Analysts with BlueFin Research Partners recently claimed that Apple (NASDAQ:AAPL) wouldn't adopt Taiwan Semiconductor Manufacturing Company's (NYSE:TSM) second-generation 7nm chip manufacturing technology, known as N7+, for cost reasons.
N7+ promises a 20% reduction in chip area and a 10% improvement in power efficiency. It will also employ a new lithography technology known as extreme ultra-violet, or EUV, to handle the patterning of some of the smallest features on the chips.
Although this report seems a little bit puzzling, as Apple tends to aggressively move to next-generation chip manufacturing technologies, BlueFin Research Partners does have a good pulse on the Apple supply chain and has reported accurate information in the past. So, I'm not willing to discount it.
With that being said, I'd like to go over some statements that TSMC made in its most recent earnings call, in mid-April, to try to figure out if BlueFin Research's information is correct.
TSMC's N7+ comments
TSMC co-CEO C.C. Wei said that N7+ "will start to ramp up in the second half of next year quickly," adding that this ramp-up would be due to wafer shipments to smartphone customers.
On the surface, it might seem as though it's obvious that Apple will use N7+. After all, Apple sells premium smartphones and can afford to use the best manufacturing technology available. However, comments from TSMC's other co-CEO, Mark Liu, seemed to paint a less rosy picture for the kind of volume that TSMC would ship on N7+.
In response to a question from analyst Randy Abrams, who wanted to know how important it would be for TSMC to see "wide adoption" of N7+ to gain experience in manufacturing chips using EUV, Liu indicated that such wide adoption wouldn't be necessary.
TSMC says that it merely needs some volume on N7+ to prove the viability of EUV technology in a production process, suggesting that the company doesn't need to come close to shipping as many N7+ wafers as it plans to ship N7 wafers.
Perhaps some of TSMC's other customers besides Apple are proving somewhat hesitant to adopt N7+ after N7, but if Apple were signing on to use N7+, then TSMC would need to be gearing up to produce pretty large volumes of N7+ chips during the second half of 2019.
There's probably some truth, here
Ultimately, I think, based on TSMC's seemingly lukewarm commentary around the demand for N7+ (compared to its gushing comments around N7 adoption), there's a reasonable chance that Apple may, in fact, be sticking with standard N7 for the A13 chip in 2019 rather than going with N7+.
But Apple would be hurting itself by not adopting the newer, more advanced technology as it promises meaningful area and power efficiency enhancements over N7. Moreover, I can see some of Apple's competition -- namely HiSilicon, which is Huawei's chip development organization -- migrating to N7+ to try to get an edge over the competition.
If Apple doesn't go with N7+ and instead sticks with N7 for the A13 chip, then it'd better have some nice design tricks up its sleeve to make the A13 more powerful, more feature-packed, and more power efficient to compensate for the lack of an improved manufacturing technology.