M is for Mystery

I recently tweeted that any discussion related to wearable technology needs to begin with a description of the job it would be hired to do. Without a reason for building a product, you are building it simply because you can.

The reason a product deserves to exist is that it can do a job that needs doing and that few, if any others can also do it. This happens when the job is unstated and difficult to perceive. Put another way, the difficulty behind jobs-to-be-done based design is that jobs are never plainly evident. In contradiction to invention, where the problem being solved must be as clearly stated as its solution, value-creating innovation meets new and unarticulated needs. Even when created, the value is more subtly perceived, often only after prolonged use.

Which brings me to the M7. Apple chose to highlight a component (curious in itself) which it bills as a “motion coprocessor”. It claims to be measuring motion data via accelerometer, compass and gyroscope and processing the information in some way. By bundling the sensors and their management into one integrated chip battery consumption is reduced and these motion monitoring functions are performed more efficiently.

But what for? The examples given during the presentation for an iPhone with M7 don’t add up to a lot of benefit. An iPhone could be used as a fitness companion but it would not a very good one. Compared with, for example, a Nike FuelBand, an iPhone could not track your activity well. It’s often not moving, sitting on a desk or in a purse or pocket while you are doing exercise. It’s too big to take into a basketball game. It can’t “observe” your activity because it’s not worn during many activities and if it is worn it is in a position which does not inform much about what you’re doing. Phones are too big to be used as physical activity monitors.

Hence the question of what is the M7’s job to be done. As part of an iPhone, it does not seem to have one. Saving a bit of battery life is not a job, and certainly not one that needs to have billing in a media special event. The answer must be that the M7 was developed for some other, as yet unstated reason.

When the A series chips were created Apple leveraged the in-house design and cost reduction to make a wide range of products with more than 700 million examples built. Designing a chip needs a broad application domain.

Perhaps this is why Apple chose to describe the iPhone 5s as “forward-thinking”. The M7 and the Touch ID are like research projects whose actual value will be realized at some future time, in probably different contexts. The M series of chips may become Apple’s “low end” microprocessor as the A series climbs the trajectory into core computing tasks (read: phone, tablets, TV, laptops).

M might be the chip for the wearable segment, woven into a whole new fabric of uses and jobs to be done.

  • David V.

    By bundling the sensors and their management into one integrated chip battery life is reduced […]

    Typo: Either “battery life is extended” or “battery drain is reduced” (or some equivalent wording).

  • David V.

    But what for? The examples given during the presentation for an iPhone with M7 don’t add up to a lot of benefit.

    Apple acquired WifiSLAM earlier this year. The following presentation by them suggests that M7 could be used for micro-location purposes:

    which would play well with iBeacon.

    It’s certain possible that it would be used for a wearable device too, but I think it does have a reason to exist even if iPhone-only.

  • Accent_Sweden

    One minor correction Horace: By bundling the sensors and their management into one integrated chip battery life is reduced and these motion monitoring functions are performed more efficiently.

    I believe this should read “battery life is extended” or “battery life is improved”.

  • Luis Alejandro Masanti

    What I love of Apple is how they have a master plan and hiddenly built it step by step.
    Everybody is speaking of “electronic payment.” Apple first delivered Passbook, all the needed parts except payment: they tryed the infrastructure.
    Now, they release Touch ID: this is (IMHO) a beta test of the fingerprint tech: just for secure your iPhone. Once tried on all its aspects, it can be part of a payment system.
    And now, M7. As you said, it can be the heart of the iWatch or of anything else, but by now it is just a building block of a Master Plan.
    (Albert Einstein said: “I want to know God’s thoughts; the rest are details.” I would like to know Apple’s Master Plan, the rest are building bricks.)

    • charly

      Payment system need market share to work. Less than 10% of the people have an Iphone in countries like the UK so passbook is death there. And the UK is the market with the highest percentage of people owning an Iphone outside America.

      Apple can only successfully introduce a payment system if they work with other companies but Apple is not known for working good with others.

      • sharrestom

        Payment system work better with customers that have disposable income than without, so marketshare may not have all that much impact at this point in time. The default payment system today seems to rely on NFC, and NFC has not and will never take off in the U.S.

        iBeacon plus hardware and software development partners will be rolling out BLE systems by Q1 2014, and some of the latest Android devices are BLE compatible. Then we will get an idea of success.

        BTW: I finally went to the effort of going through Disqus; previously I guest posted as tmay.

  • Janne Käki

    Could the M7 be the primary sensemaking unit for the yet-unreleased wearable Apple products, outsourced (via Bluetooth LE, perhaps) to another device (the iPhone) which has far more battery to drive it? Could it be that the iwatchamacallit will not include one at all?

    • Mark Howard

      I agree with this. The iWatch will not have this processor, it will be a companion device to the iPhone.

  • drkhrse

    I think I read on 9to5 that it some of the features of the M7 will be used for things like Maps integration. It can sense when you are driving and disable searching for wifi networks to save on battery life. Also detecting when you park and auto record the location of your car just in case you don’t remember where you parked.

    As far as the usefulness of the chip now, an app like Moves already demonstrates how much data can be gleaned from just the accelerometer data. The problem with using it now was that it killed my battery pretty quickly. I agree that the majority of the features of the 5s will be in future applications.

    • Janne Käki

      Moves is probably using not only the accelerometer (and other sensors available via Core Motion) but also Core Location, at least in the more lightweight forms such as monitoring significant location changes and geofencing.

  • Christian Peel

    I do not think we need to think about future products to see what job the M7 has:
    * Better location information (i.e. combining accelerometers, compass and GPS)
    * Always-on location info without swamping the battery
    * As a marketing tool to those who are thinking about fitness bands

    I personally hope that Apple puts a always-on heart-rate monitor into any watch they make. If it’s just the M7 I’ll be disappointed.

