The new age of Capital Intensity

In the post reviewing Samsung’s Capital Structure I noted that its component divisions have historically taken 90% of capital investments and that the overall capital intensity for Samsung Electronics has increased in proportion to its component revenues.

In another post regarding the capital structures of other technology companies with different business models I noted that Apple has changed its capital structure to a significant degree over the previous three years.

In the following graph I combined these observations to show how capital expenditure patterns may be used to discern the underlying business model.

I would group “cloud” or “platform” based companies like Google, Amazon and Microsoft as a cohort which, although spending significant amounts on capital equipment, do so mainly to support services. Their primary employment of capital is to sustain the infrastructure of data centers necessary to deliver the services underpinning their business models.

I would group Intel and Samsung as “component” based companies which use capital to build production capacity that delivers, hopefully, a control point in the value chains they participate in. Note that I am assuming that Samsung Electronics’ CapEx primarily targets semiconductors and displays as it did during the 2006 and 2007 period when data was available.

Therefore, given the patterns over the last six years, it might be possible to classify a company’s core business by just looking at its CapEx: The services group showing growth but relatively stable patterns and the components group showing more volatility, macroeconomic sensitivity and an overall higher level of spending, especially as a percent of sales.

The question that pops up then is in what category does Apple fall?

My observations for over a year have been that Apple is increasing its capital spending to such a degree that it is neither fish nor fowl. It operates data centers but presumably does not operate fabrication plants. And yet it spends more than Intel and Google combined.

Contrasting Apple and Samsung Electronics tells an even more interesting story. In the following graph I’ve added two additional items of information: First, although CapEx spending is recorded in the cash flow statement, Apple reported in its yearly (10K) report that it spent an additional $2 billion for equipment which did not pass through cash flow.

If we were to add $1 billion to each of the previous two quarters, the graph would look like this:

The second piece of information is that Apple projected its spending for the following fiscal year. I’ve added that spending in a pattern matching the last fiscal year.

Apple is still spending less than Samsung. In its fiscal year 2012 Apple spent about $10 billion while during the same time frame Samsung spent nearly $22 billion. However, Apple has begun to bump against Samsung’s spending level on a quarterly basis.

I consider this extraordinary evidence of an extraordinary shift in strategy.

For a year we’ve been debating here to what degree Apple has a “capital intensive” business and what exactly could it be spending capital on. One early hypothesis has been that it’s on data centers or on real estate for spaceship campuses. These can be easily dismissed on the knowledge that they can’t cost that much (and that the spending happens gradually). Another hypothesis is that the tooling for milled aluminum cases and other mechanical components is expensive and Apple has to buy a lot of it. But, again, the level of expenditure is in the league of Intel and Samsung which make vast quantities of semiconductors in very expensive plants. It’s hard to imagine milling machines being cost equivalent to chip foundries.

The comparison with Samsung, whose vast capital commitments are legendary and are its primary strategic asset puts Apple in a new light. Perhaps we can’t yet call Apple a semiconductor company or even a fabricator but they are putting their “money where their mouth isn’t”. We can infer that it’s not going where the company currently operates. The capital is being deployed almost silently and, though vast in scale, barely gets a mention from analysts. Not even a single question has been raised at any earnings call about this spending.

So what tale does this capital tell?

One interpretation is to apply the value chain disruption theory and ask whether Apple is acting to counter a potential disruption. In other words, that this is a specific threat response from an incumbent. There is some logic to this because  the antagonism between Samsung and Apple is so strong that it might justify a reaction of this kind.

However, I believe that there is a more measured process at work over a long period of time. The one distinguishing advantage and attribute that Apple has  is its preference for integration. This preference is not a secret and Tim Cook repeats it often. The integration is evident with their move into retail and component design, in other words in both directions along a value chain. So my hypothesis is that this is part and parcel of an ongoing re-alignment of their role in the value chain.

This second explanation suggests that Apple simply wants to control all the areas which affect the value they produce. First because they feel that there are areas that are not good enough and the product/experience cannot be good enough if its constituent parts are not good enough. Second because without control there are constraints to what it can do strategically.

Whatever the reason, I read the evidence of increasing commitment and engagement in parts of the value chain as part of a continuous evolution of Apple’s role. Furthermore, it’s something that should be seen as a signal of a new era in how technology companies operate. We see hints of “vertical integration” with Microsoft building hardware, and Google buying Motorola and Amazon selling devices. Apple did all these things and now it casts an eye over the next frontier: components.

If you want to learn more about the future of the technology industries and the evolution of value chains, consider participating in Asymconf.

  • Semicon / LCD lines cost 4~5 B$ per line… Apple doesn’t seem to be fully investing whole LCD lines or semicon lines yet. Maybe co-investing with existing vendors?

    • KirkBurgess

      I presume the cost of building one of these lines would be spread over the multiple quarters of its construction time, so $4-$5 billion would not be visible.

  • Kizedek

    “Note that I am assuming that Samsung Electronics’ CapEx primarily targets semiconductors and displays as it did during the 2006 and 2007 period when data was available.”

    Horace, a couple of posts ago you were discussing the huge capital expenditures that Samsung was making in the area of marketing/promotions — some 12 billions or so. Which we took to mean that they are targeting the final product of the mobile divison more recently, versus targeting the ramp up in production of commodity “semiconductors and displays as it did during the 2006 and 2007 period” as noted above. Are you rethinking, or is Samsung making huge expenditures in both areas?

    If both, and if Samsung’s expenditures in marketing/promotion are so great (more than CocaCola, etc.), are the cap ex amounts for semiconductors even greater in relation, in order to both serve customers like Apple and make the huge number of Samsung phones they are selling?

    If Samsung is spending so much (both 12 billion for marketing and promotions, AND piles of cash for semiconductor expansion), then the big question is:
    How does all this spending compare qualitatively with Apple’s spending? Is Samsung simply “buying” marketshare by both investing in higher production capacity of their own not-so-terribly-innovative commodity parts (so that they don’t have to buy them in), and by making sure the finished phones get into customers hands?

    Or, is Samsung being strategic with its cash like Apple, which is doing things like:
    buying up specialized CNC milling machines and installing them on partner assembly lines; buying PA Semi; buying the flash controller company in Israel; building out infrastructure with data centers; etc.?

