In 2014, chip tear-down reports revealed that Apple's (NASDAQ:AAPL) then-new A8 applications processor was manufactured exclusively by Taiwan Semiconductor Manufacturing Company (NYSE:TSM). This was something of a surprise to many because TSMC's main rival in chip manufacturing, Samsung (NASDAQOTH:SSNLF), had been the exclusive manufacturer of Apple's A-series chips for several generations.
With the A9 applications processor, Apple split the orders between TSMC and Samsung, but the A10 Fusion and A11 Bionic were both built solely by TSMC. Moreover, Apple's upcoming A12 processor, which should power the trio of iPhones Apple is expected to launch this fall, is expected to, yet again, be manufactured only by TSMC using its 7-nanometer technology.
According to a new report from DIGITIMES, which cites industry sources, Samsung is "seeking to win back orders from Apple in 2019" with its own upcoming 7-nanometer manufacturing technology, called 7LPP.
Let's take a closer look at what DIGITIMES had to say, then I'll offer my thoughts on what's likely to happen.
Samsung's risky bet
TSMC's 7-nanometer technology uses traditional immersion lithography tools, so as long as TSMC can come up with a good manufacturing recipe using these proven tools, there's little risk of any show-stoppers for potential customers looking to use TSMC's initial 7-nanometer technology.
In fact, TSMC says the technology is in mass production right now, so the risk that the technology won't work is zero.
Samsung's 7LPP technology, on the other hand, relies on a new type of lithography known as extreme ultraviolet, or EUV, lithography. The upside to EUV, assuming the tools for it are ready, is that it can lead to a simplification in the manufacturing process compared to traditional immersion lithography. The downside is that it's not clear if the technology is ready for prime time yet.
According to microprocessor analyst David Kanter, EUV is "definitely not HVM ready" -- "HVM" is an acronym that stands for "high volume manufacturing."
DIGITIMES seems to echo that sentiment, saying, "industry watchers said that the EUV 7nm process involves yield-rate issues and quality risks." The report goes on to say that although Samsung has reduced its contract chip manufacturing price quotes by 20% "to attract orders" from major companies like Apple, Samsung has seen "little response from them, probably due to the risks involved."
In other words, potential customers don't seem to think that (presumably) more aggressive pricing on Samsung's part is enticing enough to get them to bet their businesses on Samsung's relatively risky 7LPP manufacturing technology.
This makes perfect sense
Ultimately, the companies that contract out the manufacturing of their chips want to be sure that when it comes time to ramp the chips they need into mass production, there won't be any show-stoppers. If the chips they need can't be manufactured in the quantities required on time, then those contract chip manufacturing customers' businesses could be in huge trouble.
They could suffer from supply bottlenecks, leading to their own customers being displeased with them and causing a near-term revenue shortfall. Moreover, the cost structures on the chips that do get manufactured could be unreasonably high if the yield rates on the technology are too low (although more powerful customers could conceivably force the contract manufacturer to eat those costs, in that case).
In effect, it's crucial for chip design companies to sign on with contract chip manufacturers they can trust to deliver the chips they need at an acceptable cost structure in the right volumes and on time.
TSMC clearly took a less risky path with its 7-nanometer technology than Samsung did, and I think that because TSMC's 7-nanometer technology is in mass production now, while Samsung's is expected to go into production sometime later this year (if Samsung actually hits its arguably aggressive schedule), potential customers are going to think twice before abandoning TSMC in favor of Samsung -- even if they could, in theory, save a little bit of money by doing so.
Ultimately, I don't think Samsung will be able to wrest away even a portion of Apple's chip orders from TSMC. TSMC has been a trustworthy supplier of Apple's that has met Apple's demanding product introduction timelines for years now, and it seems to be making technical decisions that are sensible and as minimally risky as possible.
My guess is that Apple will want to keep its relationship with TSMC strong given TSMC's track record, and I'm confident that TSMC won't want to lose Apple's business, either, so I don't expect TSMC to go nuts in taking advantage of its leadership position in chip manufacturing by trying to charge Apple too much.
Ashraf Eassa has no position in any of the stocks mentioned. The Motley Fool owns shares of and recommends Apple. The Motley Fool has the following options: long January 2020 $150 calls on Apple and short January 2020 $155 calls on Apple. The Motley Fool has a disclosure policy.