September 23, 2004

How Vinod Khosla created Sun Microsystems

While I knew the one line description "Vinod Khosla was the founding CEO of Sun Microsystems and was earlier part of the founding team at Daisy Systems", I hadn't come across a more detailed version of Khosla's pre-KPCB exploits before Joe Kraus talked about it on his blog.

Here are some extracts from the Harvard Business School case study (by Dr. Amir Bhide) that I found interesting:

How a Stanford secretary "linked up" SUN's co-founders:

I'm probably more of a conceptual engineer, and I can draw block diagrams for almost anything I can think of, but I can almost never implement them. So I started looking for someone who had done this kind of stuff before. I heard of a project at Stanford called the Stanford University Network, or Sun.workstation project. I called the computer science department, and some secretary who did not want to bother a professor gave me the uame of a graduate student from Germany, Andy Bechtolsheim.

Apparently, Andy, who was also at Carnegie at the same time I was, but I did not know him there, had come to Stanford to do his Ph.D. in CAD tools. I think he realized there was no appropriate machine to develop CAD tools, following the same discovery process I had gone through, so he decided to build one himself. His specs fit mine almost to a T.

Andy had developed the workstation concept in a fair amount of detail and had a prototype implementation of lt. Stanford had assigned the technology to him because, in their great wisdom, and after calling DEC and Prime, they had decided it had no value.

So, for over a year he had been licensing the technology to six or seven companies. He had invested $25,000 of his own money into building prototypes, and as a grad student licensing it at $10,000 a pop, he thought that was just wonderful.

Bechtolsheim offered Khosla his usual $10,000 license. Instead, Khosla tried to persuade Bechtolsheim to join forces to start a company to build workstations based on his designs.

I said to him, "I want the goose that laid the golden egg, and I don't want the golden egg." I thought that kind of resource is very rare to get. So I would rather have him than any one design he would come up with. I had nothing very concrete to offer. I told him we could build a big company, that we could raise a few million dollars. He would be a founder of the company.


As an entrepreneur, VCs were good to VK.

Andy Bechtolsheim agreed to participate by late January 1982. The two started working out of Andy's office at Stanford and in a couple of weeks had produced a brief plan.

It was a real concise statement of the reasons for making an investment: how the economics had changed, what the product would be, when it would be out, and how big it could be and why the market made sense.

The next day, February 12, we met with two venture capitalists, one of whom, Bob Sackman, had helped me write the Daisy business plan. Within three or four days, they agreed to give us $300,000 in equity. They gave us a $100,000 check right away and said, "You can get going and let's work on the paperwork." On February 22, we formally incorporated the company and received the remaining $200,000. The price of the stock was $2.75 a share. We also gave them an option to put an additional $2.2 million for a total of $2.5 million at $5.60 a share, that option to expire on June 30, 1982. By that date we were supposed to hire a marketing person, write a business plan, and demonstrate a prototype.

Bob Sackman led the thing, and he trusted me. It was really on trust. There was very little due diligence on their part- - they just believed in the concept and said, "Yes, we think you can do it.


Importance of containing burn:

We were unlike most start- ups. Most start- ups have everything- - marketing, sales, support, advertising, and PR- - in place even before they have a product to sell. They get up to $600,000 to $800,000 a month of expenses before they've really started selling anything. In that range, given you're starting out with low gross margins because your product costs are high, you've got to start selling $1.5 million worth just to break even in 8 month.

Click Here to read the full case study.