Dead End: The Mac Clones
Abstract
From 1995 to 1997, Apple licensed Mac OS to other computer makers. The theory was Microsoft’s playbook: let hardware partners grow the platform. The practice was that clone makers collected Apple’s best customers, paid Apple a roughly $50 royalty per machine, and took sales on which Apple would have earned hundreds of dollars. Steve Jobs ended the experiment within months of his return, spending $100 million to buy out the biggest licensee. It was the only time Apple tried being an operating-system company, and it nearly finished the job the rest of the 1990s had started.
Why Apple Licensed
By 1994 the argument for licensing looked unanswerable. Windows machines from dozens of vendors were crushing the Macintosh on price and availability, Windows 95 was coming, and Apple’s worldwide market share was sliding toward single digits. Inside and outside the company, the consensus was that Apple had made a historic error in the 1980s by refusing to license Mac OS the way Microsoft licensed DOS and Windows to everyone (Bill Gates and Microsoft). CEO Michael Spindler decided to correct it: in late 1994 Apple signed its first Mac OS licensing agreements, and the first licensed clones shipped in 1995.
Power Computing, founded by Steve Kahng, was the first and most aggressive licensee, selling built-to-order Mac clones directly by phone and mail in the Dell style and reaching roughly $100 million in revenue in its first year. Radius (founded by former Mac engineers), UMAX with its SuperMac line, DayStar Digital, Pioneer in Japan, and Motorola with its StarMax machines followed. In all, some thirty licensees and sub-licensees shipped about 75 distinct clone models.
Why It Failed
The licensing math never worked. Apple collected a flat fee plus a royalty that started around $50 per clone sold. An equivalent Macintosh sale carried several hundred dollars of gross margin. Licensing would have paid off only if the clones expanded the platform, winning buyers who would otherwise have bought Windows PCs. That is largely not what happened. Clone makers concentrated on the high end, where margins were fattest and where Apple’s own machines were weakest on price. Power Computing advertised faster PowerPC machines at lower prices than Apple’s and delivered them quicker. By 1996 Apple’s executives could see the clones cannibalizing their most profitable segment while overall Macintosh market share kept falling: Apple was paying to compete against itself.
Compounding the problem, Apple’s platform was shrinking while it invited competitors in. Microsoft’s licensing model worked because Windows volumes grew year after year, spreading the ecosystem’s costs over more machines. Apple licensed a premium niche platform with declining volume, so every clone sale came mostly out of Apple’s pocket rather than out of Microsoft’s.
Jobs Kills the Program
Steve Jobs returned to Apple in 1997 with no patience for the arrangement; he regarded the clone makers as leeches on Apple’s engineering. The contracts gave him his exit: the licenses covered System 7, and Apple’s lawyers concluded they did not extend to the new Mac OS 8. Apple declined to license the new system on terms any clone maker would accept. On September 2, 1997, Apple bought Power Computing’s Mac OS license and customer list for $100 million in Apple stock, and Power Computing shut down soon after. Motorola dropped the StarMax line, one casualty of a broader chill between the two companies. Only UMAX secured a Mac OS 8 license, and it expired in July 1998, ending the clone era after barely three years. (BeOS, itself chasing survival, had ported to Power Computing and UMAX hardware and lost its cheapest machines in the shutdown; see Dead End: BeOS.)
The decision looked ruthless and was. It was also correct: Apple’s recovery strategy under Jobs depended on integrated hardware-software products with healthy margins, beginning with the iMac a year later (Steve Jobs and Apple). A clone market would have undercut every one of them. Apple never licensed its desktop operating system again, and when the company later did open its platform to outside hardware ideas, it did so through the App Store, where it kept 30% and the hardware margin both.