Today in Tedium: Why did Intel move away from naming its widely used microprocessors after numbers—you know, like the 286 and the 386? They were perfectly succinct, along with easy to remember—what’s not to like? As it turns out, Intel was very much into the idea of using these numbers, and really wanted to stick with it. But legal and trademark-related problems got in the way. The same spirit of reverse-engineering innovation that gave Intel its longtime advantage in the world of desktop computers ultimately cost it the right to hold onto its numbers. Tonight’s Tedium explains why. — Ernie @ Tedium
The year that Dale Earnhardt, Inc. received a trademark for the number 1, a trademark related to the racing team’s No. 1 car. It’s a trademark which the late NASCAR racer’s company still owns today. It’s a reminder that, yes, you can in fact trademark numbers. In fact, trademarks for numbers are actually fairly common. So, what was Intel’s problem?
The subtext of the trademark battle over Intel’s numbers: A long-lasting feud with AMD
The roots of many of Silicon Valley’s biggest conflicts did not start with IBM. Certainly, IBM’s entrance into the personal computer market firmed up some of the players, but the roots go a bit deeper.
For years, Fairchild Semiconductor, a firm that was closely associated with the early years of the integrated circuits, cast a long shadow on Silicon Valley. Among its founders was Robert Noyce, who created the first commercially practical integrated circuit made from silicon, and Gordon Moore, both of whom were co-founders of the company. Noyce, an executive who was passed over for a role as the company’s president, took Moore, the company’s head of research and development, and founded a new company called Intel in 1968.
Intel quickly made a name for itself through its microprocessor business, including with its 8086 processor, first released in 1978.
But Intel initially didn’t make all of its processors in-house, in part because there was a high demand for its product in the early computer years, and in part because microprocessors are complex things. IBM specifically asked for a secondary source for the 8088 when making its early PCs, and when that point came, Intel leaned pretty hard on other manufacturers, among them a company called Advanced Micro Devices, or AMD, which was also launched by alumni from Fairchild Semiconductor, led by Jerry Sanders.
This strategy worked well for Intel at first, but in the mid-‘80s, Intel decided that it wanted to be the only company to sell 80386 chips, and ended its “second source” licensing program.
In a passage from the 1996 book The Power of Boldness: Ten Master Builders of American Industry Tell Their Success Stories, Moore explained that the issue was that it felt the second-source deal with AMD was no longer a fair trade:
Most members of the executive staff remembered the poor experiences that Intel had with the i286 second sources. We wanted to obtain equivalent value in exchange for a i386 second source. Intel and Advanced Micro Devices (AMD) had agreed each company would earn points that could be used to evaluate technology exchanges. But by 1987, there were no acceptable AMD products for exchange, and nothing was expected for at least two years; AMD had no way to earn second source rights to the 386. We decided that we would not transfer the 386 to AMD unless we believed that we were actually going to get valuable products in return. Thus, there was no second source arrangement with any 386 competitor other than one that allowed IBM to make a portion of its internal consumption.
This move, which Moore admitted was “a real change in the way the industry operated,” was controversial at the time. And it put Intel at odds with a 1982 agreement that it made with AMD to allow the tech company to second-source its CPUs.
Tech analysts, like John C. Dvorak, did not like what they saw and called the move short-sighted.
“If the 80386 remains a single-source system, then you can be sure Intel will single-source all its future microprocessors, and the paranoid designers will look elsewhere fast,” Dvorak wrote in a 1987 column for PC Magazine.
And because Intel was making its own processors entirely under its own corporate umbrella, suddenly its trademarks were worth much more to the company than they were previously.
There was just one problem with that.
“In a three-year period, AMD made over a billion dollars in operating profit. And I was very proud of that billion dollars. But always over my head was the sword of Damocles, that if we lost the court case, AMD was history. So there was tenseness, a tautness in my stomach through all those good years. Even though we were making a billion dollars, I couldn’t enjoy the good times.”
— Jerry Sanders, the founder of AMD, discussing his company’s legal battles with Intel in the late ’80s and early ‘90s in an interview with Forbes. Sanders noted that Intel’s moves during this time represented “duplicity,” an attempt to push AMD out of the processor picture. There were image problems on AMD’s end as well: A 1989 New York Times article portrayed the company as moving slow on technology during the mid-’80s and having an overly lavish image. But ultimately, these battles would work out in AMD's favor.
An array of Intel and AMD processors from the '90s. (Labete/Flickr)
Why Intel legally lost the ability to trademark its numbered processors
AMD, clearly, was smarting from Intel’s move to push it out of the processor picture, but it wasn’t out. The company, outside of the courtroom, did what it could to keep up.
