Google is reportedly taking a page out of Apple’s playbook and expressing interest in co-developing Android chips based on its own designs, according to a report today from The Information. Similar to how the iPhone carries a Ax chip designed by Apple but manufactured by companies like Samsung, Google wants to bring its own expertise and consistency to the Android ecosystem. To do that, it would need to convince a company like Qualcomm, which produces some of the top Android smartphone chips today using its own technology, to sacrifice some of its competitive edge. Google did not respond to a request for comment.
Within a few years, Google will be competing head-to-head with Apple, with its own line of smartphones, tablets, laptops, and maybe even desktops, all running Android.
While the CPU will remain ARM I don’t think we need more and more custom designs making it hard to write programs that can be cross-compiled.
Really, it is not the hardware around the CPUs I worry about, it is whether we start seeing custom CPUs with instruction sets that do not inter-operate.
Is there work to standardize this?
Why would Google shoot themselves in the foot by making it less likely that people can port their programs from other platforms to Android?
Whatever, your still coding in Java and C++ using the compilers provided by Google, so what’s the point ? And if ever there are some new CPU ‘extensions’, you still have an equivalent of CPUID to check for them and run the right routine.
Yes, it was invented by the mainframes back in the day to make it easier to replace systems.
It is called portable executable format and gets compiled to native code at installation time.
Sounds familiar?
Edited 2015-11-06 07:49 UTC
Android’s is based on ART, which compiles from hardware independant bytecode to whatever the underlying system is. For compiled languages the tool chain will simply produce a fat binary with support for every supported platform built in.
Apple does the same thing – when you compile for iOS you can compile a binary with 3 or 4 different formats now. CPU tech is a commodity, and the toolchains all abstract the details out these days. It’s not a big deal, and the freedom that creates leads to real innovation.
You mean like ARMv7/ARMv8 vs x86-64 (aka amd64) vs MIPS32/MIPS64, which are all currently supported by Android 4.x/5.x/6.x?
The future is already here, you can use whichever CPU architecture you want, and run Android on top, without losing access to any apps in the Play Store.
CPU architecture fragmentation is not an issue with Android.
No SH4/SH5 support ? I’d like to see Android on Dreamcast
That would be their sokution to enable limitless upgrades of the software, enabling any new linux kernel to be run.
A custom implementation of ARM architecture can bring big performance advantage over a standard implementation (see Apple, Qualcomm, Nvidia) but it is quite a difficult thing to do, and extremely expensive unless you sell many million units.
I don’t see Google investing all that money to spoil their relation with Qualcomm, Nvidia, Intel or Mediatek, which are extremely valuable providers for Android’s ecosystem.
Rather, I do see Google moving into ARM’s game of providing the plans, but only for must-have peripherals they can sell or force these providers to implement (hardware for VP10 codecs, security, voice recognition, biometrics, sensor processing, etcetera).
From an article on appleinsider. Ok so they might be a bit biased
http://forums.appleinsider.com/t/189953/google-looking-for-chip-par…
In negotiations with chipmakers, Efrati wrote that Google in particular wants more sophisticated camera processing, enabling features such as faster photo capture and the ability to constantly record the environment, sending images and video to Google for cloud-based analysis.
Why? Because you are the product. Google will know not only where you are but who you are associating with and why. This must be a dream come true for the spooks.
The article links to other reports (https://www.theinformation.com/with-apple-in-mind-google-seeks-andro…) but the full report there is behind a paywall. It would be nice to know if the quote I posted above came from there.
sorry, but this is most certainly a compelling reason NOT to buy any device with this capability.
AppleInsider is worse than Fox News. Anything they say can automatically be assumed to be bullshit. If you see a claim on AppleInsider, you can safely assume the exact opposite is actually true.
Thom,
You are probably right but does the claim not make sense?
With companies trying to make laptops with this sort of feature and an ever increasing ‘always connected’ society something like this is not that too far fetched (IMHO).
I see it as a fairly logical next step but would love to be proved wrong.
I already keep the location services on my company iPhone switched off 99.999% of the time. Actually, Data, Wi-fi and BT are also disabled unless I need them but there again I am a bit paranoid because I did have my identity compromised. My own phone is dumb and has no GPS chip.
I have all those features turned off because I mostly use my tablet to read books and play movies.
I want my mac battery life.
PS. Also on the cruise ships the WIFI connection is expensive to use.
There’s plenty of useful stuff that dedicated image processing hardware could be for. Video stabilisation (like hyperlapse from Microsoft/Instagram), background or object removal (for video chats), text recognition (for translation). There’s applications in augmented reality gaming too. They’re already experimenting with Project Tango, they might want to bring those features to phones eventually.
I really don’t think there’s much benefit for Google in trying to use this for advertising purposes. They already have a profile of you to use for targeting advertising. Why risk alienating people by analysing images and video (for advertising purposes)?
