Under the Monterey agreement, IBM could only use SCO’s UnixWare code in its own AIX for Power product once an IA-64 version shipped and became generally available. SCO alleged IBM released a deliberately crippled, unsupported “beta” of Monterey on May 4, 2001, purely to satisfy that clause, then shipped AIX for Power with the disputed code that same day.
An internal IBM email that later surfaced didn’t help. One employee wrote that releasing the beta was necessary to gain rights to the Santa Cruz code IBM wanted for its base AIX product. He added, “I know the fine lines we are walking here.” A federal appeals court found the claim credible enough to revive it in 2017. That’s over a decade after the original filing.
One real grievance doesn’t excuse a $5 billion shakedown
How the whole campaign actually ended
Credit: Valery Evlakhov/Shutterstock.com
If you zoom out, you’ll realize something interesting. That narrow claim was more of a footnote inside a case trying to tax every company on the planet running Linux. In August 2007, a federal judge ruled that Novell, not SCO, actually owned the UNIX copyrights SCO had been suing over. Court records later showed McBride had personally emailed Novell in 2003. He asked them to amend an old agreement to hand SCO the copyrights it claimed to already own. SCO filed for bankruptcy weeks after losing.
The Monterey claim limped through appeals for another decade. In 2021, IBM finally settled it for $14.25 million, no fault admitted, closing out a case that started with demands as high as $5 billion.
Do that math for a second. Five billion demanded, fourteen and a quarter million collected, eighteen years later. Along the way, leaked internal Microsoft memos revealed exactly how worried Redmond was about Linux. This says a lot about how threatened the old guard felt by what Torvalds built for free. Even SCO’s one legitimate grievance wasn’t worth what the gamble cost the company.
Losing to Linux was fair, but suing over it wasn’t
Project Monterey deserved a better ending. It was a serious attempt to save UNIX from its own fragmentation. What still bugs me is what SCO tried to do with the wreckage instead of just walking away. The code that actually won runs on your phone, your car, and probably whatever server delivered this page to you. On top of that, it’s worth more today than it’s ever been.