    • Marketing budgets are different from capex budgets. Marketing semiconductors to consumers is unorthodox as is spending billions buying tooling for assembly of components. That’s not to say that they are not done, Intel does have a budget of billions for marketing (i.e. Intel Inside) and Apple could be spending hugely to assemble iPhones, but I doubt the numbers add up to what we’re seeing here. I am under extreme strain trying to convey just how big these numbers are. Perhaps another tack: what Apple and Samsung spend in one year on equipment for use in building what they sell is worth more than the market cap of all their competitors combined.

      • newtonrj

        If we agree that Samsung is copying the Apple product line, why can Apple copy the Samsung production line? -RJ

      • SSShu

        Blueprints for a product that you piece together and can use is easier to mimic than say…. plans for an entire production/assembly line system, factory and staffing regimes.

      • Agree Horace. I can’t think of what could possibly eat up billions of cash other than equipment for a semiconductor fab (or maybe an LCD fab) or both.

        I once heard that fabs are the second most expensive man-made structures – oil refineries being the first ($9B+ each) and I suppose we discount Apple’s participation in the petroleum business (at least for this year 🙂

      • Indeed, and Oil refineries take far longer to build so the rate of spending on fabs is by far higher. (I tried to draw an analogy to nuclear aircraft carriers but they’re also too slow to build.)

  • For the moves we know, cpu design, both your reasons stand: the A6 is better than other cpu on sale (that were not good enough for apple) and strategically they are now at least six months ahead of competitors in computing power.
    Acquiring PA-Semi and owning an agreement with arm they were able to produce they own cpu, all the licenses and patents were available.
    To produce sd-ram or displays or socs they have not acquired the necessary licenses or companies that got the patents, for now and for what we know.
    Furthermore the spending is high but lower then that necessary to build new production lines for the hundred of millions of components required.
    It should be coinvesting with existing vendors, and here is the question: coinvesting is enough to avoid disruption from suppliers?

    • OpenMinde

      Again, if Apple want to produce their own semi components, they need to buy Intel. Intel fab produce more components than just CPU. Intel have sdram, flash, soc, etc. And Intel own all the necessary licenses or cross-licensing enough patents to get coverage. Buy Intel, send CPU design to AMD or Microsoft, keep fabs.

      • I don’t see Apple to enter a new market buying Intel, they just want to make better in their market.
        Buying Intel to destroy it and using only the fab for their components production makes no sense, first it could be not possible, there are contracts to be respected, second antitrust could not agree.
        They could make an agreement, covinvesting, buying a participation, that could solve the reasons identified by Horace, but could it avoid disruption?
        Intel is closed in an diminishing market, if apple brings intel in the mobile market could that be a mistake in the long run?

      • Peter

        Partnering with Intel would certainly be dangerous as it may guide them to dominance in the mobile parts market, which will then also be available to Apple competitors. Buying Intel outright might become as massive management distraction in terms of antitrust and integration issues. Better to go for a much smaller partner.

      • I agree

      • I don’t think Intel’s problem in the mobile market would be fixed by a deal with Apple to make Apple’s chip designs. Intel’s problem is mostly due to its own insistence on keeping the x86 architecture while moving into the mobile chip space.

        I don’t think there’s any reason why Intel couldn’t license ARM cores or architecture and build a killer ARM SoC on their cutting edge processes. They’ve got the fab scale, they’ve got the best fab technology, they’ve got the designers, and the ARM core is easily licensed. Nothing is stopping Intel here except, like Microsoft, they can’t let go of their success in the 1990s. So they’re only going to do mobile with an x86 design, and they’re having trouble mapping that to a low-power world.

        As far as I can tell, they think they can make up architecture inefficiencies by having a better process than the ARM vendors. This *may* work, but it hasn’t so far. And they keep trying…

        So I don’t think Apple moving to Intel for fab would do anything particularly good for Intel in the way of Intel disrupting the mobile market.

      • If they partner Intel will make killer arm soc for Apple, learning how to do them with an head start and a big order, then they will sell arm cpu to others and perhaps design arm chip better than Apple itself, closing apple’s advantage

      • The thing is, I don’t think Intel would learn much in making ARM chips for Apple that they couldn’t by just taking an ARM license and doing some experiments on their own fab line, unless you think they’d reverse engineer Apple’s A6 design. But I expect Intel’s CPU designers are probably even better than Apple’s, it’s the one thing they’re really good at. (Architecture not so much.)

        They might gain a little from Apple doing a big ARM SoC order, but I don’t think they’d get nearly as much benefit out of that as, say, Samsung. And I don’t think they want to go that route anyway — they want architecture lock-in based on x86, and *everyone* can make ARM chips, so with ARM designs, Intel’s advantage diminishes to simply whatever lead they have in process, typically feature size and yield.

        If Intel designed their own custom ARM cores, like Apple’s A6, then they might have a bigger advantage. So far, though, Intel has abandoned any CPU they had which wasn’t x86-based. They *really* love that architecture lock-in. Unfortunately, with the decline of Windows and the rise of the x86-free mobile market, that lock-in is dying.

      • As I understand it, Intel’s current chip architecture is a hybrid between CISC and RISC. The X86 CISC instructions are translated (through a proprietary process) into RISC instructions and then executed. Those RISC instructions could certainly be executed on an ARM chip.

        If the translation process could be performed by, or incorporated into, the ARM architecture then you would have an Intel compatible ARM chip — with all the advantages that would entail.

        I don’t know the current status, but IBM used to have some pretty good foundry capability. It is possible that Apple could invest in or use IBM as a foundry. This could be done at a very high level in both organizations with few people and little chance of leaks.

      • The Intel “RISC” you refer to is, I believe, actually a lower-level micro-code machine specifically designed to implement the x86 instruction set efficiently, and which almost certainly has only a vague resemblance to the RISC instruction set in the ARM processors. Just as the ARM RISC is only remotely similar to other notable RISC machines, like the MIPS core. So no, I wouldn’t expect the Intel micro-operation “RISC” instructions to come close to executing cleanly on an ARM core.

        What you’re talking about is basically dynamic machine-code translation, which has been done before in software (e.g. the Rosetta PowerPC environment in the Intel Macs), but I don’t think it’s been done in hardware, and I don’t think it’s likely to be. It’s generally much easier to recompile or translate at the source or binary code image level.