I so tired of people preaching “you are the product”. I know how Google’s business model works. I’ve always understood it, and I think they deliver good value. Why don’t you go tell it to someone watching free-to-air TV, or reading a newspaper or magazine, or watching a movie in a cinema, or taking public transport, or walking on a street in a densely populated area, or visiting most of websites on the internet?
Uh-huh, because simplistic paranoia is ALWAYS a good criteria on which to base purchasing decisions. And while you’re at it, make sure you never vaccinate anything! Ya know, because of the astrological effects of quantum chemtrails.
IMHO, your assertions only hold true if the original journalist is an Apple Fanboi.
If the quote I used is true then what?
Do you want Google to tell the NSA ‘This guy seems to be associating with some undesirable “Ragheads”‘?
I really hope that it does not come true.
As someone who spends a good proportion of my working time in the Middle East working on projects that the likes of the NSA know all about (because the country they are located in is regarded as an ally to the west), I’d hate to be branded like that. Even if it were untrue ‘Mud Sticks’ etc.
If this does come true then I forsee many companies placing blanket bans on all personal cell phones inside their properties. The chance for industrial espionage is just too high.
Just to be clear, I hate all forms of snooping no matter who does it.
IMHO Google will produce a royalty-free Open Source reference SoC.
No
The ARM instruction set is not free.
GPUs and advanced CPUs are made of mountains of patents.
It is just not possible.
Anything is possible for Google. They can literally buy ARM ($22 billion market cap) and licence the patents if they want.
Nvidia and AMD already provide reference platforms for all their GPUs.
They can make a free reference design.
But the patents needed to build the chips belong to AMD/ATI, nVidia, Intel, Qualcomm… so some fee must be payed, the reference design for the chip cannot be free.
Apple got bitten recently about a CPU design patent from University of Wisconsin.
And buying ARM, I doubt that it would be so easy.
ARM works because it is independant.
Take it into the sphere of a Megacorp and many competitors will run away to other low power CPU’s.
Pat ARM the license and away you go and do your own thing with it. By paying your license to ARM and not Google or Apple then you know that you are not funding the profits of your competitors.
I actually heard that from ex- Freescale employees.
ARM is not seen as a direct competitor by TI, ST, NXP, Broadcom, Marvell…
Freescale (now merged with NXP) is in a complex situation as they are contemplating abandoning their own PowerPC architecture. Most others seems to have already made the jump.
Using the same ISA as everyone allows to benefit from an huge ecosystem, but it also makes competition harsher as there is fewer differentiating features, and a big customer could as well buy an ARM license to make its own chips.
ARMs are also used in incredibly diverse stuff. From the minimalistic embedded CPU in a SD card to state-of-the-art phones CPUs. Potential buyers would probably tend to focus on a narrower range, e.g. the high margin high end CPUs.
See how Intel abandoned low end enbedded x86 around ‘2000, making an huge opportunity for ARM for many industries that had to redesign from scratch without any hardware nor software compatibility.
Designing a CPU/GPU isn^A't easy. Does Google have experts at this?
They can hire some. In any case they can licence most of the technology direct from ARM.
Nope. But then again, neither did Apple until fairly recently – they weren’t even remotely capable of building their own chips, which is why they had to acqui-hire that capability by purchasing PA Semi.
… until you can develop an Android application using Android it will always be a toy.
… until you can develop an Android application using Android it will be a toy.
I don’t think anyone will take Android on a laptop seriously until this happens.
Looks like Windows 10 is the only thing so far to have truly merged desktop/mobile.
You can actually develop Android apps in the cloud, so it is possible to develop Android on Android. See codenvy and others.
I admit it is not perfect yet, but it is possible.
That may be harder said than done. Doing hardware successfully (in the business sense) is actually pretty difficult. That^aEURTMs why so many hardware companies (PC and handset OEMs) struggle to make a profit. There is a world of difference between making a few flagship products (such as the Nexus range) intended to act as an inspiration for design motifs but not actually intended to sell in large numbers, or indeed make a significant profit, and running a major global hardware business and sell vast numbers of products at a profit. Running a very large global supply chain and distribution system is in itself a tremendously complex business, let alone getting product designs and costing right, building a deep product pipeline, etc. Plus of course both Microsoft and Google face the same dilemma, the harder they try to copy Apple the more they disrupt their own OEM ecosystem and move into a potentially adversarial relationship with some important ecosystem partners.
Its uncanny how both Microsoft and Google are being drawn along the same road of integrated product development. Both the Wintel and Android ecosystems are based around a very similar model: one company makes an operating system and a large number of OEMs assemble a wide range of products using that operating system. The key difference between Microsoft and Google is that the former sought to monetise its business via selling software licences and Google seeks to monetise its business through high valued added advertising founded on data mining. Both however face a similar problem which is how to exercise control over an heterogeneous collection of OEMs, who actually make the devices that run their respective operating systems, in order to ensure a certain standard of platform quality, security and timely software updates. Increasingly new device features will require a close integration of hardware and the software (right down to the chip level), and that is hard to achieve using the ‘one OS-many OEM’ model.
