A new chip manufacturing plant, known in industry lingo as a fab, costs anywhere from $2 billion to $5 billion. To be a serious, profitable player in this business, you need several fabs and the orders to keep those factories humming all the time.
A few years ago Samsung Electronics made a bold decision to jump into this brutal, high-risk business. It already had factories that produced the memory chips found in all manner of computing devices and decided to complement those with even more complex factories that could build the computation chips that go into smartphones, PCs, and the like. This move put Samsung in direct competition with industry giants Intel, Taiwan Semiconductor Manufacturing, and GlobalFoundries and launched a global war that has yet to play out fully.
The strategy here, as my colleague Sam Grobart points out in his cover story this week on the company, is classic Samsung. It finds a tough-to-make component—in this case, computer chips—and takes the plunge into learning how to make it. From there, Samsung gains insight into how the companies that build consumer products around that component operate. And soon enough, Samsung shows up with its own end products.
The chip manufacturing story, though, is particularly compelling. Samsung managed to land Apple as its first massive computation chip customer when the original iPhone was heading to market. Apple needed someone to make the ARM chip that would give the iPhone its low-power consumption and zip, and Samsung said it was up to the task. “Apple basically told Samsung what to do, and they designed and built the whole thing,” says David Kanter, a chip industry consultant.
When sales of the iPhone exploded, Samsung morphed into a chip-making powerhouse. It needed to pump out tons of chips, which meant its factories were full. This scenario persisted as Apple released more iPhones and then iPads, all of which run on Samsung-made chips.
This arrangement has grown beyond awkward, however, as the years have progressed. Samsung now makes huge investments on new chip fabs designed just to handle the Apple work. Meanwhile, it’s competing directly against Apple with its own smartphones and tablets, and the companies continue to sue each other over all kinds of things. It’s a love-hate-make-up relationship.
Rumors have floated for months that Apple may turn to Taiwan Semiconductor as a second chip supplier or migrate entirely to TSMC. But such a shift would come with serious consequences. “With Apple, it’s not just about performance,” Kanter says. “It’s about having a supplier you can trust. TSMC has a lot of leaks and is kind of challenging in that respect.” (By rights, Samsung’s chip-making group is not meant to talk to other parts of Samsung about customers’ plans. Samsung has largely proved that it’s up to this ethics challenge by trailing Apple with lots of key chip advances and component features.)
Outside of Samsung and Taiwan Semiconductor, Apple could turn to Intel for its chips. This again presents a major challenge. Intel is used to making its own chips instead of acting as a contract manufacturer that builds other companies’ designs.
Still, it seems unlikely that Samsung and Apple can maintain the relationship that underpins their roles as kings of the mobile revolution. “I would imagine that Apple would like to figure out a way to decrease its reliance on Samsung, and not just for computation chips but also for displays and other things,” says Kanter. “By giving Samsung so much business, it allows them to invest in developing things like retina displays, which they can use for their own phones.”
If Apple does shift from Samsung, you can expect to see a massive global shakeout among component suppliers. But the big plus Samsung has earned from its rising popularity is that it now has enough volume from its own smartphones, tablets, and other devices to consume a lot of its chip and other component volume.