    • Walt French

      I glanced at a listing of GPS/runners’ watches recently for comparison. My guess is the watches are not much more sophisticated than the M7—they run a claimed 6–50 hours of GPS/logging on a button battery—except for the heart monitoring function.

      A couple came with (optional) electrode gel, the better to get a good pulse. I confidently predict that no Apple product will have a heart monitor if it involves smearing goo on your wrist.

      • Christian Peel

        I’m hoping for heart-rate monitoring along the lines of wriskwatch or better; I don’t think it requires a gel.

  • echotoall

    M is for milieu.

  • Diogo Pacheco

    Both Apple and Google are putting a lot of effort in the improvement of Software and Hardware of motion sensors. Like David said, micro-location / location inside buildings is one of the next steps to be filled and both companies are now creating mapping services on which several other services are being built.

    As an example, Google launched “Ingress” (Virtual reality location based build on top of GMaps) and the game is providing the data to support hardware changes and also is perfecting the way that all the sensors information is collected and merged to define location. When playing the game the consumption of energy skyrockets and normally, even with a new phone, the battery only lasts 3-4 hours. Here is a link to the Google IO 2013 where they announce the integration of all the data of the sensors to improve the location –

    By integrating everything in one single chip instead of doing that by software. (Something that is going to be difficult in the Android world where each manufacture as a different way to set up the motion/ location sensors) the M chip does seem to a right set toward wearable technology (Just need to look at the battery life of Samsung Gear).

  • M7 is billed as a coprocessor. Not a processor. IMHO, it serves two purposes. Allow the handset to still have a decent battery life by working in a strict power budget and augment the capacity of the device to tackle specific tasks. The A7 is quite similar, in that prospect, to the X8 found in the Moto X.

    • Kenton Douglas

      It’s the same. The Moto X also does this with voice.

  • dancharvey

    The original Nike+ was predicated on the observation that runners ran to music. That they always had their iPods with them. Are we so sure iPhones are left behind?

  • peter

    Thinking about wearable technology and its uses; so far I have listed: telling time, finding items with GPS tags, GPS directions while out walking/hiking/shopping, pager messaging, ultra-local weather (rain in 10 minutes type), voice search (nearest coffee place), train platform & gate announcements, bar code scan while shopping, Facebook check in of location, fitness, and small payments.

    Some of these things are pretty nerdy, but one or two might make the whole thing worthwhile. Many of the items on my list would require very accurate in-door GPS positioning, which is something the M7 could do.

    If the M7 is the wearable computing (co-)processor then the absence of one in the next generation iPad would be a clear indicator.

  • anon_coward

    the M7 is optimized to do a specific kind of math. my old iphone 4s when i used navigon or waze it would get very warm, probably because the math of doing all the GPS calculations was intense. the M7 is optimized for this and will result in better battery life.

    the magic of apple is that they will put this into their SDK and apps utilizing it will come out pretty fast. where with android you will have to wait for years since every phone will have a different motion chip like in the moto x and s4.

    • peter

      “will have to wait for years” — The adoption of new technologies in the PC market (3.5″, USB, CD/DVD, DVI, etc) should give you an indication of the speed at which un-integrated hardware/software manufacturers can move (and the trouble that exists in agreeing on standards between commercial entities, see Coase).

      Samsung and other ambitious manufacturers must feel the pressure from not ‘owning’ the operating system when it comes to integrating new technology; 64-bit, motion chips, or otherwise.

      • Kenton Douglas

        They don’t need to ‘own’ the OS to integrate new technology.

      • peter

        True, but software developers might be encouraged to use snazzy new features if there were a standard API in the OS. Failing that, apps that use the new technology will be few and far in between for too long a time.

      • Kenton Douglas

        True, but there is a outline framework API available (, but the vagaries of the differing hardware implementations is the real issue here (and to Android ‘fragmentation’ overall)

    • charly

      Samsung, who has a larger market share than Apple, can and will do the same.

      • paperbirch

        We have seen data showing Android with a larger market share than Apple. It would help facilitate the discussion to provide sourcing for your statement on Samsung having a larger market share than Apple. It would also help to know if you have knowledge of Samsung’s internal roadmap to support your assertion that “Samsung ………… can and will do the same”.

  • markwilcox

    It’s probably a bit early for informed speculation since we don’t really know what the M7 can do. However, in a world of wearable computing where very small, cheap, dumb wearable devices continuously transmit data to your smartphone, having a low-power chip that can process and store that data in large batches, so the main CPU has to wake up much less often (maybe only when there’s something interesting in the data) is going to have a massive impact on power consumption. It’s not just saving a bit of battery life – it could easily be the difference between some wearable concepts being viable or not.

    FWIW, Apple are probably also seeing strong growth in sleep monitoring apps – normally devices get to be mostly off at night and Apple mentioned power optimisations they’re doing in that area, having them poll the network less frequently – having the main processor waking up regularly all night to monitor the sensors is a battery killer.

    I agree that none of this qualifies for billing in a media special event, but what about the 5S does? 64-bit doesn’t seem like the sort of thing that Apple would normally trumpet either, particularly when it brings very little direct benefit to the device – it’s twice as fast as the last one is all they usually say. My theory is that the 5S has two markets – those that love the tech and want the best phone on the market (knowing about 64-bit and motion co-processors as reasons why their phone is technically superior is important for this group) and those that want to be seen with the latest greatest device (i.e. status/snobbery value – replacing the 5 with the 5C and only having a metallic body at the very top end is key here).

    Not saying that the M7 might not be the brains of future wearables, just that doesn’t seem like a great reason to call it out as a “co-processor” in the main iPhone media event either.

    P.S. “Phones are too big to be used as physical activity monitors.” – I thought this and because of it dismissed an app concept that syncs your music beat to your exercise tempo (has been shown to reduce perceived exertion) about 6 years ago. Now almost every time I go out I see someone running with a gigantic smartphone strapped to their arm!

