We have been examining the decline in Apple’s gross margins over the past four years, and working out the possible causes of that decline. We are not really Apple analysts, but we wanted to understand why Apple’s margins are declining at the same time that they are shifting to their own silicon. In theory, shifting to their own chips should help their gross margins by effectively eliminating the middlemen, the merchant silicon vendors, aka Intel and Qualcomm.
In this piece we want to look at the math behind Apple Silicon, and it turns out that math is fairly surprising.
Some very rough math based on Google searches from reputable sources (no Statista). Apple will sell about 200 million iPhones this year, 18 million Macs and 30 million iPads. We do not know how much they pay other chip vendors, but a rough guess would be the iPhone and iPad processor costs around $30-$40 and a MAC CPU around $100. These could be off by a fair amount, but not enough to change the outcome meaningfully. Multiply all that out and we get about $10 billion in silicon cost. Apple’s cost of goods sold for hardware last year totaled about $130 billion, so silicon is a little bit less than 10% of that.
As a sanity test, we compared the cost of the bill of materials for an iPhone, which that linked report estimates is about $406, with the A14 processor valued at $40, again a little bit less than 10%. We are 100% certain that some of our numbers are wildly off, but they will be close enough to still make our case.
Now let’s look at the cost of Apple’s own silicon. Some rough estimates, Apple’s N series processors for phones and iPad are around 90 mm2 – 100 mm2, and the new M1 is 120 mm2 (tiny!). Using an online calculator these yield 666, 500, and 600 die per wafer, then factor in some errors and we get a total of 431,575 300-mm wafers. TSMC probably charges something like $15,000/wafer yielding a total cost of $6.4 billion. So Apple is saving $3.5 billion or so versus using merchant silicon. This alone is worth 1.7% pts of gross margin. For comparison, since 2017 Apple has lost 3.5% points of gross margin (for hardware), which fell to 31.5% in the latest fiscal year. Meaning, without the shift to internal silicon, Apple would be below 30% hardware gross margins by now.
However, we still have to factor in the rest of Apple’s costs. The biggest of these is the cost of the design team. We could not find a good estimate of how big the Apple Silicon team is. When the current head of the team joined, we know he had a team of 1,000 in Israel. We have to assume that is only part of the team and that it has grown significantly. So let’s say it is 5,000 people, could be a lot more, probably not that much less. That alone is probably $2 billion in cost. Then factor in fees they pay to ASIC vendors, probably Broadcom, who handle the back end work of readying the product for manufacturing, plus test, assembly and packaging costs and we get to $9 billion in total cost for Apple Silicon.
Compare that to the $10 billion or so it would cost to buy merchant silicon and something becomes clear – Apple is not saving that much money by building its own silicon.
So then why bother? The answer is that Apple Silicon gives Apple a distinct competitive advantage. Or to put it in terms that Apple is more likely to use, its silicon gives its customers a better experience. Read any of the reviews of the new M1 CPU for Macs, and the merits of this approach are obvious – faster, much less battery use etc. The same is true for iPad and iPhone, where the N Series has given Apple a big advantage in power consumption and image processing. We then have to factor in Apple Watch and their home speakers, and whatever else they have planned (AR glasses), and it is pretty clear that Apple is not into silicon for the money. In fact, we could make the argument that factoring in all costs and even larger design teams than we modeled here, that Apple Silicon may even be more expensive than using merchant silicon.
This highlights how expensive the consumer electronics business has become. Years ago, we wrote about how Apple’s iPhone strategy was to capture profits not market share, today they are going after something different. To compete in phones, companies need their own silicon or risk being marginalized to low price, low profit segments, companies like Transsion (we wrote about Transsion last year for our subscribers, drop us a line if you would like to sign up for our newsletters). At the same time, Oppo/Vivo and Xiaomi are investing heavily in designing their own processors. And this shows why Huawei is no longer in the phone business, now that they cannot build their own chips. With Apple Silicon, Apple has built immense barriers to building phones.
Pingback: Apple Si ? The Surprise Twist ? It's not About Money - Apple News·
Pingback: Apple is Building its Own Modem (Redux) | DIGITS to DOLLARS·
Pingback: Margin Stacking and the Cost of AI | Digits to Dollars·