I’ve talked before about the disadvantages of the Intel 8086. But in spite of its weaknesses, it won. The CPU you are using right now descended from it. Even if you’re reading this on a smartphone, the server that sent the page to your smartphone has an Intel 8086 descendant in it. So what are some advantages of Intel processors? They won for a reason.
In the 1980s, there was a saying that nobody got fired for buying IBM. Intel positioned itself as the logical successor to that, building an ecosystem around its CPUs that allowed it to conduct marketing campaigns as the safe, reliable, compatible choice. It was true enough of the time that Intel built momentum that no other company could match.
Intel won because it priced its CPUs cheaply, but not too cheaply. Plenty of companies came and went who made better CPUs. Digital Equipment Corporation is a fine example. But Intel CPUs were cheaper than CPUs from companies like DEC, HP, Sun, and others. That gave Intel a larger market.
In the 1990s, about a half dozen companies tried their hand at undercutting Intel’s prices, selling clone CPUs. Intel held them off in two ways. First, the Intel chips usually were faster at math operations necessary for 3D gaming. Between this advantage and the competitors’ much lower margins, Intel slowly drove all of them from the market except AMD.
Pricing CPUs too cheaply cuts into the capital necessary for future investment. Commodore once had a vertical integration that could make Intel jealous. But Commodore rarely invested money into making its manufacturing more efficient.
Aggressive, predictable price cuts
Intel generally cuts prices every quarter, and has been doing so for decades. Computer manufacturers love this, as it lets them design a model to hit a price point months in advance. Intel can do this because they own a large number of chip plants, nicknamed “fabs” in the industry. They always have some number of fabs idle, undergoing modernization to get ready to produce the next generation of smaller, faster, and cheaper chips. When a fab is ready for the new manufacturing process, Intel shuts down and modernizes another one. Intel rotates the fabs in and out of production so it can always make those price cuts.
Intel hardly needs IBM now, but in 1981, winning IBM’s backing was critical. IBM used the Intel 8088 CPU in its IBM PC. Microsoft really wanted IBM to use the Motorola 68000 CPU and run Xenix, Microsoft’s Unix, on it. I’m not making that up. But the Intel 8088 was a lot cheaper than the Motorola 68000. The IBM PC was anything but cheap, but the Intel 8088 let IBM get the profit margin it wanted.
You know the rest of this story. The IBM PC grew popular, thanks to IBM’s big name and a killer app called Lotus 1-2-3. Other companies cloned the IBM PC and turned it into an industry. Hundreds of imitators were more than happy to accept thinner profit margins than IBM. Intel went from sharing the desktop CPU market with Commodore, Zilog, and Motorola to having all of it. From there it moved upward into high-end computing too, where it had been non-existent.
If IBM had listened to Bill Gates and built a 68000-based Xenix machine and called that the IBM PC, Intel would probably be in a different position today.
Intel’s formula wouldn’t work if the chips weren’t any good. Intel had a good-enough chip at a reasonable enough price to attract IBM. And by the time any Intel chip neared obsolescence, Intel had a new, good-enough chip ready to replace it. That allowed Intel to hang in there until its chips caught up with other competitors one by one. It took a quarter century to catch them all. But markets favor the good-enough, cheap-enough solution. Intel did a better job of positioning itself in that sweet spot over the long haul than anyone else.
From time to time, AMD makes a better chip. AMD did it in 2005 with its Athlon 64 line, and then did it again in 2017 with its Ryzen line. The competition keeps Intel honest, but AMD’s advantage, when it gets one, typically only lasts one chip generation. Intel has built up such an advantage since the 1970s that it will have to do a lot of things wrong for a very long time to squander it.
Intel countered the clone threat by positioning itself as the quality, compatible choice. The problem was that when you put an Intel chip on a low-quality motherboard, you still had a junk system. Intel solved that problem by going into motherboard production, ensuring you could get a reasonable quality system by buying both an Intel CPU and board. It worked. Low-cost vendors putting the cheapest stuff they could find together tarnished the reputation of the clone CPUs, while consumers realized they could get a good balance of price and quality by asking for Intel everything.
Once Intel established the market and a few other brands producing consistently reliable boards emerged, Intel stopped producing motherboards and went back to being a partner rather than a partner/competitor.
Knowing when to enter or leave a market
Last but not least, Intel has a very good track record of knowing when to cede a market to others. Intel ceded the memory chip market to the Japanese in the 1980s when it realized Japanese makers could make better memory chips cheaper than it could. Intel entered the chipset market in the 1990s, giving itself more control over the whole system, in addition to another revenue stream. Intel’s entry into the motherboard and chipset market helped stabilize the PC ecosystem in the 486 era, when compatibility was something of a Wild West.
Intel navigated this pretty well, doing as much as anyone to stabilize the PC market. Producing reliable chipsets and motherboards helped burnish the reputation of its CPUs. The mantra in the 1980s was that nobody ever got fired for buying IBM. Intel eventually became the logical successor to that. There were times when other companies could provide a better solution than Intel. But tracking who was better than Intel this year was a lot of work. And in the end, that’s why Intel survived, because Intel built an ecosystem around its CPUs that supported its marketing as the safe, reliable, compatible choice.