Intel’s disruptive move with the 386

A couple of weeks ago I went to see a lecture by Harvard prof Richard Tedlow on how Intel changed the industry with the launch of it’s 386 processor. I wasn’t sure this was going to be all that exciting, but the the Computer History Museum in Mountain View is an interesting place, so what the heck.

It turns out to have been quite a good session. There were many Intel employees and alumni, including a good number of folks from the 386 team, as well as Tom Halfhill from the Microprocessor Report. While Intel’s 386 was a significantly improved processor on the earlier 8088 and 80286, it was the company’s business decisions that proved most significant. The 386 transformed the industry, creating a new order and putting Intel and Microsoft in the driver’s seat for the first time.

But it helps to understand the historical context of where Intel and the industry were in 1985. Intel was by no means the leader in processors. It had just suffered a dramatic decline in sales, suffering the biggest loss in the company’s history. And the cost of developing the 80386 was approximately more than 10 times that of earlier processors. Worst of all, the 386 was still considered by many to be technically inferior to the Motoroloa 68000 family of processors.

At this time, the industry convention was that when a computer vendor purchased CPUs, they would require a “second source” licensing agreement. Second-sourcing came about from the days when military contracts were driving the semiconductor business, and they wanted to ensure that they were not reliant on a single vendor. In effect, for Intel or anyone to sell their chips, they would agree to enable a competitor to license the technology. So in the early days of the 8088 and 80286, Intel played by the same rules as everyone else, setting up companies like AMD and Fujitsu as secondary sources for compatible chips.

But with the 386, Intel took a risky move. Having spent four years and $100 million developing the 386, it decided to not license the technology to competitors. And as it turned out, the company’s biggest customer, IBM, didn’t seem all that interested in the 386, fearing it would hurt sales of Big Blue’s high-end systems. But Compaq bit and became the first computer company to use the new processor. Seven months later, IBM was back at the table, but it never again had a leadership position in PCs. The rest, as they say, is history. By 1992, Intel had 83 percent of the microprocessor market, and more than a $1 billion in annual profit.

But the surprising thing about Intel’s move was it was not at all obvious at the time. Andy Grove admitted in interviews that there was never really any monumental business decision. It was just a series of moves that put the company on a better course.

Two related items for those interested: Tedlow’s book “Andy Grove: The Life and Times of an American” provides a detailed biography of Grove. Tom Halfhill wrote an interesting editorial last fall about what would happen if AMD failed and Intel was truly a monopoly.