Apollo Sun Silicon Graphics and NeXt the Ascent the Metamorphose and the Fall


(T) I will always remember the Monday morning in 1989 when I was “Up in the Air” traveling from Paris to London with my general manager and learning from the Financial Times the acquisition of Apollo Computer by HP for $476 million. At that time, the computing industry was still very young and M&As were not common practice. Not to mention the high price paid by HP for this acquisition that surprised both of us. Apollo was the first workstation vendor two years before SUN Microsystems.

During my studies in computer science, our school lab was equipped with Apollo workstations and we were the first class to learn C and UNIX. Before us, other computer science grads learned Fortran and Multics (although we learned in my class C, Pascal, Lisp, and Prolog, at that time the only programing language used in the industry was Fortran).

After my computer science degree, when I interviewed for a position at Apollo Computer, I was asked what was the major difference between Apollo and Sun. I tried to answer the question by pointing out differences into their implementations of the Unix operating system but I lost myself in my response. My interviewer with this question was only trying to point out the culture difference between Apollo which was an East Coast-based company and Sun which was a Silicon Valley-based company.

Apollo was founded in Boston in 1980 by William Poduska an ex-Prime employee (Prime Computer was a mini-computer manufacturer; other mini-computers manufacturers from the 1980s included Digital Equipment, Data General and Wang).

The first Sun (Stanford University Network) workstation was the product of Andy Bechtolsheim in 1982 when he was a Ph.D. graduate at Stanford University. Andy needed a more powerful computer than those available in the marketplace for his Ph.D. and designed his first workstation from spare parts, using the Motorola 68000 for the CPU and Unix for the operating system. With two other Stanford graduates, Vinod Kholsa and Scott McNealy, and Bill Joy a graduate from UC Berkeley and a major contributor to Unix BSD (Unix implementation from UC Berkeley), Sun was launched. One of the best ways to start a company is to develop a tool that can become a product for which a large potential market is indeed in need. That was exactly why Sun was started. And, the workstation market grew fast and became quickly one of the fastest growing markets in the computing industry.

In the meantime at Stanford (again!), Jim Clark a Computer Science professor designed the very first VLSI chips that accelerated geometric computations required to display three-dimensional images. Jim foresaw a market segment in need of a new kind of workstation that could be used for high-end visualization applications. And, Silicon Graphics was born as a new competitor to both Apollo and Sun.

After leaving Apple in 1985, Steve Jobs with a few Apple employees created: NeXt with some initial seed money from Ross Perot. Initially, Steve wanted NeXt to develop a new generation of powerful computers for the academic market. But NeXt changed its course and became the last entrant to the workstation market competing with Apollo, Sun and Silicon Graphics.

The NeXt workstation was, without doubt, the most elegant workstation ever designed. It used the Mach kernel developed at Carnegie Mellon for its operating system: the NeXtSTEP. The NeXt workstation was used by Tim Berners-Lee at the CERN and became the first Internet server. While Steve secured a $100 million investment from Canon to manufacture NeXt workstations in Fremont, CA, NeXt never reached any significant market share compared to Sun and Silicon Graphics. Finally, Steve withdraws NeXt in 1993 from manufacturing hardware and re-engineered the company around NeXtSTEP and its object-oriented user interface which later became OPENSTEP and got adopted by many leading software development organizations.

And 1996, Apple acquired NeXt in order to use OPENSTEP as its next generation of operating system. After the acquisition by Apple, Steve first became the Apple interim CEO and later in 2000 the Apple CEO while still being the CEO of Pixar. It goes without saying that Steve learned from his youngest mistakes both at Apple and NeXt to transform Apple which was close to going out of business in 1998 to the most innovative high-tech company of this century and probably for the high-tech industry.



Jim Clark at Silicon Graphics did a fantastic job to execute on the vision that he imagined while being a professor at Stanford University. Silicon Graphics’ IRIS workstations captured the high-end graphics and visualization market. The initial machines were based on the Motorola 68000 processor like Sun’s. In 1992, Silicon Graphics bought MIPS to redesigned its workstations around MIPS’ RISC architecture. The quality of the Silicon Graphics 3G graphics became a brand for the production of Hollywood movies’ visual effects, in particular, Steven Spielberg’s Jurassic Park was designed on IRIS workstations. But Jim Clark, very savvy at capturing new market trends, foresaw very quickly that PCs were not far distant to compete effectively with Silicon Graphics workstations at a much lower price point. In 1993 in disagreement with the other board members about the direction of the company, he left Silicon Graphics to start with Marc Andreessen Netscape. After the departure of Jim Clark, Silicon Graphics moved aggressively into the high-end supercomputer market to compensate the lack of sales in its workstation business by first leveraging its internal high-end computing technology and further by purchasing in 1996 supercomputer vendor Cray Research for $740 million. But over time, the supercomputer market became under attack from high-end commercial servers from Sun and IBM which later became under attack themselves from clusters of off-the-shelves Linux-PCs based systems. Silicon Graphics finally abounded its MIPS RISC processors to adopt Intel’s RISC processors. It re-branded itself to SGI. But over time, SGI’s high-end servers sales decreased to a point where the company in 2008 filled for bankruptcy and sold its assets to Rackable Systems for $25 million. SGI’s last headquarters on Amphitheatre Parkway in Mountain View, CA became Google’s Googleplex.