      • In the design of a6 Apple put years of experience on mobile devices use and power saving, things you can’t find in a lab. Entering arm arena with a big order making the best design can mean one or two years of time saving instead of going alone.
        I don’t think It will happen I am only saying that Apple should carefully evaluate all options and future scenarios before making an Intel move.

      • If they’re just running the wafers, I don’t think they’d get much more than they could by deconstructing an A6 out of an iPhone 5. I really think Intel knows its process limits and design rules very well, so running a lot of wafers at scale probably won’t tell them much new.

        On a new process, yes, they’d learn a lot. But they’d use their best processes for the high-end x68 lines first, so they’ve already learned most of the problems there. This is different from TSMC, which would be bringing in a bleeding-edge process for Apple’s new SoC.

        The other key point here is that a lot of Apple’s power management comes from system-wide integration and tradeoffs, not just the silicon. Intel taking Apple’s ARM design (assuming they could, somehow) and slapping someone else’s software on it might not do nearly as well.

        From what I can recall, Android phones with similar chipset specs to Apple’s in the earlier days did poorly on battery life, or needed bigger batteries. Presumably this was due to software inefficiencies.

        Intel would have to learn to do software too — something it’s never been good at.

      • I don’t think that deconstructing an A6 will give the same amount of information as running the wafer for a new A6 using different cutting edge processes, different from what apple/samsung is using now.
        Going from design to wafer is not an easy task and for the current A6 it has been hand optimized by apple targeting the samsung foundries.
        Using a new wafer process will require a new design and new optimization, intel knows the process, apple knows the chip, they will have to exchange a good amount of information to make it work.
        The parameters for the optimizations are given by the valuable use case experience of apple in both running apps and power savings.
        The necessities to cooperate for the wafer design will transfer to intel a lot of useful information that intel does not have the opportunity to gain alone in a short time.

      • katnrica

        “Intel would have to learn to do software too — something it’s never been good at.”
        – Are you sure? As of a couple years ago, Intel employed more Linux engineers than any company on Earth, except Red Hat.

      • It’s possible they’ve changed, but when I was dealing with them more regularly (some years back now), they were *really* bad at software. Also, I definitely wouldn’t correlate “number of engineers” with “good at” anything — there are plenty of counterexamples around for that. Or maybe I’m just too cynical in my old age…

      • KirkBurgess

        Intel has a dilemma – if it moves to ARM it becomes just one of half a dozen large companies all using the same architecture. Even if it becomes the largest, it will still have nowhere near the margins & marketshare it enjoys in the PC market.

        Intel’s dream is for mobile device manufacturers to switch from ARM to a proprietary intel chip design, but that will take a giant leap in design (at this stage it doesn’t just have to be a little better than ARM, it has to be a lot better).

        What’s more likely to happen is that not only will mobile device manufacturers keep on using ARM based chips, but PC manufacturers will also make the switch to ARM for the cost/weight/battery life advantages compared to the good enough raw power of x86. The fact that most users jobs to be done now no longer require x86, will accelerate this transition (Office going ARM was the final nail).

      • Intel already worked on ARM chips called XScale. It ended up selling everything to Marvell. Don’t know why.

      • I’d forgotten they’d inherited StrongARM when they bought DEC’s chip line. I worked with the Intel IXP network processors once, I should have remembered the control plane processor was an XScale ARM….

        Intel still owns the IXP line, according to Wikipedia they only sold off the mobile XScale line to Marvell, possibly due to the start of the Atom development. So they still have some ARM knowledge in-house, though it may not be in active development any more.

      • I worked at Intel for 11 years starting in 1995.

        Intel certainly saw the mobile wave coming and much thought was put into how to go after that market. There was always much fear that not winning mobile early would translate into big strategic problems down the road for the core business of desktops, laptops and servers. Xscale (ARM) and other approaches were considered but in the end the decision was made to try and drive x86 as far down into mobile as possible. This ended up being what we now know as “Atom”. The decision to sell Xscale was a direct consequence of sticking with x86 for mobile.

        The innovator’s dilemma Intel faced then (and faces even more so now) is that they have successfully built the highest margin, most profitable semiconductor business in the history of the world on x86. Doubling down on this and leveraging it’s manufacturing prowess (Intel’s true competitive advantage) was I think the only realistic option it had to go after mobile when it set the strategy about 10 years ago – anything else meant threatening the high margin business and would have been very hard to sell to Intel’s Board and major shareholders.

        So here we are in 2012 and Intel is a tiny player in the biggest semiconductor market segment (mobile) and ARM is doing exactly what they feared most – eating into the core business via iPad and even ARM servers.

      • Steve Rosenberg

        It’s clear from your description that the ambitions of the legacy x86 platform at Intel prevented the ARM segment from flourishing. As opposed to making the ARM segment a completely separate effort and allowing it to cannibalize the profit engine if it occurs. “Better us than someone else” is a more than courageous decision.

      • Tatil_S

        I don’t think lower margins on ARM tells the whole story. Intel has gone into relatively low margin semicon businesses, such as WiFi chips. It could have kept its mobile ARM business during the times there was no way mobile devices could have used x86 chips.

      • Companies ingest what they perceive to be opportunities but vomit them if they cannot digest them or if they are perceived as poisonous. Indigestion comes from a mismatch with processes and poison control comes from a mismatch with priorities.

      • Another way to think about it is: why is Intel’s process so much more advanced than foundries?

        All companies essentially tap into the same global talent pool. So clever engineers is not the reason. Business process can be copied, so it can’t explain the prolonged advantage in Intel.

        I think the answer is the monopoly in Intel’s x86 architecture: the old Win-tel alliance. The monopoly profit from Intel architecture allows it to out-spend everyone. Foundries like TSMC’s profit is simply not sufficient to compete with Intel on process technology front. The superb manufacturing process in turn makes Intel design more competitive, and the cycle continues.

        So Intel’s design arm and the manufacturing arm are co-dependent. That explains Intel’s hesitation to go into low margin foundry business. If the monopoly profit from Intel design does not come in, very soon it won’t have the money to pump into process technology development and the whole temple collapses.

        A very tricky innovator’s dilemma indeed.

        Going back to Apple, it has tons of capital waiting for investment. It too should be able to buy into process technology development if it chooses to. TSMC will never be at the forefront of process technology. But Apple can, by the virtue of its cash pile.