    • anon_coward

      runkeeper and other apps will map your run and will be optimized for M7 very soon. Nike+ seems to be integrated into the 5S already along with some of the accessories that Nike sells.

      i’ve run with the 5 and if you lock the screen the run route mapping gets screwed up. unlocking means you waste battery and maybe miss a text or an email.

      • johnkeippel

        That’s odd. I lock mine all the time and the route mapping only ever gets screwed up when I’m in downtown Chicago. Along the lake it’s nearly always spot on.

  • pesc

    Wild speculation:

    Apple has its best disruptions when they add new I/O methods; mouse, touch.

    The M7 in a iWatch will allow the wearer to command a device (Apple TV) with gestures. The M7 will continuously monitor the iWatch and will know when the movement is a gesture.

    Science Fiction Speculation: The true innovation in the iWatch will be the sensors in the watch band which will be able to pick up any finger movements. Allowing for a rich gesture vocabulary.

    • Insider

      Why is everybody is so obsessed with the watch like in wrist watch? A wrist watch or arm band like device is a niche not mainstream. The facts show that fewer people wear watches since the mobile phones came. So unless you really need that kind of application, a wrist band or watch is an annoyance just because it’s an extra device that you need to carry – unless you really enjoy having something tied to your hand all the time.
      Since you already have the phone in your hand (presumably) you are suggesting that you need to to move your other hand to command the TV. Am I going to have to buy a wrist device just to command my TV? You gotta be kidding … I personally find that one finger on the Apple remote app is a good enough user experience. Having to move my wrist or the whole arm just to change a channel. Why?…

      • charly

        Less people use a watch because they use another time giving apparatus, aka the mobile phone. But the smart watch purpose is to obsolete the mobile phone

      • Insider

        Why obsolete the mobile phone?
        I can see the watch as a replacement for a phone if:
        – it works as a standalone device and it doesn’t doesn’t need to connect, draw power or tether with something else.
        – it offers a reasonable user friendly data entry/input method such as speech.
        – the reading experience / screen size is at least as good as a phone
        – it beams very little radiation onto your body, much less than a cell phone does. Let’s not forget that until proven safe all these radio emitting devices are potentially harmful.
        The last two points are unattainable with today’s technology so let’s talk about stuff that can be done rather than things that are from the scifi books.

      • charly

        Start with replying to Erik.S

        Computers are no replacement for typewriters. Computer + printer are a replacement for a typewriter.

        Smart watches will never be better than a a smart phone as their screen is way to small. But a smart watch + cheap tethered 5″ tablet is better than a smart phone. Losing a watch is much more difficult than losing a 5″ smart phone and losing your smart phone such a bad experience.

        – It will work as a standalone device as well as a how a computer replaces a typewriter without a printer in 1985

        – It does. It is called your smartglasses, 4″/5″ pocket tablet, 7″ couch tablet, 20″ screen + keyboard office or your 50″ screen + controller couch. And in the rare situation your without any of them you could use the watch itself.

        – The reading experience of my smart phone is just as good as on my laptop. But i could be lying now? Besides most of your time the watch will output its interface to another screen. Same as smart phones will do in 3 years time.

        – Communication infrastructure isn’t smart enough yet but that has nothing to do with technology

      • Ridiculous. We live in an era where computing hardware is cheap. The computer had to obsolete the typewriter because they were both expensive, and the computer even more so. The smartphone is obsoleting the portable media player because there is little reason to carry both.

        It seems unlikely that a smart watch will ever be good enough at everything people love their smartphones for that it will obsolete the mobile phone. As I noted above, watches struggle to provide a reasonable user experience for anything more than checking the time and date. The idea that they will, on their own, be replacing smartphones is laughable, particularly when the $$ savings might, at most, be few hundred, and is just likely to be negative, because smaller devices tend to be more expensive than larger devices with similar capabilities.

      • claimchowder

        Gives Moore’s law there is reason to believe that a smartwatch MOST CERTAINLY WILL offer a good enough experience if either of the following becomes true:
        – it will use a new I/O method. Input could be voice, output is hard to get right without a decent screen, so this seems unlikely
        – it can be bound to any of those (eventually) ubiquitous screens popping up in living rooms, airplane seats, cars, at bus stops, at Starbucks’ etc. by a simple gesture, and that screen can then be used like any iOS screen is used today.
        Slightly bump your watch onto the screen at Starbucks and watch your favorite podcast…

        I personally think the latter is what it will turn out to be, eventually. The problem will be the transition period where there are few screens.

        BTW you would pay for using the screens by watching an ad or by authenticating by TouchID and being charged on your AppleID.

  • Nathan Stevens

    I feel that it enables a huge amount of location/motion data to be crowdsourced and harvested without killing the battery life. Crowdsourced info that requires no interaction from the crowd beyond carrying the device around as they would anyway. The WifiSlam video indicates that the mapping is being built and improving in realtime as users wonder around in a building. Basically any device with an M7 chip will be a worker bee happily working away to help improve the maps and make the experience better everyday.

  • r.d

    MEMS chip was already discrete.
    What Apple has done is licensed the tech
    And branded it. only new thing is that it can work
    Independent of the processor.

    Apple has in fact licensed the GPU and is calling
    Apple GPU. So some custom stuff is going on.

    In fact ISP is also licensed from the GPU vendor
    And most likely the H265 module as well.

    • Jessica Darko

      The M co-processor is more than just a “MEMS chip”. It is a CPU that takes data from an accelerometer (probably MEMS based but certainly a seperate chip since the process for making MEMS is not compatible with the A7 process where the M7 lives.) and combines it with GPS, compass and gyro data.

      • r.d

        M7 has three things: compass, accelerometer and gyro.
        So that means MEMs chip.
        No where did I say anything about GPS.

        If M7 was part of CPU the Apple would have described like
        They described ISP chip being part of CPU.

