Over on Seeking Alpha, contributor Black Coral Research recently argued that were Apple (AAPL 1.75%) to transition away from Intel (INTC 7.63%) processors in its Mac line of computers, contract chip manufacturer Taiwan Semiconductor Manufacturing Company (TSM 2.27%) would be a big winner.
"TSMC is benefiting from the halo effect of solid iPhone 7 sales, but its real potential is when Apple decides to design its own silicon for its Mac," the contributor writes. "We estimate that TSMC could generate $5.5 billion in annual revenues from such a venture, leading to 20% revenue and earnings growth."
This analysis doesn't stand up to scrutiny. Here's why.
The current Mac supplier doesn't even generate that much
The biggest problem with this analysis is that it assumes that TSMC will be able to charge Apple a whopping $300 per chip for Apple-designed Mac processors.
This is absurd, because this would imply that TSMC would make significantly more from building each Apple-designed Mac processor than current Mac processor provider Intel (INTC 7.63%) makes on average from the chips it sells Apple.
A quick way to "prove" this assertion is as follows: Apple ships roughly 20 million Mac computers each year, give or take a few million. If Intel sold processors to Apple at an average of $300 apiece, then Intel would generate roughly $6 billion per year in revenue from Apple.
Per Intel's form 10-K filings for 2013, 2014, and 2015, Apple wasn't responsible for at least 10% of Intel's revenues in those years. In those years, Intel's revenue was solidly south of $60 billion, so we know that Apple pays much less than $300 for Intel processors on average.
Further, even if Intel did take in that much per chip from Apple, Apple's per-unit costs would be lower for Apple-designed chips manufactured by a third party, since the third party isn't getting paid for the chip design.
The iPhone opportunity is much larger than the Mac opportunity
The Seeking Alpha post suggests that TSMC would have a bigger opportunity to profit from Apple's building an in-house Mac processor than it does from having a monopoly on the iPhone 7 A-series processor business -- as well as the A8, A8X, and A9X chips that currently power the iPad.
That doesn't make sense. In Apple's last fiscal year, it shipped nearly 212 million iPhones and 45.59 million iPads. It shipped just 18.484 million Macs in that period.
In terms of chip area, and therefore chip dollar content to TSMC, an Intel processor isn't much larger than a high-end iPhone or iPad processor. The A10 Fusion measures in at 125 square millimeters in TSMC's 16-nanometer process; the A9X comes in at 147 square millimeters.
These are already large, high-silicon and thus high-dollar content chips to TSMC. Even if we assume Intel's average content across Apple's Mac lineup is twice that found in an iPad processor, TSMC's winning the Apple business over from Intel would amount to the addition of roughly 36 million iPad processors' worth of revenue.
In TSMC's 2015 form 10-F filing, the company said its two largest customers, one of which is almost certainly Apple, each made up 16% of its total revenue in 2015. That means TSMC's Apple-related revenue in 2015 -- consisting of A8, A8X, and some A9 chip sales -- was around $4.12 billion.
It's unlikely that approximately 18 million Mac processors could generate anywhere close to the kind of revenue for TSMC as all those other chips do combined. Indeed, even if TSMC can capture around $50 per chip for a potential Apple-designed Mac processor -- which is probably too high of an estimate -- then this would amount to annual revenue of less than $1 billion per year for TSMC.
As TSMC's annual revenue approaches $30 billion, even $1 billion in incremental revenue would be good for only around 3.33% growth. That'd be good business, but hardly a game changer for the chipmaker.
Indeed, capturing 100% of the A10 Fusion chip orders for the iPhone 7-series phones, up from a reportedly minority allocation for the A9 processors in the iPhone 6s-series phones, was probably a bigger deal to TSMC than a potential Mac win ever could be.