        Apple fabbing its own in-house designed chip running its own in-house OS will be the re-creation of mainframe era IBM. Like IBM back in the days it will be sole dominant 800 pound gorilla.

      • I often hear people flinging around ludicrous suggestions for Apple to “buy this multibillion dollar company, buy that multibillion dollar company.” Ask yourself the following question before you make the suggestion: Does Apple need that company?

  • vincent_rice

    Apple of course intends full vertical integration but sensibly seems to shy away from the social and political hassles of owning the massive production lines required. So, joint venture with Foxconn in Brazil – for the Apple TV maybe? Joint venture with Sharp for the screens required? Joint venture with TSMC for the A-series going forward?

    (One can’t help but see the full-court-press against Samsung as an indication of what partners can expect if they get out of line and start competing – the Dell/Acer scenario).

    I think some ‘learning’ may be coming the other way – A modest US factory run by Foxconn to manufacture the Mac Pro might be an excellent opportunity for Apple to understand how mass manufacturing in the US might work and how to repatriate some of the offshore earnings.

    Interesting times.

  • Ncgo4

    Alternatively, could it be that Apple is spending on the “next big thing” in a way different from past introductions. In an effort to control and keep secret the next introduction could they be doing significantly more in house, rather than through suppliers? This would make sense if they were introducing a battery breakthrough, or a unique automobile system intro or some kind of TV product or system.

    • Maybe, but with all that money flying around, you’d think someone would be talking about it — there are certainly enough leaks in Apple’s supply chain. Though maybe not so much on the production investment side, so I guess it’s possible.

      Probably the area most ripe for disruption right now is batteries — there are a lot of interesting things going on in that space at the research level, if if Apple locked onto one of them, it could make a big leap over the competition in that area.

      Less exciting from a user standpoint would be semiconductor fab integration — there’s not a lot to gain here that isn’t accessible to the competition.

      Spinning up a new product line factory would eat money, but mostly if Apple was doing it in-house. I suppose they could always be going into competition with Tesla….

      The main problem I have with the “new product” thesis is that I can’t see it eating money at the scale involved — Apple’s new product categories normally start relatively small and then ramp up as demand dictates.

    • I think this is spot on. People are forgetting the exclusive agreement Apple made for LiquidMetal. I could imagine Apple investing a ton of capital to get this tech to scale. It would give Apple a massive advantage. LiquidMetal would give Apple the ability to create cast parts with the quality of finished milled parts.

      I don’t think Apple is going to invest significantly in manufacturing and supply chain under these circumstances, particularly the intense competition with Samsung, unless it is going to derive a competitive benefit that its competition won’t easily be able to match. I’m surprised that Apple has not made moves to gain complete control of Sharp’s IGZO display tech, though it seems that it has made investments. It looks like Apple’s next strong move will be into 4K display technology. Rumors are surfacing of an iPhone 6 with a 4.8″ display at “Retina+” resolutions. There are IGZO display samples at almost 500ppi and the tech is stated to be production ready.

      I can see Apple devoting a lot of capital for LiquidMetal and IGZO to stay ahead of its competition. Outside of software, there aren’t many ways Apple can make its products more attractive to consumers and ergonomics limit how much Apple can innovate on form factor. Manufacturing innovations like LiquidMetal could significantly improve Apple’s costs to manufacture components over the long-term.

      • Interesting notion; I’d forgotten about LiquidMetal. Scaling that up ight be expensive.

        IGZO is certainly also a good place for them to go, but not for the increased resolution — it’s really kind of pointless except in a spec war, which Apple doesn’t engage in. I think Apple’s interest in IGZO is much more likely in size/weight/power efficiency, rather than DPI.

        Hmmm.. actually there’s one place those high-DPI displays would matter — ones much closer to the eyes, like eyeglass-mounted displays. Not sure why Apple would go there, though they’ve done some patents in that space. I really don’t see a killer product there, Google hype notwithstanding.

      • Agreed on IGZO, though I think the move to 4K for Apple would be a huge selling point. Look how long it took for everyone else to catch up to Retina.

        Head Mounted Displays introduce a lot of ergonomic issues and I think they’ll be niche. People won’t even wear Bluetooth headsets, I can’t see everyone trading in their phones for eyewear. Maybe as a supplement but even then I can’t see HMDs being that big.

        Some grist for the mill re: LiquidMetal:

      • KirkBurgess

        Head mounted displays will be the next killer app in gaming. The kinect glasses have already been leaked on microsofts Xbox future efforts.

        It might not sound that special, but imagine sitting around a table with some friends and playing games where instead of watching a TV you are looking at a 3D representation on the table top, which moves for you as you adjust your viewing angle. Kind of hard to explain, but basically it would look like a solid hologram overlayed on your table but would only exist on the transparent LCDs of the people wearing eyeglasses. To top it off the friends sitting around the table from you might actually be sitting on the other side of the planet, and they are also only viewable on your eyeglasses.

      • Very much agree. The Oculus Rift is making big waves in this space. HMDs for gaming is very much going to be huge.

      • 4K may make sense for TVs, where the screen can subtend a large angle at typical viewing distances. It increases the immersive effect, much like a movie theater does. However, in a handheld device at typical smartphone viewing distances, it will simply be heavy overkill and require more power and computation for no visual gain.

        Basically, for the pixels on a 4K display to be distinct, the screen would need to be twice as close as Apple’s current Retina displays, for the same display size. Or if the PPI is kept the same, the screen would have to be twice as large in both dimensions.

        If Apple does a 4K display, I’d expect to see it first on the 27″ iMac or a larger Apple TV product. I don’t think that resolution makes much sense on smaller screens. There’s a nice number-crunchy article here about what constitutes “Retina” resolution:

      • IGZO would mitigate the power issue. Theoretically, an iPhone with a “Retina+” IGZO display would have the same or slightly better battery life. Apple’s A6X processor already has the graphical muscle to run such a display. The A7 won’t break a sweat.

        However, I think the iPad is the best candidate for a 4K display. The iPad mini is said to be significantly cannibalizing sales of the big iPad. A Retina display on the iPad mini will only increase that effect. But a 4K large iPad may stem that. iPad also represents a much greater market than the display market.

      • But 4K on an iPad makes about as much sense as putting a 400hp gas engine in a Prius. You could do it technically, but it doesn’t do anything for your customer.

