Apple moved to its own silicon for better control of user experience. Its core supplier is no longer the partner it once was. And because it can innovate more quickly. That happened in 1994. Apple moved to the PowerPC and its RISC architecture away from the struggling Motorola 68000 family of chips. The Apple M1 chip is not a strategic break with a legacy. It represents a return to an earlier strategy that differentiated the Mac platform until its retirement in August of 2006: control over the entire stack, including costs and schedule.
I am not going details of how 5-nanometer M1 differs from the PowerPC. That is a history lesson best left to new chip designers. The Apple M1 is a RISC chip like PowerPC–extending the 1994 strategic choice back into the mainstream Mac line. Unlike the 1994 decision, however, Apple’s resources differ significantly today…they no longer need IBM, Motorola, or the AIM alliance to create a viable silicon alternative. Apple also no longer needs ARM, its more recent chip design partner on its way to NVIDIA via a $40B acquisition. With M1, Apple creates its own market.
M1 and the value of periodic disruptions
Periodic disruptions drive innovation, and that is what matters to mainstream developers and Apple’s customer base. Apple aptly named the new chip the M1 rather than some extension of the A or Z family that power iPhone and iPads. M1 is a first-generation chip, but because of Apple’s extensive silicon design experience, and their intimate knowledge of Intel’s CISC architecture limitations, it can build chips to meet known deficiencies in competing architectures. Thus, the new MacBook Air becomes fanless despite its thin design—yet still manages to run faster than its contemporaries with fans and Intel chips. Developers get to see their creations run more efficiently on a wider range of computers, and customers get to experience increased performance along with more meaningful access to previously constrained software choices (like games that didn’t perform well on basic MacBook Airs).
The M1 brings System on Chip (SoC) to a new level, with CPU, GPU, Machine learning, and various memory features into a single chip. From a user experience perspective, the separation of cores into high-compute and casual-compute components offers a view into the power of owning the full stack. The new silicon better balances the user experience, and purposefully engages hardware through reciprocal design. Software runs better.
Back in the era of the PowerPC, the integration of the FPU, or Floating-Point Unit, proved of value to making spreadsheets and Computer-Aided Design sing. In the Apple M1, the Neural Engine will likely prove the definitive enabler early in the cycle, as developers learn to leverage on-chip ML for new features within existing software like AR and voice recognition, but more importantly, for new classes of productivity, gaming, and wellness.
The Apple M1 and the cycle beyond
Apple says the movement to pure Apple Silicon is a multi-year journey. Strategically, they will want to accelerate that journey to reduce the dual architectures required during the transition. When MacOS Big Sur starts controlling Macs on Thursday, November 12, 2020, it will run almost exclusively on Intel-based machines. The burn rate associated with dual architectures may not overly concern Apple. Its deep pockets will easily fund the runway—but it will remain a technical and tactical distraction until completed.
At the end state of the journey, Mac users should ultimately expect a deflection point. At some point in the future, Intel Macs will no longer run the latest operating system. Some Mac owners will experience a discontinuity now as older systems won’t install macOS Big Sur. Eventually, Apple will retire the Intel line and support for its Macs based on that architecture. The slower the transition, the higher the likelihood that the next disruption in chips will challenge Apple’s innovations. That will be Apple’s next competitive inflection point.
The cycle of chip displacement will never end. The future will likely find users working through an operating system running on a chip as foreign to MacOS and the M1 as those components are to TRS-DOS and the Z80.
For more serious insights on hardware click here.
Leave a Reply