  • poke

    I’ve been thinking about the job to be done for an iWatch. I think it would be expanding the job of a regular watch. A regular watch gives you context-sensitive information (the time) at a glance. The job to be done for an iWatch is to be always with me and to tell me what I need to know here and now at a glance. Any technology relevant to ascertaining context would be useful to it. The name “watch” and the language we use for describing watches (they “tell” the time) is already perfectly suited to this category of device.

    It needs to be a “low commitment” device. People are willing to put things on their wrists but only if they can put it on in the morning and forget about it. So it needs to be small, light, robust, water-proof (or at least splash-proof), have a long battery life and be easy to recharge. The UI has to be as simple as possible (it might not even use touch). It would be a device you have a different relationship with compared to a smartphone or tablet; something you don’t have to keep track of, don’t have to think about whether to take with you, etc. As you suggest, it can always be with you, which brings with it new possibilities (this suggests that it won’t merely be an accessory that connects to your iPhone/iPad).

    This gives me three expectations: (1) technological-limitations mean the first version will likely trade features for getting the form factor right, since it’s the form factor that makes the device; it won’t do much, it will disappoint the usual suspects; (2) it won’t be released until Apple can get the right mix of features and form factor to make it worthwhile; and (3) it’s eventually going to be a big deal.

    Might we also see the M7 in the remote for the new Apple TV?

    • Jessica Darko

      Consider also that this watch may be a remote control for your environment– an audio remote for your iPod, a remote for the AppleTV, a satellite interface for FaceTime as well as provide it’s own (probably simple) apps.

      • That sounds kind of horrible. Watches struggle to provide a usable UX for anything but checking the time and date. I can imagine ways Apple could do a better job, but there is a big gap between better, and good enough for Apple.

      • dmx

        yea, when introducing the original iPhone , SJ singled out mouse and the scroll wheel, and then multi touch as new user interface that enabled new product category. I just don’t think iWatch could thrive with a touch interface, nor Siri (too intrusive to the env surround you), nor Google Glass. Some kind of UI interface breakthrough is needed to have a truly new product category.

      • peter

        I agree. The 2011 iPod nano (little square) has a touch interface that is already tricky to use. You would need a simpler touch model to use on a watch, perhaps aided by voice and air gestures but those are power intensive. This will require some thinking and testing to get right.

  • Tatil_S

    Although tracking my path while I am hiking without draining the battery is a nice feature that I might use every now and then, it is not that big a deal and I doubt it is something many people are craving for. If there are other uses that Apple has in mind, it should release the apps along with he hardware rather than leaving it at some speculative, maybe for this, maybe for that description. Otherwise, it ain’t so different than Motorola rushing to launch an Android handset with a front camera, to beat iPhone 4 to the market by a few weeks, without a compatible video chat app available for it. If owning the hardware and software stacks are so valuable, we should see it in action. No consolation prize for half hearted efforts. There are plenty of poorly thought out hardware features available from various vendors at much lower price points.

    • Jessica Darko

      You used the example of the front facing camera, and unlike Motorola, Apple had a key software application lined up for their front facing camera- Facetime.

      The situation is the same for the M7 co-processor, likely. Tracking your path while hiking alone is not compelling, but tracking your activity, all the time, however, is… at least it appears to be given the rapid proliferation of devices that are designed to do just that.

      • Tatil_S

        When iP4 came out with a front facing camera, Apple also introduced FaceTime and made it one of its selling points. When GPU capabilities of 5s were brought up, a launch partner demoed Infinite Blade 3 to showcase its capabilities in game play. When M7 was announced, well, nothing really. “We hope somebody will make some use out of it” does not quite fit Apple’s MO when it adds new hardware features.

      • same

        “When GPU capabilities of 5s were brought up, a launch partner demoed Infinite Blade 3 to showcase its capabilities in game play.”

        So this counts, but basically the same thing for the M7 with the Nike+ Motion app doesn’t.

      • Tatil_S

        Wasn’t that just a screenshot? I wouldn’t call that driving home the message.

      • Tatil_S

        Comparing a simple screenshot to the demo of an actual game play on stage and calling it “basically the same thing” is foolish. Pricing? Can it tell whether I am biking or riding the bus? Can it estimate calories based on the terrain and activity (biking vs. hiking, flat vs. uphill)? Can it show my path overlaid on a map? Does it matter when it is in my side pocket vs. on the back of a biker jersey? How does the UI work?

        Yeah, a simple screenshot does not cut it. That is no different than Moto telling people to download an app for a video chat (“Give it a few months there will be apps that will start working reliably, some may even look nice”) or PC companies adding GPS to a laptop (“We are not quite sure how anybody would utilize this, but we are confident somebody will come up with a useful program one day.”) Feature matrix engineering…

  • Jessica Darko

    Horace, you’re operating under several misunderstandings.

    First off, while they showed a chip marked with a bit M, the M coprocessor is not a seperate chip. It’s one of the many, rapidly multiplying, co-processors in the A7 SoC. These provide differentiation and improve battery life because they can be handling specialized tasks while the main processor cores sleep. Thus the M coprocessor might be on a watch or other device, but would not power it. It merely does the telemetry.

    The job it’s hired to do is to track telemetry full time. Telemetry is a combination of a variety of sensors, gyro, compass, gps, and probably several others. This is not computationally expensive for a full ARM core, but it is something that is worth doing seperately so the arm core can be powered down.

    Previously, apps were limited in their frequency and readings of the telemetry from these sensors due to the power requirements. Or put another way, Core Motion was hobbled- only really available when the cores were powered up. Now they are available all the time.

    for people who keep the device in their pocket all day long, then it’s perfectly fine for tracking activity. (A phone in your breast pocket with good software is just as good as a band on your wrist for tracking activities.)

    The M-Coprocessor enables this full time telemetry use. This is a job that the iPhone couldn’t really do before. IT could do it some times and the app developer had to figure out the details and integrate all the signals, but now this is really there all the time.

    They are “forward looking” features, as Core Motion and Touch ID will be much more valuable when all devices have them, just as the forward facing camera is much more valuable now that all devices have them…. but as an added cost they start out on the high end and work their way down the line (same thing happened with the retina display).