        If Apple puts IGZO in the big iPad, which I agree is likely, I expect the to exploit the power benefit by cutting weight via reducing the battery size, or by increasing battery life, or both. Both of those benefit the customer. Going to a display resolution that only matters if you’re holding it less than 8 inches from your eyes, does very little except give you bragging rights.

      • KirkBurgess

        I agree – the thing the iPad 10″ needs is a weight reduction – adding a higher res screen would make little difference to users (the retina screen is only 9 months old!)

      • vincent_rice

        4K on the iPad makes no sense whatsoever

      • I’m viewing it more from a strategic perspective. With the iPad mini cannibalizing sales of the iPad and a Retina mini likely to increase that trend, Apple is going to have to give people a compelling reason to buy big iPads. Believe it or not, 4K screens in the IPad’s size are probably the most ready to go into mass production and reach economies of scale quickly.

      • vincent_rice

        a) Apple doesn’t care that iPad minis are cannibalising iPad sales. Apple loves to cannibalise its own devices – it’s a central plank of their strategy
        b) 4K on an iPad brings absolutely no user benefit = pointless

      • I’m not a believer in the “PPI Myth” that has been pushed by opponents of Apple. In fact, I was able to prove the value of high pixel density in one post on the Verge that shot down the whole “PPI Myth” argument. PPI arguments are always based on the premise of a font at a standard size, but pixel density becomes a factor in relation to the smallest text elements on a screen, not the most standard.

        I don’t think pixel density becomes irrelevant until it matches print. The Retina display just touches the quality of print. A 4K iPad display would be the first display to firmly match print. So in that respect, I think a 4K display on an iPad is a big selling point. There are also very few ways to view 4K content. The iPad can, in a sense, become the first AFFORDABLE 4K TV. As 4K content becomes more popular, Apple would already have a device in the market capable of viewing it.

        As for cannibalizing, you still have to find a way to replace the lost revenues. I don’t think Apple will let the iPad mini cannibalize the iPad without considering how to prevent the total deterioration of the market for a device with substantially higher margins. 4K continues to make the big iPad both relevant and desirable.

      • vincent_rice

        So when exactly do you expect this 4K iPad to appear? Before a 4K desktop display?

      • That actually has me a bit stumped. Apple probably should introduce both a Retina iPad Mini and 4K iPad next year but I don’t think either is technically feasible in the short-term. But I can’t imagine what stop gaps it would use to stem cannibalization and encroachment by Android devices, which are evolving at a tremendous pace.

        I do think 4K displays will be available by Q4 next year.

      • The next step in resolution for the iPad would be @3x — 3072 x 2304 — 4K makes no sense in that the aspect ratio and the asset scaling would be off (though at sufficiently high densities, this last bit matters less). Some might believe that the next step in resolution is another doubling of the retina, but because developers still code for a 1024 x 768 point canvas, any multiple of that resolution would not require the introduction of a new target for them. And the bonus of @3x is that photos on the web that are @2x wouldn’t look all that bad, while photos taken or viewed with the device would show more detail.

        Would Apple change the aspect ratio? That’s possible, but I hope they don’t. Of course, I wish that they had gone to a 1080 x 720 4″ display on the iPhone, keeping the aspect ratio the same, even though it still required a new target for devs. (16 x 9 interfaces may come in handy for

        On the 4K front, I do think that it would be necessary for an iTube* (my completely hypothetical name for an AppleTV with integrated screen) greater than 40″ — not because of content, but because of UI ugliness. (And with a good scaling algorithm, 1080p could look more like film because of the fuzzy edges, which I have always noticed at the cinema.) Crispness and fidelity of high resolution photos, text and other vector objects matter everywhere.

        *Am I the only one that thinks of not just the Cathode Ray Tubes that were televisions for so long, but of a metaphorical tube delivering the world far away — like a pipeline, a wormhole?

      • tmay

        Apple’s retina Thunderbolt display will most likely be the first 4K display, and announced along side the next Mac Pro iteration. If my speculation is correct, it will re-imagine the Mac Pro as a Multiprocessor workstation built on SSD modules and Thunderbolt.

        I would argue that traditional internal space for four or five 3 1/2 inch hard drives will be sacrificed for a more petit package with a mix of 2 1/2 inch hard drives and SSD’s, and perhaps Apple’s first dedicated OpenCL accelerator. All catering to the target market of Final Cut X, Logic Pro and Adobe Creatives.

        Alternately, the Mac Pro will be the server of choice for creatives in small workgroups or organizations.

      • Very plausible. There’s going to be demand for a workstation to edit 4K content. A significant refresh of the Mac Pro w/ a 4K display makes a lot of sense.

      • Another place for a 4K displays is for TVs. Also, high-end displays used in video editing.

        The movie: The girl with the Dragon Tattoo, was the first movie developed from end-to-end (capture, edit, distribute, present) entirely in 4K.

        Some claim that this is the harbinger of the end of broadcast TV.

      • I might be crazy but I really think Apple is going to go with holographic display tech for its TVs. Solves the 3D glasses issue and has super wide viewing angles. For instance, everyone in a household could watch a 3D movie without glasses on one TV. It would also give Apple a unique advantage for its eco-system, such as special content. With Apple’s connection to Disney, there’s a lot of possibilities.

      • This is Star Trek technology, not something even available in any research lab I’ve heard of, except for maybe some simple proof-of-concept stuff. Holograms also require tremendous amounts of data, which would be a real impediment. 4K flat displays are bad enough in that respect….

        Even if the basic technology existed now in a lab somewhere, it would be too bleeding edge for Apple — they tend to jump on a technology only when it’s ready for prime-time.

      • The technology is definitely in its early stages but at least one company seems to have practical applications of it:

        I’d like to think that Apple might be much further ahead, especially if they’ve been working on it for years, like the iPad.

        The issue is that everyone is going to have 4K “SuperHD” displays in the near future. Apple won’t have much of a first mover advantage. Not to mention that the market for 4K TVs at the price at which they are expected to sell ($5000+) means that production is not likely to ramp up fast enough to make them affordable anytime soon.

        But a holographic TV would give Apple a monopoly. Even if the TVs were expensive, it is unlikely Apple’s competitors would be able to create competitive products any time in the near future. Apple already has applicable patents. I grant that it’s a stretch but I’m just not certain Apple will enter into a market that is going to get real crowded real soon.

