Origins Of the AMD-Intel Rivalry

How did the Intel – AMD rivalry start? In his book SPINOFF, Charlie Sporck, former CEO of National Semiconductor, interviews Jerry Sanders, founding CEO of AMD.

"The Japanese were taking over the semiconductor industry. That was the fear. Everybody was scared," recounts Sanders, "we had HP saying Japanese quality was superior. Things were tough."

"Andy Grove believed, in the 81-82 timeframe, that there was no way the US semiconductor industry, and that included Intel, was going to survive the onslaught of this enormous juggernaut. They had to find ways to make alliances to leverage resources."

"Grove could have just made us their second source," recalls Sanders, "we could have paid a royalty. But he wanted someone to augment Intel's design capability and he thought we could be that company."

Wistfully, Sanders adds: "It turns out we didn't do a very good job there."



  1. My first CPU was 8088 for one week after that NEC V20
    (150% Norton SI) with Taiwan board and Hercules monocrom grafik BIOS I dont remember but it run MS Flightsimulator 1.0 that was the must have checkpoint these times. kny

  2. Thanks for that, Fred – I knew a bit of the DMA controller story but not that it was AMD’s IP (was that part of the x86 deal?) or that it saved Intel’s bacon against NEC (who did better chips in general, I seem to remember). That’s a wrinkle I’m delighted to learn.

  3. I believe the 8257 DMA controller chip used in PC’s was AMD’s I.P. even though the PC part usually had an Intel logo on the package an most folks thought it to be an Intel design.
    For those of you who remember the NEC V40 chip, which had **almost** the correct peripherals for a more integrated PC ( improved performance CPU, 8259 Interrupt controller, 8254 timer ) but, curiously a DMA controller not compatible with the 8257.
    That would explain why – NEC believed had a license to build on ( and improve )Intel’s processor and peripheral chip technology **but** not AMD’s !
    This was fortuitous for Intel, as at that time NEC had the faster compatible CPU core, interrupt controller and timer.
    But, the DMA subsystem was PC incompatible and if it wasn’t for this then the NEC V40 would have taken the entire PC clone market away from Intel, thus changing history dramatically.
    Incidentally, Intel’s data sheet for the 8257 was not quite complete, for reasons one can only speculate. There was enough documentation to allow the chip to be used in standard mode but there were undocumented registers as well which were used by some popular PC games of the day ( Quake and possibly Doom II for example ) to improve the game performance.
    Later on, Harris produced a CMOS version of the 8257 which did properly document the complete register set of an 8257.
    So, maybe Intel do have something to grudgingly thank AMD for, albeit totally fortuitously.


    rupertg, I, too, heard that IBM would not commit without a second source. Actually I heard it from several sources, but none that I could cite as authoritative.

  5. I understood the second sourcing deal to have been at the behest of IBM and the PC deal, which had a policy of not committing to single-sourced components (remember those days?).
    On reflection, I don’t now how true that would be: did AMD (or anyone else) second-source the peripheral chips? The original PO was basically built from Intel data sheet example circuts, ISTR, much as the Atari ST was from Moto’s.


    For the sake of accuracy, I would like to point out that it was NOT simply a matter of “HP ‘SAYING’ Japanese quality was superior”(emphasis added). HP knew that their announcement would cause the fecal matter to strike the rotating blade so they had a whole boatload of data to back up the results of their analysis. At that time, Japanese quality WAS superior.

Leave a Reply

Your email address will not be published. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>