    But I wouldn’t call them experimental.

    this is just how you solve the chicken and egg problem- you introduce it on the high end which can bear the price, and that gets it out there for app developers to experiment with (The apps are experimental) …such that as the feature spreads across the line the apps to make use of it are already there.

    • sharrestom

      I’m of the impression from various discussions elsewhere, and from the nature of MEMS, to believe that the M7 is a separate die on the SOC, not integrated on the A7 die. It would be therefore easy to repurpose that same die in another SOC or even as a separate package.

  • Mark Howard

    The M7 was never described as accessing GPS information since that is battery intensive, so I see limited value in terms of mapping applications, indoor or out.

    The M7 tells the device the context the user is in; it’s job to be done is for the device, not the owner. The device then make relevant decisions based on the owner’s state and context. Optimizing power is an obvious one. Are there others?

    • Rene Stein

      Inertial navigation has been used for a longtime, rockets, space craft, missiles, bombs. Their arcing trajectories make it much easier to do, but the processing on a smartphone is now pretty high. So, this, along with iBeacons can be used to track motion and location.

  • Jessica Darko

    Consider this: Apple did a refresh of the AppleTV early this year with a bespoke single-core A5 processor. Why do a custom silicon spin for a low volume device like the Apple TV?

    There’s no reason Apple couldn’t have also put an M7 and other co-processors on that CPU. They don’t need them for the Apple TV that doesn’t move… but if they are doing a watch then suddenly they have a high volume product that will make use of it.

    Possibly they are doing a revision of the Apple TV that will come out this fall that will be high volume as well (But only require a single core?)

    Anyway, I wonder what the package size is on that A5. It might be small, small enough for a watch. A watch wouldn’t need an A6 or A7. But it would need an M co-processor.

  • santoscork

    The M7 can certainly be used to create entropy when performing encryption keys. I can’t verify this claim; however I can offer the following, when generating a set of keys for encryption purposes, continued use of the computer in the form of mouse and keyboard input and other interaction serves to add entropy, helpful in the delivery of good encryption keys. This does not discount your interpretation of how the M7 can play a role.

    I found this keynote of particular interest and markedly different from any Jobs or post Jobs keynote. Apple took the liberty of mentioning the leaked photos during the keynote and used it as an opportunity to dismiss them as much as use them. After Apple identified obvious knowledge of leaked product, they showed the products themselves in all their glory. In essence, this almost ignores leaks as much as it does dismiss them. In a sense, Apple leverages this ‘unofficial delegate’ as a stream of buzz prior to any keynote. It suggests that the leaks fail to encapsulate a grand vision, the heart of this particular keynote.

    Two deeply relevant technologies have surfaced, consolidation of sensors into one chip, an M7 and Apple’s Touch ID. These are two very deliberate and causal technologies that I suspect were under a considerable development curve with deep regard for a future roadmap. I would argue that all the other material Apple presented are several factors less significant because they are about immediacy. The M7 and Apple’s Touch ID present engines that invite innovation.

    • neutrino23

      I recall that Steve referred to the Time Canada(?) cover leaking the lamp style iMac a day or so before the keynote. Probably there were other examples.

  • It isn’t clear to me that the M7 is actually a separate chip, or whether it is a block on the A7.

    Assuming it is the former, while I think it is certainly plausible that the M7 heralds the arrival of a new form-factor in a new “job” category from Apple (just as it is plausible that the move to 64-bit heralds the imminent arrival of yet another device category), I think it is worth considering more mundane explanations for the M7, just as there are mundane reasons for the A7 being 64-bit.

    First, I’d suggest considering that Apple didn’t design a whole new chip, and instead are buying a commodity design or minor variation, just as they did with the SoC for the first few iPhone generations. There are dozens, if not hundreds of cheap relatively powerful but low-power chips that could work in this application that Apple could have selected from.

    Second, consider that the fitbit/fuelband devices are already overserving their core market, and creating the opportunity to a worse but better solution, such as one incorporated into a popular smartphone platform. Hikers, walkers, cyclists could all be quite happy with a smartphone based solution, provided it didn’t run down their battery. That market may be broader still if you consider the market supports multiple options for carrying and even using one’s smartphone while running, or some other high-activity sports.

    Third, this functionality can enhance existing and planned iOS features. It could augment GPS positioning while outdoors or driving. It could enable inertial location tracking when indoors, based after the last available satellite fix or from an iBeacon-device. In a related way, it could support other contextual scenarios, like having your Passbook ready when you pull your phone out of your pocket at the airport gate, or Starbucks, etc.

    Now, arguably many of these scenarios are also applicable for a new class of device, but often as a compliment or alternative to iPhone in the same scenarios, not as a replacement.

    • Rene Stein

      It seems that it would be a separate chip. When they displayed it, they displayed it as a separate chip. When they talked about keeping the finger print sensor data on the A7 itself, they highlighted the particular section of the A7. Also, Apple designs the chips themselves when it will benefit them. I don’t think it is possible to tell either way who designed the chip.

      • Jessica Darko

        Apple uses the “chip” image to imply a co-processor. They’ve used it in the past for other processors that were integrated in the Ax SoC.

  • It is just the beginning, not the end.

  • Walt French

    I’m going to give Apple the benefit of the doubt—that it is oh-so-aware that its business is all about services accessible thru its devices. These services need the hardware, software and even external networks to be in place before they can be offered.

    Take a Waze-type driving service or better maps service. Apple has said for years that it was collecting users’ location data for a future driving service. Cell-tower-based, wifi-based and even GPS-based location information isn’t really quite precise enough or reliable enough for several useful features; rapid updates from the phone’s accelerometers and gyro are necessary. But *rapid* is key; checking them only every half second or so will miss most of the useful info and the accuracy suffers severely. Any Waze user will tell you how hot their phone gets, and every maps user knows occasionally they go off where they can’t possibly be. These services aren’t good enough yet, due to not being able to afford more constant updates.