      • Okay, it’s not actually holography, despite the company name. And it looks interesting. But looking at their very limited explanation of how they do it, I still have serious issues with the data rates required to generate anything more than simple images. Synthesizing the correct light field for every possible viewing position in front of the screen is a *lot* of work. There has to be some serious compromise in the system somewhere to alleviate that.

        I do believe Apple filed a patent on a somewhat simpler 3D technology a while ago, though, it may be related. I think their approach involved user location tracking, though.

        Since there’s very limited content that could drive a “true 3D” (as opposed to stereoscopic) display, I don’t expect Apple to even think about this space for a long time, though. About the only immediate application I can see would be 3D gaming, which would be really neat, but is a much more limited market.

      • ronin48

        Great point. And Apple has an exclusive and broad license for Liquidmetal technology. Competitors couldn’t easily copy them for a long time.

      • If that were the case, and Liquidmetal was about to happen in a big way, I think they’d buy the company. Its market cap is only $24M. It takes Apple less than five hours to make that much profit. If they sat down one afternoon to negotiate, by the time the deal was signed they’d already have more money than when they started.

      • ronin48

        They don’t need to buy it. They have a broad and exclusive license for Liquidmetal technology in consumer electronics. Look it up. That’s all they need. Apple doesn’t want to have to worry about supplying or licensing Liquidmetal watch bezels to Omega.

        So Liquidmetal – the company – is not about to happen in a big way no matter what Apple does with their tech because Apple already paid for the license.

      • The current agreement runs out in 14 months. Look it up. The risk of someone else buying the company is too great, so as soon as (and if) Liquidmetal becomes viable on a large scale you can bet they pull the trigger first. The fact that they haven’t is a good sign that there are big problems yet to be solved.

        The $50M-$100M or so it would cost to buy them, considering how much Apple spends per year, is essentially nothing. So there must be nothing on the horizon. Apple spends more than that on milling equipment.

      • When has Apple purchased any of its suppliers or tech partners? Why have the license if it doesn’t intend to use the tech? Your argument works both ways. If the technology has issues, then no competitor is likely to overcome them so why get exclusive control of the tech in the first place?

        Apple is a company that uses innovation to solve big problems. The rationale that it could just buy out LiquidMetal therefore LM tech must be a dead end is circular. On top of that, why RE-LICENSE the tech for an additional two years?

        Apple almost never acquires when it doesn’t have to. Hell, Apple can just buy HP outright and eliminate one of its largest competitors.

        There are definitely questions but to rule it out entirely seems unreasonable. I doubt Apple licensed a technology that could give it a significant advantage in the market for no reason. And just sitting on it makes just as little sense as the scenario you outlined.

      • This is not the same thing as a normal supplier. They are the only source of LM. I didn’t say it was a dead end. I said they aren’t yet close to figuring out a good use on a large scale. I didn’t say they’ve given up.

        It makes no sense at all to believe that Apple is on the verge of rolling out LM in its products, of which it sells many hundreds of millions, without controlling the *sole* source of LM.

        It’s just wishful thinking to believe that there’s some miracle material right around the corner that only Apple will be able to use.

      • The license is exclusive so, for all intents and purposes, it does control the sole source. The license is explicit in that only Apple can use the tech for electronic and computing products. No of Apple’s competitors could use LM even if they could figure out how.

      • Yehat

        “I don’t think Apple wants to get into the materials science business.”

        Why do you think so?

      • Too far removed from its core business. Working with a materials science company and running one are two different things.

      • “I don’t think Apple wants to get into the materials science business.”

        Seeing as how Apple always boasts about the materials / look & feel of their products, they kind of already are. I heard in a podcast once that Apple employs some of the best materials engineers in the world.

      • The original agreement expired in February 2013, and it was extended until February 2014. No reason to assume it won’t be further extended, when the 2014 expiration approaches.

      • Suppose Samsung buys them. How dumb would Apple feel, even with an agreement? Suppose Google, or Microsoft, or *anybody* else buys them? Those would be pretty good reasons to assume the deal wouldn’t be extended further.

        If Apple thinks Liquidmetal is important, rather than just highly speculative, they’d have already bought them out. It’s not a big check to write. The fact that they haven’t *screams* that Apple is not close to figuring out what to do with it.

        If there’s anything promising with LM it’s an automatic deal.

      • Those types of scenarios can be addressed in negotiation. As long as Apple can maintain control of the technology via license, it has no reason to get into a business which diverges so much from its core business. There’s really no reason for it to be “either, or.”

        I think Apple strive very hard not to become a conglomerate. It is difficult to manage businesses you may not fully understand. Tim Cook is a logistics specialist, not a scientist.

      • tmay

        I’m in agreement;

        Apple isn’t going to set up a research operation for
        materials, and that’s what LM is right now. It’s still quite likely that Apple has a number of options such as right of first refusal in event of a potential LM sale, and even then, would have ironclad licensing extension options.

        At the time of the original license, there was a single purpose built machine, and it was being used to turn out some production, but mostly as a testbed for customer projects. These machines would be very difficult to mass produce, certainly not on the scale of machining centers, and would require quite a bit of iteration even then to perfect the alloys. Witness the issues with the golf club heads shattering.

        Winemakers would feel more at home with a LM machine at this point in time than Apple’s production folks.

      • The reason why LiquidMetal is interesting to me is because it is the one investment Apple could make that would give it a substantial business advantage that also could not be replicated.

        At least theoretically, LiquidMetal production could feasibly allow Apple to significantly drop the prices of ALL of its products while protecting its margins.

      • We aren’t privy to the terms of the agreement, so what Apple is doing with Liquidmetal could be as ironclad as its outright purchase. If it is as important to Apple as we’ve been lead to believe, I can’t imagine Apple lets it go.

      • ronin48

        Once again, it’s not the price of buying Liquidmetal that matters. Apple doesn’t buy stuff it doesn’t need no matter how cheap.

        Also, Liquidmetal has licenses and contracts with government contractors and other consumer product companies that Apple has zero interest in dealing with.

        And to repeat, Apple already has a broad, exclusive, and renewable license for Liquidmetal technology. Apple has the option to renew. Look it up.

        So let Samsung or anyone else buy them – but it wouldn’t matter to Apple. Apple has the renewable license they want already. And if Apple can’t figure out what to do with the tech don’t expect Samsung to figure it out.

