A lot has been written this week about Apple's switch from IBM to Intel chips.
Not a lot has been written about what IBM might do now.
The chip business is moving in two directions at once, toward mass production and mass customization.
A Microsoft order for XBox chips means mass production. Orders for FPGA chips onto which a process may be programmed represent mass customization.
An order for Mac chips falls somewhere between the two. Extensive development is needed for one customer, but is production really high enough to beat Moore's Second Law, the idea that costs rise with complexity, and grow exponentially?
It's possible that IBM concluded, not any more. Given Apple's proprietary model, the contract may not have been worth fighting for.
We usually think of mass customization in terms of gate arrays, like FPGAs. But Spooner writes that Power.Org, which makes those Power processors Apple is abandoning, is becoming a sizable community, with about 6,000 developers as members.
IBM wants to harness the power of this community, allowing members to work together in hardware, software and tuning, sharing the resulting breakthroughs. As with Linux IBM is looking to build service revenues here, in this case design and manufacturing service revenues.
I have said before that IBM is the leader in monetizing the open source model. But open source should not be confused with Linux. It's a business process, not an operating system, in which innovations are shared from a common pool, and costs are spread out among many different companies.
The short form. It's not communism, it's a co-op.
As hardware becomes software, and as Moore's Second Law bites, the open source business model is moving into the vacuum.