    Look up “Kalman filtering” for a taste of the complexity of updating guidance/location info. Complexity means battery power, so high-resolution updates are very expensive.

    Indoor location is important for Passport and individual stores’ apps—Nike runners’ watches on sale on aisle 7 just to your left! This relies on rapid-fire updates to that inertial/internal information even more than driving info because when walking you generate MUCH noisier acceleration/turn signals than when driving. The M7 can log every little bump & twist for very little power (runners’ watches get many hours out of a button battery) and let the A7 actually process the data more in efficient batches.

    Several iOS developers are downright excited about better location data and having the system able to monitor it in background.

    Many more functions of the phone than I first realized benefit from 64-bit math. Most signal processing, OCR, possibly those snazzy new camera features, all get faster; one dev told me his OCR could be better from having enough speed to tackle more accurate algorithms. I have guessed that the announced speedup is mostly that the CPU now can get data from memory 64 bits at a time instead of 32. (“Bitness” is usually defined by the size of the address space, and I project it’ll likely be 2015 or later before the 32-bit addresses cramp apps’ style. Maybe it’ll come sooner on the iPad or some more MacBook-like device.)

    Low-power Bluetooth is also a part of the location mix—and part of the iPhone since the 4s. In-store, in-office, in-museum BLE beacons can help calibrate locations to provide sales or multi-media experiences.

    It’s very unusual for Apple to tout hardware that has no user-observable benefit, so I take this as a positive sign for Apple’s CPU group and for the OS software group having been able to crank out major upgrades faster than expected. Now that the foundations are in place, and once enough phones show up in the wild, we should expect a Monster flood of apps and services.

    • Insider

      This is very good analysis.
      Everybody seem to have forgot about IOS in the car.
      M7 will allow IOS devices to turn even more into real-time devices and control systems. If the device is plugged in, then power savings are not as important as the ability of allowing the main CPU to centralize the data coming from the sensors and most importantly – react in real time without being overwhelmed by the floods of data coming in.
      Now – the sensors could be internal or external to the device. In the context of a car IOS could take over the control of the car based on the information coming from the sensors mounted on the vehicle. Picture cameras with built-in recognition features, proximity, distance sensors, lasers, etc. This is just a small snapshot on the things to come. That may be the FORWARD thinking.
      And iWatch: I’m watching my phone driving my car.

    • Rene Stein

      I completely agree with this. Interior navigation/better navigation with this and iBeacons is going to be one of the driving factors behind this. There are a lot of possibilities for this chip in the future. Apple isn’t showing their hand yet, but they are showing some of the cards.

  • Lun Esex

    >But what for? The examples given during the presentation for an iPhone with M7 don’t add up to a lot of benefit.

    From Apple’s iPhone 5s product page (

    “M7 knows when you’re walking, running, or even driving. For example, Maps switches from driving to walking turn-by-turn navigation if, say, you park and continue on foot. Since M7 can tell when you’re in a moving vehicle, iPhone 5s won’t ask you to join Wi-Fi networks you pass by. And if your phone hasn’t moved for a while, like when you’re asleep, M7 reduces network pinging to spare your battery.”

    The accompanying image shows the Nike+ app.

    The original iPhone included an accelerometer, which was highlighted during its introduction. Its “only” use at the time was for sensing orientation and automatically reorienting the screen.

    The iPhone 3G included a GPS chip, which was highlighted during its introduction. Its “only” use at the time was for more accurate positioning in the Maps application over cell tower triangulation and WiFi hotspot positioning.

    The iPhone 3GS included a digital compass, which was highlighted during its introduction. Its “only” use at the time was for the included Compass app, and reorienting the view in the Maps app.

    The iPhone 4 included a front-facing camera, which was highlighted during its introduction. Its “only” use at the time was for the included FaceTime feature.

    All of these things had simple, mostly singular purposes when they were introduced, and their usefulness has been expanded upon since then. Apple could have included their own in-house motion-tracking app with the iPhone 5s, but instead they partnered with Nike to show the Nike+ app, which is already mature and more full-featured than a brand-new Apple-developed app could be. This is similar to how Apple partnered with Yahoo for stocks and weather data and Google for Maps, YouTube, and web search in the original iPhone.

    All of these things were included for forward thinking reasons.

    One part of the forwarding thinking involved with including the M7 on the iPhone 5s is that it’s the initial platform for CoreMotion. This allows developers to become acquainted with it, and perhaps discover innovative uses that Apple themselves have not yet thought of. When future products include the M7 and CoreMotion the developers will already be familiar with it and they’ll be able to get apps ready to take advantage of it that much sooner. Of course one of these future products could be a smartwatch, but another one that I haven’t seen anyone else mention here yet is a motion controller for the Apple TV.

    It’s also not hard to believe that the M7 will be included in future iPod nano and iPod touch revisions. It may be, and in fact I’d say it’s likely, that the M7 has its own low-power ARM core running inside of it. This could be enough to become the sole SOC used in the next revision of the iPod nano. It could also finally enable the nano to use a version of iOS as its core OS, and potentially open up the nano to 3rd party apps. The nano already includes Nike+ support, so at the very least this might reduce the number of chips or complexity needed for integrating that.

    Finally, if Apple does do a smartwatch, having an M7 in both your iPhone and your watch will mean more coverage for motion tracking. There might be times when you only have one or the other on your body at a given moment. When you’re ready to view or process the motion data the two devices can exchange their data to build a more complete picture.

    • Lun Esex

      To pre-empt speculation that the M7 isn’t a separate chip, but is included in the main A7 CPU: Maybe, but that’s less flexible. It should be easier to integrate with other devices and other processors as a separate chip. For instance, next year’s C-class iPhone might keep an A6 processor, at a slightly higher speed, and be paired with a discrete M7.