It eventually reverse-engineered Intel’s work on the 80386 processor—an effort that took years—but before it was able to release its own version, fate intervened in a pretty hilarious way.
In 1990, a package intended for AMD marketing official Mike Webb ended up getting delivered to an Intel engineer … who was also named Mike Webb. And by sheer chance, the two men were checked into the same Sunnyvale, California hotel at the same time. (AMD's Webb was based in Austin, Texas, but was visiting the Sunnyvale-based chipmaker.) The hotel’s staff tried to deliver the package to Mike Webb, but unfortunately, they picked the wrong one.
(Who said irony was dead?)
This weird circumstance led Intel to sue AMD for trademark infringement in October 1990, for somewhat understandable reasons. And Intel was worried.
The company had had invested a lot in the 386 name. According to Ad Age, the chip-maker was so freaked out that its processor name would become generic, it forced PC manufacturers that mentioned its processors to mention its company name. This strategy eventually led to the “Intel Inside” campaign, which astutely focused on the company more than the architecture.
Writing in PC Magazine, John C. Dvorak noted that all this gave Intel the look of a company that knew what was coming.
“The possibility that the 386 may be legally cloned is becoming a reality,” he noted in November of 1990. “So much so that Intel is already taking a defensive posture by putting together advertising campaigns that emphasize that Intel invented the chip.”
And, unlike some other calls that Dvorak made over the years, that prediction quickly became reality. In March 1991, the judge sided with AMD, invalidating Intel’s trademark on the 386 by claiming it was generic.
That was a big loss for Intel, and it was one of many legal battles between the two companies over the years. The lawsuits highlighted the complexity of the companies’ business relationship. AMD claimed that, due to the contract between it and Intel, it had the legal right to Intel’s microcode for multiple generations of x86 chips, and the debate over this, along with Intel’s evolution on the processor front, drove an array of lawsuits between the two companies. (It took until 1995 for the two sides to agree to stop suing one another for a while—a battle that picked up again after AMD sued on antitrust grounds in 2005.)
But what did this all mean for Intel’s numbering scheme? It meant that it was in the market for a new name. The company attempted to register both “486“ and “586” with the U.S. Patent and Trademark Office, but ultimately abandoned the applications. The company also attempted to register the word mark i586, but abandoned the mark after it only received approval with the disclaimer “no claim is made to the exclusive right to use '586' apart from the mark as shown.”
This led Intel into the arms of a company called Lexicon, a firm responsible for coming up with memorable, unique names. The firm was given a few basic requirements—notably, to treat the processor like a key ingredient of the computer.
Lexicon CEO David Placek noted to The New Yorker that Pentium was a winner in large part because it was both a unique mark, and one that spoke to the number “five”—as “pente" means “five” in Greek.
“The first thing I thought of was the Pentagon, and I thought, Huh, that’s pretty interesting, because it’s a shape,” Placek recalled, adding: “I thought, Wait a minute—we’re going from 486 to the fifth generation, the 586.”
The mark spoke to the same things that the numbers did, but did so in a way that Intel could trademark—and that consumers would specifically ask for in stores.
“It was one of our great success stories,” longtime Intel leader Andy Grove told the magazine. “According to our own internal research, in the late nineties Pentium was a more recognized brand name than Intel—which is actually a little scary.”
Intel couldn’t legally stop AMD or other competitors from using 586 as a name—but Pentium? That was a bit easier to protect.
Since this saga went down, Intel has famously become notorious for being aggressive about protecting its trademarks—no matter how big or small the stakes.
In 2002, for example, it tried to sue a San Francisco yoga studio because the name of the studio was Yoga Inside, and Intel owned the Intel Inside trademark. In 2008, the company went after a small travel agency called Intellife Travel, out of concern the brand could dilute Intel’s trademark.
Last year, the company went after a site called InvestorIntel, out of concern people might confuse a news website focused on investing in experimental technologies with a company that sells microprocessors.
And perhaps most unusual of all, John McAfee had to sue Intel last year in a trademark battle over the use of his own name, because Intel owns McAfee and John McAfee no longer does. Intel had even stopped using McAfee for a time, only to revive it as an independent firm.
One has to wonder if Intel, as a corporation, is acting out because of the trademark that got away way back when. They lost something important back then, and they probably don’t want it to happen again.
Correction: An earlier version of this piece incorrectly stated that IBM first used the Intel 8086 for its initial IBM PC. Instead, they used the 8088, a less-expensive variation of the 8086. We regret the error.