Samsung’s Gambit, as per The Economist
Samsung recently announced plans to construct a new semiconductor fab, to be completed in the second half of 2017, with anticipated CapEx of 15.6T KRW (roughly $15B or so US at current exchange rates).
The stated purpose:
The new fabrication plant is expected to help Samsung Electronics to develop a semiconductor cluster in Giheung, Hwaseong and Pyeongtaek to meet growing demand for advanced semiconductor products. However, the Company has not determined the specific products to be manufactured.
The Economist reported on the new fab in the context of Samsung’s utterly atrocious fiscal Q3 2014 guidance (there isn’t any other way to fairly describe a year-over-year quarterly operating profit decrease in the 60% range). The article, titled “Saved by the chips”, wraps up as follows:
… Samsung’s biggest problem is that rival mobile devices are increasingly being sold alongside software and services that it cannot match. As Samsung is structurally a hardware company, it would surprise many if it suddenly came up with great apps or extra services. The firm’s best chance may be to stick with gadgets, and to try to create ones that consumers love so much that they fly off the shelves.
Samsung seems to have a different plan, however. It is betting that its chip business, which has done well in the third quarter, will provide more of its growth. On October 6th the firm announced that it will spend nearly $15 billion on a new semiconductor plant in South Korea to meet the growing demand for processors in mobile devices. Although the decline of its smartphone business will not be an existential threat to Samsung, it remains to be seen whether making chips will replace all the profits it has lost.
So according to The Economist: Chip business did well in third quarter. Will provide more of its growth. Decline of the smartphone business not an existential threat. (Growing demand for mobile device processors, we’ll take as given.)
There’s just one thing. The notion that Samsung’s “chip business” will “save” Samsung is more than slightly flawed. And I’m quite sure Samsung knows it.
A Quick Baseline
What’s a chip? Well, The Economist says “processors in mobile devices”, so if you take that at face value, that sounds a lot like systems-on-a-chip such as Exynos, Apple’s A-chips, Qualcomm’s Snapdragon series. But semiconductors as a category are broader than that – they include discrete NAND flash and RAM chips for PCs (and Macs), which Samsung has long supplied worldwide as a top-tier componentry fabricator. So let’s start with the least restrictive category.
Here’s a quick history of Samsung’s recent financials. Yes, it’s not comprehensive, but it’ll still provide good-enough context.
In Samsung’s fiscal 2012, total “Semiconductor” category revenues were 34.89T KRW, with 4.17T KRW in operating profit (it’s a subset of the “DS”, or Device Solutions, revenue category). In fiscal 2013 (same link), Semiconductor revs rose to 37.44T KRW, and associated operating profit to 6.89T KRW. So: 7.3% year-to-year revenue growth, an impressive 65% increase in operating profit. For reference, in the same interval, “IM” revs (mostly smartphones and tablets) and operating profit went from 105.84T KRW >> 138.82T KRW (~31% increase), and 19.42T KRW >> 24.96T KRW (~28.5% increase) respectively.
In fiscal Q1 2014, total Semiconductor revs increased to 9.39T KRW from 8.58T KRW in the year-ago quarter (~9.4% increase), while operating profits increased to 1.95T KRW from 1.07T KRW in the year-ago quarter (yep, an impressive 82.2% increase for that quarter).
In fiscal Q2 2014, total Semiconductor revs increased to 9.78T KRW from 8.68T KRW in the year-ago quarter (~12.7% increase), while operating profits increased to 1.86T KRW from 1.76T KRW in the year-ago quarter (~5.7% increase)
Samsung’s Semiconductor business is indeed a bright spot amidst the recent, jarring profit trend reversal. But it’s much less of a salvation to Samsung than The Economist implies Samsung believes it to be, for three reasons.
Scale, That Other Chip, and the Uncertain Future
First and arguably most important, there’s the issue of scale.
