IBM PowerPC Motherboards: Apple's Biggest Opportunity Yet
August 20th, 1999

IBM announced last week that they were releasing a CHRP motherboard reference design for free. This would allow OEM manufacturers to produce CHRP based computers and ship them with LinuxPPC. MacWeek has reported that it may even be possible for enterprising users to make Darwin, Apple's open source code for MacOS X that should be released next year, work on these as well.

There is much debate as to whether this will impact Apple's business. Is it possible that these potential CHRP machines will steal sales away from Apple?

Uhhh... No. I think that this will actually bring more customers to Apple and bring other benefits to both Apple and Mac users alike.

First, we need to go back to the Clone Wars. For those not familiar with the Clone Wars, this is the name given to the last few months of MacOS Licensing from Apple. Apple had pursued a licensing program that reached fruition under then CEO Gil Amelio with companies such as PowerComputing, Umax, MacTell, and Motorola producing Macintosh clones. These clones used motherboard designs and Mac ROMs from Apple. Due to flaws in the licensing agreements and the years of mismanagement at Apple, the clones succeeded in mainly cannibalizing sales from Apple, while at the same time Apple's sales plummeted all on their own. This occurred for the following reasons:

  • Apple had quality issues (Think 5300 and every Performa made)
  • All the problems that had been developing within Apple for the preceding decade or so came to a head during the cloning era
  • Cloners, such as Motorola, Umax, and PowerComputing in particular, were able to make faster, cheaper machines, and bring them to market more quickly (sometimes by pre-announcing machines way before they were anywhere near ready to ship)
  • The licensing fees paid by the cloners to Apple were inadequate to cover the costs of R&D, leaving Apple to foot most of the R&D bill while the cloners reaped much of the benefit.*
  • The cloners concentrated their marketing efforts in established Mac markets, instead of trying and bring in new Mac customers. It had been assumed by much of Apple's management that the cloners would, in fact, bring in new customers.
  • Microsoft's successful marketing of Windows 95 convinced many computer users that the Mac's vaunted ease-of-use advantage had vanished, thus cutting into Apple's sales

The licensing effort was cancelled in part for the reasons I mentioned above, but also because Steve Jobs can not tolerate the idea of people using his company's software on hardware designed outside of his company. Steve's vision is strong and I think that the idea of people sullying that vision with their own ideas is repugnant to him.

I am not trying to turn this column into a Clone War rant, I just wanted to give some background on that time.

To the outside world, it was a fairly accepted concept that cloning was going to save Apple. This was important to Apple's AIM partners, IBM for "I" and Motorola for "M," who wanted to see the Mac market expand so they could sell more PowerPC processors. Indeed, the concept of licensing had a big part to do with the development of the AIM objectives in the first place.

So when Apple stopped licensing, not only did it royally piss off its two main master licensees, Motorola and IBM, because their not inconsequential investments in their own licensing efforts were rendered worthless**, it also threatened their investment in the PowerPC itself.

IBM had said that they needed to sell at least 1 million PowerPC chips each quarter in order to break even. Presumably something in the same neighborhood would be necessary for Motorola to break even as well. Folks, that's somewhere around 2 million Macs per quarter that need to be shipped to make the I and the M in AIM happy.

Right now, Apple is selling Macs and iMacs in record numbers and yet they hover below 1 million units per quarter. That's quite a discrepancy between what IBM and Motorola want to sell and what Apple needs to buy.

Still, it is obvious that whether or not Apple did the best thing by canceling cloning in the manner that they did, Apple is definitely healthier than they were two years ago. What does this have to do with free PowerPC motherboard designs?

I'm getting there.

So here we have the two most important members of the AIM consortium angry at the third, that's Apple, and just about ready to give up on them. Motorola decided to put much of their efforts into embedded PowerPC projects, a decided turn away from the PowerPC desktop model.

IBM also grew fairly interested in embedded processors as well and eventually sold its share of the Somerset Design Center, the birthplace and main research facility for the PowerPC, to Motorola. IBM also diverged from Motorola's G4/Altivec path, choosing instead to focus on continued development of the G3.

Rumors surfaced from both companies that their commitment to providing Apple with PowerPC chips may be waning. This was all during the dark days of the Clone Wars and the unsure times that followed when little word was coming from Apple and her new iCEO, Steve Jobs.

The CHRP project is also a factor in this equation. CHRP stands for Common-Hardware-Reference-Platform. CHRP was going to be the new Mecca to which all computer users would come. All manner of operating systems were going to run on CHRP, including our very own MacOS, Windows NT, OS/2, AIX, Netware, and Solaris. So the idea was that you could grab a CHRP box and run any of those operating systems! How nifty!