        Liquidmetal, the company, is not on the verge of anything unless they sell new licenses for new applications.

      • Do you have a reference for the “renewable” part of that deal? Saying “look it up” is just plain rude. It’s your reference, you cite it.

        Are you saying that Apple has the right to renew in perpetuity? Are you saying that, say, Samsung buying the company would have no effect? Are you suggesting that even if some of Apple’s billions were being spent ramping up for Liquidmetal that the company’s stock would stay flatlined at 11 cents?

        You seem to somehow be missing my point when you say”if Apple can’t figure out what to do with the tech don’t expect Samsung to figure it out.”

        I don’t see how I can state it yet again more clearly, as that is almost the opposite of what I’m saying. So I give up.

      • LM wouldn’t have much skin in the game. Outside of the license, it’s doubtful that it could secure other revenue from the association, particularly if Apple is making all of the key capital investments. So there really isn’t a reason for LM’s stock to reflect anything in relation to its association with Apple. If Apple is solving major issues related to scaling the tech, any innovations will belong to Apple, not LM. So LM probably won’t be able to use them to fundamentally improve its own business.

        It’s also possible that Apple has negotiated a strong deal contingent on it ramping up LM tech and it is keeping LM posted up. There are a lot of credible scenarios that could see LM ramped up, regardless of the state of Apple’s association with the company.

      • ronin48

        This isn’t a site for stock pumping. We’re sorry you’re taking a loss on LQMT but please take your comments to Seeking Alpha or The Street.

      • Well that’s just nasty and you ought to apologize. I have no position in LQMT, and if you read more carefully you’ll see that I don’t think anything is happening with it.

        Normally the discourse on Asymco is at a higher level. I was attempting some logical reasoning to show that any imminent developments with liquid metal were unlikely, but apparently I’ve really touched a nerve with some of the people who are hoping for some sort of miracle.

      • Yehat

        “I don’t think Apple is going to invest significantly in manufacturing and supply chain under these circumstances, particularly the intense competition with Samsung, unless it is going to derive a competitive benefit that its competition won’t easily be able to match.”

        Unless they invest in breakthrough manufacturing process.

      • Davel

        I want to thank all of you for a wide ranging and intelligent thread

    • neutrino23

      “could it be that Apple is spending on the “next big thing””

      Maybe, but unlikely. The amount they are spending is enormous. $500M a year could finance a lot of R&D developing new products they are spending many times that per quarter. This spending is so large it pretty much has to be production related. If production for a fantastic new product were ramping up in volume we’d hear something.

      • Correct. This is not R&D. This is capital equipment purchasing. The bulk of which is declared as Machinery, Equipment and Internal Use Software. Capital equipment is a depreciable asset.

  • nonfictionsection

    ” Apple did all these things and now it casts an eye over the next frontier: components”. Extra kudos for the dramatic and powerful ending.

  • Tim Cook loves to say that the brains of the iPhone/iPad are made in the US. How does he answer that question when they move away from Samsung to TSMC? Does AAPL build a foundry on US soil and have TSMC operate it?

  • KirkBurgess

    Could Qualcomm also be a casualty of Apples entry into component production?

    I get the feeling they might be….

    “@FOSSpatents: Qualcomm attacks Apple in ITC filing, calls it a patent infringer who should be embarrassed supports Samsung on #frand”

    • If Apple is going to try to integrate the baseband into their A-series SoC, yes, Qualcomm could be a big loser, in theory. But Apple would need to license a lot of Qualcomm patents to do this. As far as I recall, they own a lot of the early IP on CDMA, and were the cause of a number of protocol standards workarounds like Europe’s 2G GSM and China Mobile’s oddball TD-SCDMA 3G protocol.

      • Tatil_S

        What if they stick with GSM and LTE? It would require separating Verizon and AT&T phones, but 25 LTE bands sort of require that anyways. It is also possible for them to cooperate with Infineon (now owned by Intel). In light of recent rumors, that could be a package deal, use Intel fabs and Intel owned baseband IP.

      • Qualcomm is a major player in the LTE standards patents, according to various news reports I’ve seen. I believe the 3G standard is also heavily encumbered by Qualcomm patents (hence China’s weird 3G standard).

        I’m pretty sure if the protocol uses CDMA in some form, Qualcomm has patents that read on it. Or at least they think they do. That’s pretty much everything beyond 1G voice and 2G GSM. They’re an 800-pound gorilla in cell radio IP.

      • John

        According to iSuppli, the highest cost components in the iPhone 5 are (descending order):
        $44 display/touchscreen
        $34 wireless baseband/RF/PA [from Qualcomm]
        $33 mechanical/electromechanical
        $18 cameras
        $17 A6 processor
        $10 DRAM
        $10 Flash

        Now consider which of those Apple has brought strong customization to: I’m not an Apple expert (actually a wireless design engineer), but the $34 Qualcomm chipset MDM9615/RTR8600 seems like it’s a pretty expensive piece that’s out of Apple’s hands. Also, almost all of the “fast-followers” that Apple has patent fights with use the same (or same family) Qualcomm-provided chips. If Apple worked towards dropping Qualcomm (not a simple task, and they haven’t purchased a signal processing house to my knowledge in the same way they did PA Semi or Intrinsity), they would not only bring that customization in house, but they would put a major dent in the economies of scale that Qualcomm passes along to Apple’s competitors in the space. Put differently, Apple’s volume purchases from Qualcomm subsidize the fast-followers’ purchases of baseband/RF chips.

        As mentioned above there is a serious patent minefield around in-house development of baseband / RF chips, but Apple seems willing to face up to it, and if Apple were starting to show its hand to Qualcomm about future purchase contracts, Qualcomm would be the first to know and start fighting on the patent field.

      • I do wonder how much of the Qualcomm chip price is actually the license fees for the spectrum of GSM/CDMA/LTE patents it incorporates. Of course, I don’t think Qualcomm is giving people a particularly good deal (they don’t have to, they’re almost the only game in town these days)….

        Besides the IP issues, the other barrier that comes to mind is carrier approval. I’ve been assuming that the baseband and its radio and protocol code is pretty standard across all the phones that use it, which I would expect means easier carrier approval, since they’re always worried about rogue nodes messing up their networks. I’d assume a new design would have to go through really exhaustive testing with all the carriers. Not that Apple couldn’t do that, but it would eat a lot of time and effort.

