Last November, we wondered what had happened to Apple’s’ gross margins (part of a series and here). We were reminded of this listening to Ben Thompson’s and Jon Gruber’s excellent Dithering podcast last week reviewing Apple’s Q2 earnings report. Thompson keyed in on the uptick in Apple’s gross margins, and they attributed this to the shift to the Apple Silicon M1 CPU. We wanted to test out that thesis. Thompson’s post on the estimable Stratchery shows how unusual the bump in gross margins is (paywall, but yo should really subscribe). The second quarter is usually a period when margins drop from the preceding holiday quarter.
In our earlier posts, we noted that Apple’s gross margins had remained stable despite a sizable increase in high-profit services businesses. Apple’s hardware margins had declined fairly sharply over just a few years. We attributed this to two factors: a mix shift to lower-gross margin products (i.e. watches, AirPods and lower-priced iPhones) and an increase in component costs, which is largely attributable to their adoption of OLED screens in the iPhone..
But then in their latest results, Apple’s gross margins jumped to 43% from 40% in the prior quarter. That is a big move, especially for a company Apple’s size, $90 billion in revenues. The Dithering crew hypothesized that the shift to the M1 was a big enough jump to move Apple’s margins up by $3 billion. This is probably not the case.
Apple no longer breaks out their unit volumes, but they do break out hardware and services. Hardware margins did increase slightly, from 35% to 36%, but services margins increased by more, 68% to 70%. In the recent quarter services revenue went from 24% of gross profits to 31% of gross profits, while hardware decreased from 76% to 69%. So the big increase in margins came from the contained shift towards services revenues.
Let’s look at it a different way. How much do their gross margins improve with their own silicon. Apple probably paid Intel $100 per CPU, and their average MacBook price is probably around $1500. If their hardware gross margins are 36%, that works out to a cost of goods of $960 per laptop. (We are eliding over a ton of simplifications and assumptions here). Apple probably pays TSMC about $25 per chip (see that last post for the math). So they are increasing gross margins by 8 points ($75 in savings/$960 BOM). Apple probably sold about 5 million MacBooks last quarter, call it 4 million with the M1 (this would be a lot more fun if they still disclosed unit shipments), which works out to $300 million in gross margin improvements, which means the shift to M1 contributed 0.3% (not 3%, 0.3%) of gross margin improvement in the quarter.
In the third installment of our series we worked out that Apple’s silicon project was not really motivated by better prices for silicon. And these results demonstrate that. The upfront cost of internal design are significant enough to mute the amount of margin recapture they generate. We are sure the purchasing team is delighted with the results, (cheap champagne for everyone), but this is almost rounding error for Apple. If Apple really wanted to drive cost out of their laptops with silicon they would figure out a way to design their own memory chips.
The real benefit for Apple in designing their own silicon is the strategic benefit – M1 helps them sell more MacBooks, the A Series makes their phones faster, more responsive and less power-consumptive. For Apple, everything starts with the customer, and building their own silicon really brings about a better customer experience. Saving $75 on a $2000 laptop is nice, but selling one more $2,000 laptop because of the silicon is much better.
Photo by Craig Whitehead on Unsplash