Needless to say, it did NOT happen. Microsoft and Sun dropped out rather quickly, though I believe one can still run Windows NT 3.5.1 and an early version of Windows NT 4 on CHRP. MacOS 7.6 runs on CHRP, though a specific version is required. I actually have a CHRP prototype StarMax 6000, but that unit includes a Mac ROM and is not entirely CHRP. It does have a PC parallel and Serial port though.

All of this longwinded background information is meant to explain that the two manufacturers of the PowerPC processor were very unsatisfied with Apple as a business partner. Having them be happy is a good thing. Having them be happier about producing PowerPC processors is even better.

Enter the motherboard designs from IBM. For the first time since the Glory Days of the Cloning Era, IBM, and possibly Motorola, now has the opportunity to see the market for our favorite processor grow faster and larger than Apple's market share. This means that IBM, and possibly Motorola, might become happy about producing the damn things.

It also means that prices, already lower than those charged by CISC-hampered Intel, could possibly go even lower, and cause IBM, and possibly Motorola, to invest more resources into producing faster versions of the PowerPC more quickly.

Keep in mind that the fastest shipping G3 does not come from Apple. Instead, the fastest G3s are being shipped by the likes of Powerlogix, Mactell, XLR8, and Sonnet in the form of 466 MHz and 500 MHz upgrade cards. This is partially because Apple has always been slow about bringing faster products to market, and partially because IBM and Motorola are somewhat slow to produce their newest shipping chips in large enough quantities for Apple to use in their product lines.

All the while, the PC world is enjoying the perceived advantage of 600 MHz Pentium III chips. This perception problem does little to enhance Apple's image in the MHz oriented public, though Apple has proven it can sell machines with little more than half that MHz rating in great volume.

Where does this leave us today? LinuxPPC will run on CHRP. Linux is popular. The PowerPC is very powerful. Apple does not make the kind of hardware that the majority of Linuxheads are into. Linuxheads want to tinker and they want lots of slots so they can do geeky things. They also want lots of drive bays because it's cool. In the server world, these things are often very necessary.

Please do not misunderstand me. I would love to be running LinuxPPC on a Mac to mess around with and handle some of my home networking needs. It's not necessary, but it sure would be fun. I know my way around Linux, and am very attracted to running it on a Mac. Lots of other Linuxheads just want to run the most bad-ass hardware they can get their hands on. At the same time, they are NOT going to shell out the kind of dollars necessary to buy a Blue & White. There are also large parts of the corporate world that will never buy a computer from Apple. They might by a Linux box from LinuxBox, Inc. though. That means their choice is not between LinuxBox, Inc. and Apple, it is between LinuxBox, Inc. and Compaq.

That Compaq box will have a nifty Intel processor in it. I would rather see their hard-earned corporate dollars go to PowerPC development than for Merced development.

The seething world of the Linux faithful are also not likely to buy a Mac. They may buy a comparatively inexpensive big beige machine from LinuxBox, Inc. Hopefully this effort from IBM will even lead to Motherboards and processors available for individual purchase so that Linuxheads can build machines themselves. That's a prime motivator for many Linux users. Indeed, that is a prime motivator for many PC users in general. It is also not a market that Apple competes in or is likely to compete in soon. Once again, IBM's plan could lead to companies moving the PowerPC into markets where Apple does not tread.

This will not take sales away from Apple, but it will offer some great exposure for the PowerPC processor to lots of places that would probably not even consider running a Mac or MacOS X Server. That exposure will help Apple eat away at the Wintel hegemony and even erode the ridiculous value that "Intel Inside" holds with the computing buying public. That will lead to more Mac sales, at least in an indirect way.

The cheaper prices and increased resources that may result if this whole cockamamie plan actually catches on will also directly benefit Apple and her customers.

I think it could even increase the credibility of Apple and the Mac, especially MacOS X Server and Client, in some portions of the business world. Powerful Linux boxes running on the PowerPC chip will be powerful advertising for the company that makes a great GUI that runs on that same PowerPC chip.

In effect, this effort from IBM could do what cloning was supposed to do, and that is to bring Apple, and AIM, more customers, customers that Apple can not reach on their own.

* This argument I do take some argument with. Several of the cloners put a lot of effort into tweaking out motherboard designs to get the highest performance possible. There were R&D costs associated with their operations, though Apple had the lion's share (the vast majority in fact).

** Motorola alone wrote off some US$90 million dollars in the process of canceling its cloning effort, a loss they are still bitter about according to sources within Motorola.

Your comments and hate mail can be sent to [email protected].