        From an engineering sense, pulling the baseband into the SoC makes sense. I’m not sure it’s as good an idea from a development/management perspective.

      • OpenMinde

        Bob Mansfield heads a wireless/network division. And Qualcomm attacks Apple on SEP patents Tell tale sign of Apple would split from Qualcomm to avoid giving faster followers something to hold on. Also Apple partners with GlobalFoundries/IBM in upstate New York. Interesting time.

  • ronin48

    This should be easy to figure out by following the money.

    If a giant new chip equipment buyer (i.e. Apple) has suddenly emerged then we should see evidence reflected in or from KLA Tencor, Lam Research, Applied Materials, ASML Holdings, August Technology, Brooks Automation, Helix Technology, Novellus, Peter Wolters, Voumard Machines, and the like.

    If there’s no evidence, then Apple’s CapEx growth must be going towards something else.

  • Johnny Boy

    Didn’t Apple buy an Israeli based company specializing in DRAM or something? Could this be the source for the cap ex?

    • They designed flash memory controller chips. I assume their designs will go into Apple’s A-series processors at some point. Supposedly they could extend the life of flash significantly, which would allow cheaper/larger flash memories. This could at some point be a differentiator, but it shouldn’t be eating a lot of capex.

    • Anobit? It was capitalized at $76 million in venture money.

  • I think I’ve got it.
    Apple is doing the same trick with semincon just like it did with Foxxcon.
    They are investing on equipments (movable assets) only. (No immovable properties !)
    With Foxxcon, it was mainly CNC milling machines.
    With semicon, they must be CVD/MBE machines & steppers.
    It’s a win-win solution: Apple secures production capacity, and vendors can off-load depreciable assets from their books. (i.e. high asset turnover for vendors…)
    I also think Apple avoided owning a factory in China because they heard some horror stories….
    You really can’t own the land in China … if you get in trouble with the local government, you can be kick-out in ways that no one has expected….

  • Ittiam

    Excellent article Horace…

    Probably Apple is becoming the most vertically integrated company in the tech industry – from components to devices to cloud services….

    Samsung-Google together would also stretch end to end, but with one major drawback. Google’s wants to lower costs of computing to increase reach, which is in direct conflict with that of Samsung (and other hardware manufacturers).

    Google’s (and Samsung’s) long term problem is to manage this conflict in order to keep competing with Apple

  • SSShu

    When you mentioned the 10Billion$ number…this article just popped into my mind.

    Did they just spent 10B$ on – quite literally – a foxconn branded factory?

    • The article says the plan is for Foxconn to invest $5-10 billion over 5 to 10 years with investment to start in October.

  • Really interesting. If I was an editor at the WSJ or FT I would put a couple of my best reporters on this story and have them follow the money . . . spending billions each quarter would not be easy to hide.

  • Jeff B

    Isnt it because they are building out so many retail stores as well? Also, you have to measure capex against sales or gross margin dollar growth

    • Retail stores cost approximately 9% of CapEx.

      • Micromeme

        Hi Horace,

        Is there any chance they are considering maps+annotations a capital asset and spending properly to develop or add to it? to catch up with the google database and curate it properly would cost a lot (even if they were investing in it through a third party contractor). see

        and madrigals other reporting for some comments on scale of effort to curate maps + annotations

      • The operating expenses for a world-class mapping service can be obtained from Nokia’s reports on its (formerly Navteq) Location solutions. In Q3 2012 that division had operating expenses of EUR 175 million.

  • I guess this link could potentially explain..

    Most likely Global Foundries / IBM?

  • Pingback: De repente, a Apple passou a investir bilhões de dólares em projetos secretos | Tech Insider()

  • William Cox

    Horace, just thought you should know that Apple is indeed popping up with all sorts of plans to move things in-house. AMD and Qualcomm both have graphics divisions in Orlando, and now Apple now has its chip specialists from Intrinsity (Co-Designers with Samsung of the A4 chip) in Orlando, setting up shop for a new plant. They’re recruiting heavily from their competitors’ senior staff; I would look for Apple to abandon PowerVR as their GPU supplier in a couple of years and move everything in-house.

    • r.d

      complete BS.

      If PowerVR is the problem than ARM is also a problem.

      Just like It was previously speculated that Apple was hiring ATI

      executive. it turn out to be OpenCL and Retina Display.

    • Jonas

      Unlikely the GPU to be replaced. Apple already owns 10% of Imagination who make the graphics chip. Pulling their business from Imagination would wipe over $100 million from their investment. It costs Apple 10’s of cents to license the GPU which has given them a graphics leads over everybody. Apple look totally committed to the new powervr chips (Rogue) which will give them a massive boost in 2013 and get them retina on the mini and a thinner ipad 4. If anything Apple are working much closer with Imagination who now employ over 1000 staff. I would say the base band technology which costs 10’s of dollars is the target and is a much bigger strangle hold over Apple. Apple came late to 4G and I bet they remember that.






  • Bill Esbenshade

    Horace — I used to think Apple was ramping Cap Ex to possibly go to a six month product cycle for the iPhone. I now think it’s related to the challenges of ramping production for a low cost iPhone. I also think a six month cycle is unattractive from a design perspective, for the following reasons:

    Apple’s longer cycle times re the iPhone probably help prevent overserving products. Longer cycle times give customers time to absorb and acclimate to their existing iPhones. They give customers time to learn different phone features/functions. When they’re ready to buy a new phone, these customers are more ready to move up the product’s improvement trajectory.

    Longer cycle times are also consistent with good design. Dieter Rams noted that “imaginative design always develops in tandem with improving technology, and can never be an end in itself.” (Wikipedia).

    Longer cycle times ensure that meaningful technology improvements drive iPhone/iOS design changes and new iterations of the iPhone. Technology improvements lead to designs which enhance ease of use and product aesthetics. Improved ease of use helps prevent functionally overserving products. This approach also makes it easier for Apple to raise customer expectations of what’s functionally “good enough.”

    To justify a short product cycle, companies like Samsung saddle the latest iteration with new, pointless features and unnecessary design changes. They’re forced to do this because technology improvements to justify a new design and new product are insufficient. Poor design and too many features lead to a product that functionally overserves.

    Dieter Rams: “Good design emphasizes the usefulness of a product whilst disregarding anything that could detract from it.” (Wikipedia).