Thirty years ago, the first Mac clones rolled off an assembly line in Austin, Texas.

If you’re not of a certain age, you might not even believe that there were once Mac clones. For most of its existence, Apple has been a singular company, selling products that were a fusion of custom hardware and custom software.

But for about three wild years in the 1990s, Apple defied its own nature and allowed other companies to build computers that ran the Mac OS and compete directly with Apple. It was an era that made some long-standing contributions to the history of the Mac, but also one that Steve Jobs dramatically ended pretty much the moment he returned to power at Apple.

But…why?

The mid-’90s were a weird time for Apple. Microsoft and Intel dominated the computer industry so much that Apple’s tiny market share just kept shrinking, and Apple struggled to make money. You might think that allowing other companies to compete with Apple would just make things worse, but desperate times called for desperate measures, and Apple CEO Michael Spindler (who had taken over for the deposed longtime CEO John Sculley) decided to take those measures.

The clone strategy was designed to allow third-party hardware makers to create systems to serve markets that Apple didn’t serve very well, allowing Mac OS to penetrate into areas where Windows was winning and turn the tide. But that didn’t really happen. Instead, Apple found itself boxed in on both sides.

At the high end, Georgia-based DayStar Digital built the first multiprocessor Mac clones targeted at the professional publishing market. DayStar built expensive computers that were faster than any of Apple’s systems, and sold them to some of Apple’s most profitable customers.

Meanwhile, Texas-based Power Computing took a page from Dell and started ramping up highly configurable beige clones, including both low-end and mid-range models. They weren’t pretty to look at, but neither were Apple’s Macs during this era. Power Computing’s killer feature was the fact that you could pick your device’s specs to order and then order it on the web (or via fax!) and it would be shipped to you. This sounds completely normal today, but Power was likely the very first company to sell computers via a web-based configure-to-order system.

Most clones were based on existing Mac motherboard designs, especially in the early days. Basically, Apple did all the core hardware and software engineering–and then let the clone makers take all of that and innovate their way into taking sales away from Apple.

When Steve Jobs took control of Apple in 1997, he put a stop to the clone licensing program immediately. (Macworld’s cover story was “Why Apple Pulled the Plug.”) “Licensees does not begin to cover their share of the expenses to engineer and market the Mac OS platform,” Jobs wrote in a memo to Apple employees. “Our Board is convinced that if Apple continues this practice the company will never return to profitability, no matter how well Apple performs, and the entire Macintosh ‘ecosystem’ will continue to decline, eventually killing both Apple and the clone manufacturers. This scenario has no winners–and customers end up with no Macintosh choice.”

Jobs was pretty ruthless in killing the clones, too. Apple’s license with clonemakers covered the 7.x versions of classic Mac OS, so Apple changed the name of its next update to Mac OS 8, so that clone makers didn’t have access to the latest version of the OS or any of the new chips that were only supported by Mac OS 8. Clone makers wanted to start building laptops, but the existing license didn’t cover that, either. And Apple announced it wouldn’t renew any of those licenses, which eliminated any hope of long-term business.

Jobs didn’t find the entire clone industry as a negative, though: Apple bought the core assets of Power Computing for $10 million in cash and $100 million in Apple stock. Part of that was probably an attempt to fend off lawsuits, but Jobs specifically cited Power’s “expertise in direct marketing and sales” and “pioneering of direct marketing and sales (a direction we want to move it)” in announcing the sale. That was recognition that Power, with its web-based direct sales strategy, had hit on something that Apple needed to replicate. Which it did!

While it’s a popular perception that the clone market was one of the causes of Apple’s near-bankruptcy in the late ’90s, Jobs wrote that it wasn’t true. Clones could really only amount for a tiny percentage of overall Mac OS computer sales–and meanwhile, “during the past two years the total number of Mac OS computers sold has declined by almost 20%.” The bottom was dropping out, clones hadn’t solved the problem, and Jobs saw them as a drain on resources that Apple simply couldn’t afford.

The legacy of the clones

There are a few surprising impacts of that brief clone era, even three decades later.

Obviously, Power’s approach of custom-building computers to order became the standard across the industry, and it certainly inspired Apple to change its processes to allow custom-built Macs to be directly ordered, rather than requiring users to pick from a set of stock configurations at a local or mail-order retailer.

The name Mac OS and the familiar two-smile symbol that now represents the macOS Finder both emerged from the clone era. Originally, there was no name for the Mac’s operating system. We generally just called it “the System,” which is why the groundbreaking 7.x update was called “System 7.” With the arrival of clones–which couldn’t be called Macs–there needed to be a way to turn the Mac operating system into a product, with a recognizable label and brand name, so that clone makers could explain that their devices weren’t Macs but ran Mac software. The result was the Mac OS name and the accompanying dual-smiley logo.

And there’s the entire approach the Mac takes to multi-threading software across multiple processor cores. It all emerged from the work Apple and clonemaker DayStar Digital did to create the DayStar Genesis MP and, eventually, Apple’s own Power Mac 9500/180 MP. In those days, the idea of putting additional processor chips inside a Mac to boost performance was novel, but if you were a high-end publishing professional, you would pay a lot of money to run Adobe Photoshop as fast as possible. So DayStar shipped a clone with two or even four PowerPC chips, and Apple and DayStar collaborated on the software that would allow a Mac OS device to send work to multiple processors.

That legacy lives on today in all of our multicore Macs. And just now, I searched my own M4 MacBook Pro for the words “DayStar Digital” and found several header files, deep down in a directory installed by Xcode, that bear the notice: “Copyright 1995-2011 DayStar Digital, Inc.” Thirty years later, that little legacy of a forgotten Mac clone lingers inside my Mac.

When Tim Cook posted on Monday that “there’s something in the Air” this week, we expected a new iPad Air (which happened), maybe an 11th-gen iPad (which also happened), and a new MacBook Air (coming Wednesday). But surprisingly, we’re hearing that Apple is tossing in a one-more-thing-like announcement into the week in the form of a new Mac Studio.

According to Bloomberg’s Mark Gurman, Apple is planning to update the high-end desktop for the first time in nearly two years. But he also includes a headscratcher: it won’t include an M4 Ultra. Rather, Apple may introduce the M3 Ultra, which most people had thought was as good as dead.

Apple has used M-series Max and Ultra chips in the Studio and Mac Pro but of the same chip generation. Gurman reports that Apple is going to use an M4 Max and an M3 Ultra, rather than M4 versions of both chips. It seems like an odd thing to do, as the M3 Ultra will be faster than an M4 Max, but the gains won’t be as dramatic as they should be.

Why wouldn’t Apple just go with an M4 Ultra? Perhaps there are manufacturing issues. Or perhaps Apple wants to create a distinction between the Mac Studio and the Mac Pro, which is rumored to get the M4 Ultra. We should find out more tomorrow, but for now, you can read all about the rumors of the M4 Mac Studio and the M4 Ultra Mac Pro while we try to figure this one out.