For the first two years, Sun’s CEO was Vinod Koshla but Vinod left Sun to join top venture capital firm Kleiner Perkins Caufield and Byers and asked the energetic and charismatic Scott McNealy to become the CEO. In 1988, Sun hits a $1 billion dollar revenue. For its first decade, Sun was predominantly a vendor of workstations for engineers and scientists competing with Apollo and Silicon Graphics on lower margins. Sun branded itself as a leader of open standards and computer connectivity with the message “the Network is the Computer”. Sun invented a new network file system: NFS and integrated very quickly TCP/IP, the Internet protocol, into its UNIX kernel. In its second decade of operations, Sun moved aggressively into servers and storage leveraging the move to client/server architectures. Scott McNealy’s strategy was pretty simple: “not being a car dealer but thriving as a car manufacturer” in other words for Scott owning core computing technology was an absolute priority for the company; as such the company abandoned Unix BSD to create its own operating system: Solaris and developed its own line of RISC processors: the Sparc.

During its second decade of operations, Sun grew very quickly by moving from the technical computing market to the commercial computing market and having its line of servers becoming well-adopted by database and enterprise applications vendors such as Oracle, Ingres, Informix, Baan, SAP, and many others. Sun execution became flawless. It launched Java in 1995, a new network programing language created by James Gosling that was quickly massively adopted by all software engineers and developers around the world and became a ubiquitous tool to develop emerging Internet and intranet applications. Sun was very successfully ridding the Internet wave. Its servers were used widely worldwide for Web applications. Scott’s marketing messages were simple as was his strategy: “Sparc is good, Solaris is good, Java is good and Microsoft is bad”. Scott never considered seriously IBM, DEC, HP as competitors. For him, the only competition to Sun was Microsoft. And, Sun’s mission was to relieve the computing world from Microsoft’s oppression.

But after the Internet bubble, major companies started slowly to replace Sun’s Internet servers with clusters of Linux PCs. Sun was much too slow to react against the adoption of Linux and lost gradually its profitable competitive edge. Sun’s technology investments were unprofitable: Sparc was losing against Intel’s processors, Solaris was losing against Linux and Scott was never able to make a business from Java. In addition, IBM and HP became stronger competitors taking market shares from Sun. Add to that, a lot of acquisitions that hurt Sun (Cobalt Networks for $2 billion, StorageTek for 4.1 billion and MySQL for $1 billion) and a complete demotivated workforce which has been laid off quarter after quarter as “RIF” (reduction in force) and Sun entered and stagnated in its third decade of operations from one of the most flawless high-tech companies to execute that it was during its second decade to become one of the high-tech companies to fail to execute properly. And, I still do not understand why Scott could not turn around the company or found someone who could turn around the company.

Finally, in April 2009, Oracle announced the agreement to purchase Sun for $7.4 billion. Sun became completed absorbed by Oracle on January 27, 2009. And with that ending, all workstation vendors have disappeared from the marketplace.

So what are the conclusions of the ascent, metamorphose and fall of Apollo, Sun, Silicon Graphics and NeXt. Let me suggest three conclusions.

The first one is that, if technology is complex, business strategy is simple. There only a few reasons why a high-tech company succeeds or fails. And, it is always a simple combination of a few unique facts that are centered around technology, product, market, and management execution.

The second one is the incredible technology contributions of those companies in:

  • Software: Unix, C programming, the Unix Shell, the text editor VI, Java, distributed operating systems…
  • Networking: NFS, TCP/IP, Ethernet, mail…
  • And in hardware: RISC processors, graphics processors, multiprocessors, parallel computing, computer clusters…

And, my last one is the impact of the high-profile founders and CEOs of those companies. With Scott McNealy and Jim Clark so prominent figures of the high-industry, with Bill Gates has retired from Microsoft, and Steve Jobs being unfortunately sick and less involved in the operations of Apple, the high-tech industry is definitely losing those who started the computing industry.

Note 1: I have been both an employee and a client of Sun Microsystems.

Note 2: The picture above is from Sun; from left to right, Sun’s founding team: Vinod Khosla, William Joy, Andreas Bechtolsheim and Scott McNealy with the Sun-1 workstation in 1982.

Note 3: The first picture in the article is the NeXt workstation from Tim Berners-Lee and the second one is his description of his invention of the Web on the NeXt User Interface. Those pictures are from Wikipedia.

Copyright © 2005-2010 by Serge-Paul Carrasco. All rights reserved. Contact Us: asvinsider at gmail dot com.