      As Horace speculated and I also mentioned above, the M7 could have enough power on its own to be the sole SOC for a number of devices like the iPod nano, a smartwatch, a motion controller for the Apple TV, and maybe even the iPod shuffle.

      Considering that the A7 is touted as having 40x the processing power of the CPU in the original iPhone, the M7 alone as a single chip could easily have at least as much power as that original processor, in a package that’s a bare fraction of the size and power consumption. This could give plenty of capabilities to other cheap devices and accessories.

      Only time (and teardowns) will tell whether the M7 is actually in a discrete package or is integrated into the A7’s package.

    • Lun Esex

      ALSO: The fingerprint scanner on the iPhone 5s has a very limited number of initial uses, too. It can unlock the phone, and allow you to make iTunes Store purchases. Why not more, like integrating it into a mobile payments solution, and/or allowing 3rd parties access to it?

      Like additional functions for the M7, most likely those things will come. Apple is methodical and deliberate.

  • henry3dogg

    This is about one thing. It’s about battery life.

    Battery life is about keeping the processor in deep sleep for as much of the time as possible.

    Remember in the wwdc13 keynote, they talked about changing the process scheduler to group process wakeup to produce a smaller number of wake ups, because each wake, even for involves a significant power drain overhead beyond the work actually done.

    Then think about heath monitoring apps that want to monitor your physical activity all day. If you use the processor nearly continuously, then your battery will be dead in no time. And if you only wake up rarely, then the analysis will be very crude.

    So what you want is a very very low power processor designed just to integrate the outputs of 7 sensors, log them, and raise an interrupt when certain conditions are met in the output.

    So yes, the M7 is important.

  • Z

    “M might be the chip for the wearable segment, woven into a whole new fabric of uses and jobs to be done.”

    I think so too.

    Both the M chip and the Touch ID scanner look like they might be ideal for the rumored iWatch project.

  • peto1

    Horace, you took the words right out of my mouth. The moment I saw the M7 (and in particular how much smaller it was by comparison with the A7) and heard what it did, I thought to myself, there it is – that’s the chip that will drive the iWatch or whatever other wearable computing device Apple comes up with. Putting it in the 5s first allows them to test the technology on a very large scale. I love it …

    • pk_de_cville

      I can one up you on this:

      What if the M7 is Apple’s Trojan Horse into wearables…

      Apple’s plan might be to ‘open source’ these babies to any wearable team that designs it into their product. They buy an M7 with M7 SDK for $50. And much of the work is done for them; they need to only build an excellent piece of wearable hw that people want because it reflects their sense of personal style and the particular jobs they want done.

      M is for “All your wearables are MINE.”

      Even more evil…

      Is the M7 actually Intel’s just announced Quark chip?

  • Great discussion and it def seems plausible over the mid-term. Immediately, this move, architecturally, of unbundling some processing for total system power optimizations is a trend that will continue, as analog and mixed signal circuits drive the need for local pre-processing. That Apple chose to brand it is relevant in the short-term, the M7 moniker’s marketing value is non-trivial, and supports Apple’s “hardware + software” messaging that give its platform and approach its differentiation, and edge.

    The M7 contains 9-axis motion sensors that the iPhones already contained. Power optimizations (a very expected incremental innovation, architecturally) aside, the M7 seems a lot (no knowledge here) like an Apple badging of the Invensense MPU-9×50 with probably some Apple specific code running in the “MPU”. I’m not aware that Apple has this kind of MEMS IP in-house, so at best it could also be a SiP using other partners’ silicon for the MEMS tech, and perhaps a little deep sub-micron uP from Apple’s A-chip team.

    So here’s the marketing angle… Even before the M7 was announced, I was thinking that the new core motion frameworks in iOS7 give a new level of application-level data services to 3rd party apps, and that this would spawn a new generation of apps that offer novel integration of user motion data (steps, states of walking, running, driving, etc). Marketing the M7 to the mass consumer as “feature-tech” will allow Apple and the iOS platform to be accredited with more of the innovation delivered by those third party apps. And Apple may be finding it easier, when speaking to the marketplace, to talk about a chip, rather than a set of iOS SDK frameworks and APIs.

    “Nike+ Moves made possible by M7” is a refrain that gives Apple a lot of credit for a growing class of apps that will be more often iOS-first, and of higher consistency of quality across the board, making iOS the destination for the most innovative motion-driven apps and app features.

    Also, isn’t it the case that the iPhone 4s and the 5 don’t have the M7, in name, yet don’t seem to be lacking any of the core motion features in iOS7? This to me signal more of a badge engineering exercise than a truly differentiation set of functionality for this generation.

  • obarthelemy

    I have no clue what justifies a separate chip as opposed to some space on a power-gated SoC.
    Apparently Motorola have done the same, with separate “contextual awareness processor” and “natural language processor” on the Moto X launched last month. But they’re using a 3rd-party ARM SoC, so integrating those 2 was more difficult.

    This does seem to go against the tide of SoCs. What next, distinct GPUs ?

    • There may be some specific power savings or reusability aspects involved. I suspect the M7 (or derivatives) will appear in other devices.

    • Walt French

      We don’t actually know yet where the “M7” is, do we? Methinks it’s a tad premature to declare the end of the SoC integration; even if it’s quite distant it could be simply a timing/coordination issue of so many new developments in the phone—it’d be safer to put on a separate chip if there was any doubt about availability during development.

      I think it’s probably separate, though: the sensors themselves are specialized silicon, some very interesting geometrical structures.

      Distance between subsystems is fine if there’s neither a high volume of data, nor speed concerns, and that’d seem the case with the M7 connection to the CPU. *Very* different for GPUs. If I read it correctly, Intel’s newer Z600 Atoms integrate the PowerVR GPUs onto the CPU die.

      Interestingly, I haven’t seen confirmation/denial of whether the A7 uses PowerVR’s newer “Rogue” GPU tech to get the advanced features. One source has suggested that Apple may have gone in-house on GPU capability, too. I doubt it but there are indeed a boatload of new design features that merit attention in the A7.

