Worked Example: Owning the Roadmap — The Product Playbook

The free-throw example is deliberately small. Here's a heavier one, a real build-versus-buy bet with a company's future riding on it.

This is written as the strategy Apple committed to in 2008, when it decided to design its own chips instead of buying them. Every phase is grounded in what actually happened. Read it in the template's order, and watch how a giant, expensive, multi-year bet still comes down to one crux.

1. Diagnosis

Apple's entire advantage rests on controlling how hardware and software fit together. Yet the most important component in every product, the processor, is designed and scheduled by outside suppliers whose priorities aren't ours.

In mobile, we depend on third-party application processors built for the broad market, not tuned to our products. We can't get a chip designed around a specific iPhone's display, battery target, or feature set, because the supplier serves everyone at once. In the Mac, we depend on Intel's roadmap. Their gains in performance per watt, the metric that matters most as our lineup shifts toward laptops, are slowing, and their delays and quality problems are starting to gate our own product timelines.

The root issue isn't chip quality in any single year. It's control. As long as another company owns the processor roadmap, our product cadence, our differentiation, and our power and performance envelope are all capped by someone else's choices and someone else's calendar.

2. The Crux

The pivotal, addressable obstacle is our lack of control over the processor roadmap. We need to design the silicon ourselves, co-developed with our software and shipped on our schedule.

Important. The processor sets the ceiling on performance, battery life, differentiation, and timing. Win here and the gains cascade across every product line.
Addressable. Chip-design talent and ARM-based architectures can be acquired and built into a capability under our direct control.
Tractable. We don't have to conquer the hardest target first. We can prove the capability on a contained, high-volume product and expand from there.

Our crux. We don't control the processor roadmap. Solve that and the cadence, the differentiation, and the performance envelope all stop being someone else's call.

3. Guiding Policy

Build a world-class in-house silicon organization, and prove it on mobile before we ever touch the Mac. We'll design custom systems-on-chip tuned to our own products, manufacture through foundry partners rather than build our own plants, and treat silicon as a living annual roadmap, not a part we buy. We get there fast by acquiring proven chip-design teams instead of hiring one engineer at a time.

What we will deliberately not do right now.

We will not build our own fabrication plants. Manufacturing is enormously capital-intensive and it isn't our crux. We partner with foundries and keep our force on design.
We will not start with the Mac. It's the hardest, highest-risk target, with the deepest software-compatibility burden. We earn the right to attempt it by succeeding in mobile first.
We will not sell our chips to anyone else. Staying internal lets us optimize for the product instead of the merchant market, and keeps the team lean and focused.

4. Coherent Actions

Acquire the talent. Buy proven low-power design teams (P.A. Semi, then Intrinsity) to seed the organization rather than building from zero.
Install one accountable owner. Put a single executive (Johny Srouji) in charge of all silicon, reporting to the top, so the roadmap has one owner.
Ship custom silicon in mobile first. Design an SoC tuned to a specific flagship, the A4 for the iPad and iPhone 4, where volume is high and the software stack is entirely ours.
Establish an annual cadence. Commit to a new generation every year (A5, A6, then the 64-bit A7) so competitors are always reacting to last year's move.
Remove the manufacturing vulnerability. Shift foundry work toward a partner that doesn't compete with us in devices (TSMC), so a rival can't gate our supply.
Only then, transition the Mac. Once mobile silicon outpaces the industry, design Mac-class chips (the M-series), bridge software with a translation layer, and replace the Intel lineup on a defined timeline.

The pivot action. Ship a custom SoC in a high-volume mobile product first, the A4. Every later move, the annual cadence, the 64-bit leap, and ultimately the Mac transition, depends on first proving we can design a competitive chip on our own terms.

5. Proximate Objective

Within roughly two years, design and ship our first fully custom, Apple-designed system-on-chip in a flagship mobile product, with performance and battery life tuned to that product in a way no off-the-shelf processor can match.

The Mac is deliberately not in this objective. It's the eventual payoff, not the starting line.

How it ran

Not the strategy itself, just how it got run.

Metrics that matter. Roadmap control, the share of our product timelines gated by an outside supplier's schedule, driven toward zero. Performance per watt versus the best third-party alternative, year over year. Real-world battery life at a given performance level. Whether a new generation ships on schedule every year.

Timeline. Phase 1 (2008 to 2010), acquire design teams, stand up the organization, design the first custom SoC. Phase 2 (2010 onward), ship the A4 in the iPad and iPhone 4, lock in the annual cadence. Phase 3 (2013), leap ahead with the first 64-bit smartphone chip, the A7, forcing the industry to react. Phase 4 (mid 2010s), mature the foundry partnership and scale the architecture across the line. Phase 5 (2020 to 2023), execute the Mac transition, starting with the M1 in late 2020 and finishing when the last Intel Mac is retired in 2023.

Communication and feedback. Keep the roadmap tightly held internally, and reveal capability through shipped products, not announcements. For the Mac transition, give developers the tools and a translation layer early so the software ecosystem moves with the hardware.

What to notice

This is a fifteen-year, company-defining bet, and it still has exactly one crux. Control of the roadmap. Everything else is in service of that, and the things Apple chose not to do, no fabs, not Mac-first, no selling the chips, are as deliberate as the things it did.

Notice the proximate objective too. With the entire Mac line eventually on the table, the first target was one mobile chip in two years. The huge outcome, an entire computer lineup off Intel, came last and arrived more than a decade later. They didn't lead with it. They earned their way to it.