Life is great when we have lots of choices. But this isn’t the case in the CPU world where we basically only have two choices on PC, AMD or Intel. So how did this happen?

Over 70 million PCs were shipped during the third quarter of 2020 alone, with as many computers as there are in the world, no other companies want a piece of the pie? 

To understand why there are only two CPU options for our PC, we have to go back to the first PC, the original IBM Personal Computer from 1981. IBM chose the Intel 8088 CPU to power the machine, which was based on the x86 instruction set. This ended up being an enormously consequential choice as the IBM PC exploded in popularity, and pushed lots of its competitors out of the market because it was a versatile, well-built computer that offered great value for money at the same time.

This meant that software developers wanted to write programs for the IBM PC and compatibles that utilized x86 CPUs. Meaning Intel quickly became a very powerful name in the microcomputer CPU space. So powerful in fact, that they ended up licensing out the x86 architecture to other companies to keep up with demand, without having to manufacture x86 chips completely on their own, but still make money. Ironically, AMD was one of these licensee companies and although Intel and AMD obviously remain rivals to this day, AMD still has an x86 license, which is used at various times, to beat Intel at its own game.

In the mid-80s the chip manufacturing industry convention was that when a computer vendor purchased CPUs, they would require a second source for the chip, a legacy business move from military contractors who needed a reliable supply. In effect for Intel or anyone to sell their chips, they would agree to enable a competitor to license their technology.

Obviously, AMD's Ryzen line-up is the one currently giving Intel great challenge but this was also true back in the 1990s when AMD started improving upon the x86 design and competing directly with team blue, rather than just being Intel’s second source chip supplier. Although AMD wasn’t the only x86 licensee that tried to make inroads into the market, they did have the knowledge and resources to become a serious contender, as they were already a publicly-traded company that had multiple chip fabrication facilities.

Other firms that had access to the x86 architecture, simply didn’t compete all that well. One notable example is Cyrix, who tried to go toe to toe with Intel’s new Pentium line-up in the mid-1990s. Cyrix promised big-time performance but their chips were rarely delivered. And they made an infamous mistake when they decided to focus on integer performance to compete with the Pentium. At the time, Cyrix thought that the trend of most desktop programs using mostly integer-based processing would continue. But what actually happened is that the low-cost but powerful Pentium became so popular, developers instead coded for its floating-point unit.

So, Cyrix’s challenge didn’t last very long. And other potential competitors were typically late to the game compared to what teams red and blue were offering. Think about how Apple switched from PowerPC to Intel partly because Intel chips were simple more powerful per watt. And of course, the next major innovation in desktop CPU was 64-bit processing, which was developed by none other than AMD, who subsequently cross-licensed that technology to Intel, paving the way for the modern ear of x86-64 computing employed by virtually all modern PCs and making it even harder for smaller chip makers to get a foothold in time to be relevant.

Now, of course, because most of these issues revolve around the x86 architecture, chipmakers who have focused on other instruction sets have done quite well. Qualcomm for example is a huge force in the mobile space with its ARM-based chips. And Apple has made headlines recently for releasing its M1 processor which offers very impressive performance for Mac users.

For the PC loyalists, they can't expect this duopoly to disappear anytime soon.