  • obarthelemy

    Also, battery life seems to be a permanent issue, with a phone’s thinness being hailed as a major feature, while battery life is the really important, though oft sacrificed, feature.

    • neutrino23

      The battery on the 5s has more capacity than the battery in the iPhone 5.

  • oo

    “M” is for Mansfield?

  • Jeff G

    M is for “Me”, likely never going back.

    I never owned an Apple product in my life until May 2011 when I bought an iPad 2. Then, I bought my first iPhone, the 4S within a few months after it’s release (which was an upgrade from my Motorola Droid II).

    I certainly don’t consider myself a ‘fan-boy’, but who knows what that really even means? All I know is that I loved the iPad, and then the iPhone 4S, and their iCloud synching, and automatic backing up. I loved the ease, the lack of viruses, the lack of having to update virus software. I love not having to pay for upgrades, I love how easy they apps are to use, and of course I love the mobility. I also love that whenever I need something done, or smash a screen, or have questions, I pop into the Apple store, and get expert, face-to-face help from people whose have solved my particular issues 100% of the time – I walk out with the job done.

    I have learned by reading message boards and comments on articles that I am dumb. This is good to know because self-awareness is important, and if I’m dumb I want to know. I have degrees in Finance, holistic healing and relationship counseling. I’ve written a book and run an investment practice, raising 3 children, have a house in Scottsdale, and have invented office products. If I had known that buying an iPhone would make me dumb, I might have thought twice about it. I know engineers, cardiologists, pharmaceutical reps, CEO’s of technology companies, lawyers, accountants, judges, nurses, teachers, town managers, honor students and CFO’s who are all apparently very dumb too.

    Anyway, dummy that I am, the value proposition to upgrade from 4S to 5S seems very compelling to me. I know subsidies are also demonized, but I have an upgrade ready and waiting. I can (stupidly) walk into the Apple Store and walk out with a phone that has a better camera, longer battery life, bigger screen, fingerprint security, and a bunch of other cool stuff all for a couple hundred dollars, which I write off on my taxes. Chips with M and A, and speeds though, and technical jargon are not really my concern (CUZ me so stupid).

    Since that fateful day in 2011 Apple products have become the key tools performing my “jobs to be done”. Email, social media, internet, Keynote, Pages, Numbers, my banking app and finance apps, and numerous other utilities have usurped 90%+ of my computer needs. I prepare presentations on my iPad and give them to classes at the local community college. I use Keynote remote on my iPhone to wirelessly control the presentation, and view my “presenters notes” privately. I have totally duped the students, my clients, readers, friends and family into believing I am something more than stupid. Let’s call it stupid luck.

    My wife’s stupid company even bought all of their stupid, high paid reps iPad’s. I also see the stupid Pentagon has approved the iPhone for use by it’s stupid employees. Honestly, don’t these people read the chat rooms? My stupid (straight A) kid even wants my 4s to upgrade from her stupid iPhone 4 and I am tired of writing stupid.

    As full disclosure; I also have a mechanic who does all the work on my car, because I am too stupid to do it. I just want it to start and run well and do it’s job. I don’t want to be an expert in spark plug timing, diagnostics and virus solving…whoops that’s PC’s.

    Anyway, Thanks Apple for making me look and feel smarter than I really am. I’m looking VERY forward to my space gray 5S and perhaps feeling even slightly less dummer than I really am.

    • obarthelemy

      Just an FYI: the same is true about the Android ecosystem, except the retail presence, and adding choice, both in hardware and software.
      And I’m not sure free upgrades are such a boon in the long term: I’d love to be able to reward and encourage the devs that create the apps I use heavily by re-giving them a bit of money from time to time, instead of having that free-rider feeling 5 years on. The “free updates” model works as long as the market is expanding quickly… that may be coming to an end.

  • The biggest annoyance I’ve had with my iPhone 5 is that the GPS drains its battery really fast (it will completely drain the iPhone 5’s battery in less than two hours). This is fine if I’m in a car and it’s plugged into a charger (although it still runs quite hot) but terrible if I don’t have a charger (e.g. I’m on foot). A related annoyance is that the GPS system is very stupid about reflected signals, tunnels, and so forth, often coming to the bizarre conclusion that I have teleported several blocks to the left or somehow left a road while in a tunnel.

    The M7 chip might allow Apple to better leverage GPS (and iBeacon) data using an older technology called inertial location. An inertial locator determines your position relative to a set point by measuring acceleration and orientation changes over time (like dead reckoning using a compass, ruler, and calipers along with data about a ship or plane’s course and its speed). It’s always puzzled me why the iPhone doesn’t sanity check GPS data (and for that matter use fewer GPS data points) by relying on inertial navigation using its compass and accelerometer.

    This approach would in turn also allow navigation where GPS is simply impractical, such as inside buildings, tunnels, or in deep valleys.

    • Klasse

      Excellent point, thanks for sharing!

  • santoscork

    small typo “as a fitness companion but it would not a very good one”

    be is missing

  • santoscork

    I ran across this rather interesting IBM research piece, “The 5 in 5”. It proposes cognitive computing through machine analysis of human cognition, sight, taste, smell, touch and hearing. It’s not too far fetched to realise machines that could be able to assume their surroundings to some degree, proving users with strategies, warnings, advice, recommendations on things other than books, movies, music and the like.

    Perhaps mobile and an adjunct wearable and/or sensors hung around in our homes or cars could be detector or sensors to things that we would not otherwise be able to discover with ease.

  • ituner

    I wish the M7 had a direct link to the BLE radios so that the device can communicate and harvest radio data at much lower power budget (no need to wake up the bit processor for such ‘mundane’ operations….

  • Phil Ramsey

    Has no one guessed it yet?

    Apple will soon release some sort of games console and the M7 chip will allow the 5S to act as a controller for Wii Fit type games!

    Crazy or genius?