Apple's cash flow is every year more derived from being a services business. Even its revenue streams from hardware look more like an annuity than a transaction as customers replace devices. You can't value Apple like the widget factory in your textbook at school. P/E is useless. 
Hardware-based based moats are more precarious and requires continual innovation. Hardware that is “one and a half times better” is still an advantage. google.com/amp/s/25iq.com… And that hardware may be 🔑 to generating the competitive advantage that can be created by software.
"With near-zero variable costs, software offers the ultimate in scalability" It has low capital intensity and high gross margins causing better valuations. Why build hardware? Because creating distribution for software is damn hard to do successfully. abovethecrowd.com/2003/02/10/sof…
Can you create distribution for software without selling hardware? Absolutely. But doing either is vastly harder and more rare than people imagine. "a16z is a sucker for people who have sales and marketing figured out.” Marc Andreessen 25iq.com/2014/06/14/a-d…
If you look closely at the world's most valuable businesses today you will inevitably find very hard to replicate product distribution capability. That edge can be tens of thousands of enterprise sales people, proprietary hardware, innovative retail, freemium executed well, etc.
"Value of a business = value per customer x number of current and future customers. ...Business value is a function of customer economics and the current and future users (market size). This equation is simple, but unpacking the essential elements is not." Michael Mauboussin
"Wait! Are you saying I can't value a company simply by looking up a P/E ratio on Yahoo Finance? That an annuity business can't be valued like a widget factory? I need to do some work?" I'm saying a P/E ratio is about as useful in valuing an annuity business as a dill pickle.