As the financial data above indicates, Samsung’s Semiconductor business is growing in the high-single-digit-to-low-double-digit range. Even if we were to assume a revenue growth projection of 10-12% year-over-year, the scale just isn’t there.
Of course, in “absolute” terms – a ~$40B US (and growing) business is nothing to scoff at. However, Samsung Electronics is a ~$220B run-rate company – of which the IM business, in the previous fiscal year, made up over $130B. One guess where Samsung’s financial pressure is chiefly derived from for fiscal Q3 2014 (and the quarter prior). And let’s not forget, the stunning drop in Q3 operating profit year-over-year incorporates any offsetting gains from the Semiconductor business. Which itself is unlikely to sustain consistent operating profit gains over 60% year-over-year – it’s not nearly on pace for that in the first half of fiscal 2014.
Semiconductors are absolutely key to the foreseeable future of tech, no question. But Samsung is not new to electronics, nor is it new to semiconductors. If it’s going to suddenly (and anomalously) hyper-grow that category – much less “balance out” the massive downtrend in IM – there remains much work to do.
Second, Samsung’s Semiconductor unit growth appears to principally derive from memory products.
In other words, not processors in mobile devices. What are “processors in mobile devices”? Well, in Samsung’s business lingo, “System LSI” – “logic products” which, these days, are probably mostly application processors, such as, well, Apple’s A7 chip most likely.
“What makes you think DRAM/NAND memory is the higher performer in Semiconductor?”, you might ask. Samsung’s own business line outlook.
From the fiscal Q1 2014 presentation (page 5):
And from the fiscal Q2 2014 presentation (page 5):
Expect tight supply and demand market condition to continue under strong seasonality
– DRAM: Expect strong demand from major applications amid limited supply growth; Demand to increase driven by new mobile product launch and Chinese mobile market growth
– NAND: Expect strong mobile and Enterprise/PC SSD demand growth, while the overall demand expansion to be driven by increased contents for major applications, including card
System LSI: Expect weak customer demand for high-end mobile AP to continue, amidst new 20nm mobile AP product launch
※ 14nm product development is on track to begin mass production by year-end; well in-line with our plan to secure customer base
Does System LSI sound like the higher performer here?
Third, that $15B or so semi fab? It’s about 2.5 years away from operation. A lot can happen between then and now. Also, as I quoted above, Samsung “has not determined the specific products to be manufactured”.
The Importance/Challenge of IM
True, this is an Apple-centric blog, but I’m not counting Samsung out despite its worse-than-expected negative profit trending. I didn’t count Samsung out last quarter (though its relevance in smartphones is definitely on the wane), and I agree with The Economist that there’s no existential threat to Samsung…at this point.
Thing is, much the same way as iPhone is a “liability” to Apple due to its “revenue dependence” thereon, Samsung gets far more revenue from handsets (and tablets) than its Semiconductor business could possibly replace anytime soon. Semiconductor operating profit was far outpacing revenue growth (recently), which is helpful, but DRAM and NAND, as commodities, are inherently vulnerable to commoditization – Samsung’s not the only game in town, and a certain $175B-ish client company likely isn’t too keen on sourcing more componentry from Samsung than necessary. An additional problem? Samsung’s mobile devices, which include RAM, NAND, and often Exynos SoCs, are directly related (though abstracted by accounting) to Samsung’s Semiconductor business. Should Samsung’s IM shipments ever drop, or the mix trend to the lower end of the market, that impacts the scale, variable costs, technological advancement, and operating capacity of the semi operations.
Ultimately, The Economist’s own advice, however difficult to successfully implement, may be Samsung’s best way forward in the foreseeable future: “The firm’s best chance may be to stick with gadgets, and to try to create ones that consumers love so much that they fly off the shelves.” Unlike The Economist, I think Samsung is well aware that growing the “chip business” is not enough given the sheer revenue scale and (until recently) robust profits of IM. But as Samsung now sees, stabilizing the IM business will not be an easy task.