Over on SemiWiki, Robert Maire asks whether the "winner" of Apple's (AAPL -1.34%) A9 processor business is the "real loser." He suggests that given the way Apple treats its suppliers, coupled with the aggressive pricing the company can secure, this might be the case. One crucial observation in the piece is that Apple is playing suppliers Samsung (NASDAQOTH: SSNLF) and Taiwan Semiconductor (TSM 1.91%) like "a Stradivarius" and that the real winner here is Apple.
While I fundamentally agree that the kinds of volumes Apple brings to the table allow it to push suppliers to deliver better technology at competitive pricing (helping Apple's own profit margin), I disagree that the "winner" of Apple's business ultimately ends up "losing" from a financial perspective.
Don't confuse gross margin percentage with overall profitability
Generally speaking, a customer that is willing to buy in bulk can get a per-unit discount. This is true in just about every business. The idea is that if a customer can guarantee large orders, the vendor is willing to take a per-unit discount for the sake of adding significantly to both revenue and raw gross profit.
Let's assume that it costs Taiwan Semiconductor $2,500 to build a semiconductor wafer. Now, assume the company has two different customers, call them Customers A and B. Customer A is willing to buy 1,000 wafers, but Customer B only needs 100 wafers.
If Taiwan Semiconductor sells 1,000 wafers to Customer A for $4,000 apiece, then it generates $1,500 in gross profit per wafer for a gross profit margin percentage of just 37.5%. On the other hand, Taiwan Semiconductor can sell 100 wafers to Customer B at $5,000 apiece. This means $2,500 in gross profit per wafer for a much larger gross profit margin percentage of 50%.
However, it's pretty easy to see that in terms of total profit dollars, Customer A's business is more valuable than Customer B's, even though on a per-unit basis Customer B's business looks better.
When would this make a difference?
The only situation in which taking Apple's lower-margin-but-higher-volume business would be a "bad deal" would be if a supplier were capacity constrained. In other words, if it had the option to sell 100 wafers at $5,000 apiece to 10 different customers or 1,000 wafers at $4,000 apiece to just one customer, then it might make financial sense to sell to those 10 different customers.
However, I don't think either Taiwan Semiconductor or Samsung -- the two companies said to be supplying Apple's A9 chips this coming fall -- face this sort of trade-off. If I'm correct about that, then it's hard to see either company "losing" in supplying Apple's A9 chip.
A diverse customer base is key to foundry success, but Apple business is good
Generally speaking, customer diversification is good, and I'm sure neither TSMC nor Samsung wants to be critically dependent on Apple's orders to keep their respective chip manufacturing businesses viable.
However, it's hard to deny that Apple brings a lot of volume, and -- perhaps more importantly -- a lot of volume at leading-edge technology. As Intel likes to point out at its investor conferences, semiconductor manufacturers that don't have leading-edge volume eventually lose the scale required to continue investing in leading-edge technologies and factories.
Despite the relatively low per-unit margins that Apple suppliers typically receive, winning Apple's business is still, at the end of the day, very strategically important to the remaining semiconductor foundries -- particularly if the iPhone continues to gain share against high-end Android/Windows phones.