A common, and in my view deeply false, view of why Microsoft triumphed and Apple floundered during the PC platforms wars in the late 1980s through to the early 1990s was that all tech products get commodified and that the ^aEUR~open^aEURTM nature of the Microsoft ^aEUR~one OS – many OEM^aEURTM model allowed the Wintel ecosystem to out innovate and out compete the cumbersome Apple integrated model. Based on that (false) analysis of what went on during the PC wars many commentators have repeated the same (false) analysis in relation to the new mobile computing device ecosystem. So it has been often argued that Apple is up against some sort of inherent limit of the integrated model and will, inevitably, find itself once again marginalised by the power of the non-integrated multi OEM model.
Its interesting that not only has that not happened in the mobile device markets (where Apple continues to grow strongly, where it takes pretty much most of the profits in the system and where iOS has a more successful platform dynamic than Android) but that it never actually happened in the PC market. Now in the PC market Apple^aEURTMs integrated products continue to grow sales and take the bulk of the profits in the PC hardware ecosystem, OSX suffers essentially no tangible disadvantage from Apple having ^aEUR~lost the PC platform wars, and the Wintel PC OEMs look sick and anaemic.
So all in all I am not surprised that both Microsoft and Google have tried, and may try a lot more in future, to emulate Apple^aEURTMs integrated product design approach. I would be hugely surprised if either managed to succeed in building a profitable hardware business with any real global scale, let alone one that could dent Apple^aEURTMs grip on the top end of the market.
Thinking about all this reminded me of this article by Horace Dediu
http://www.asymco.com/2013/02/19/why-doesnt-anybody-copy-apple/
Google makes money from advertising. They can sell hardware at or below cost (just like Microsoft).
Using an online sales model and OEMS like ZTE Google can be very profitable selling Apple quality hardware at half the price.
In order to seriously compete with Apple Google would have sell hundreds of million of high end devices a year. If they sell several hundred million devices at or below cost Google will go bust. It would capsize their business – and in order to achieve what?
Google made USD40 billion gross profit last year.They could literally give away 100 million high end ($400) devices without going broke.
There is no future for mass produced high end phones. By 2020 the average smart phone will be like current feature phones – $30 disposable devices sold in supermarkets.
Lots of people said exactly the same thing about PCs a decade ago. You can buy a PC for a couple of hundred bucks now. And yet still Apple’s PC business is thriving and performing better than pretty much any Wintel OEM.
Odd isn’t it?
OSX now has an insignificant 8% (and falling) share of the total PC market (compared with 90% for Windows). By any objective measure that is a disaster.
Except oddly if one measures Apple (and the Wintel OEMs) as businesses – which is what they are – in which case Apple is by far the most successful.
Alternatively one could look at the this from the point of view of the impact of PC platform market share on end user experience, which is zero.
End users of OSX suffer no significant disadvantage, and Window users secure no significant advantage, from the relative market share of OSX and Windows.
I have yet to see anybody explain to whom in the real world platform market share of either PC or mobile devices actually matters, and why it matters.
After I wrote the above I came across this article on the PCMag website about the Google deigned chip rumours. They claim its to do with Qualcomm screwing it up and Google want a fall back position (an alternative chip design) if Qualcomm fails to get their act together. Sounds plausible but I have no idea if the story is accurate or not.
http://uk.pcmag.com/google-nexus-9/72844/opinion/google-wants-to-ma…
Apple is getting absolutely slaughtered everywhere except the US and China. iOS devices will probably have a single digit global marketshare within five years.
Slaughtered whilst taking 90% of the profit in the entire global handset business? Slaughtered whilst the iOS platform easily outperforms Android using any metric of platform dynamics? There isn’t an OEM on the planet, PC or device, that wouldn’t swap its current business for Apple’s in a heartbeat.
Exactly the same situation Nokia was in – just before they went broke.
Apple can’t survive the absolutely inevitable commoditisation of mobile hardware (it killed their desktop business back in the 1990s).Within five years (probably sooner) smartphones and tablets are going to become disposable objects with $30-50 price tags.The profit margins will effectively be nil.
Apple is already a zombie company.From the outside it appears to be thriving but it is simply a failed business with no real future.
Edited 2015-11-07 04:48 UTC
You really have a downer on Apple don’t you?
Please take a moment to think outside your box for a moment and wonder what the phones we would be using if it wasn’t for the likes of BOTH Apple and Google?
The truth is that they need each other. If Apple goes away by the turn of the decade whowhat will take its place?
Finally, can you define what you mean by a Zombie company. If you mean what I think you do then you are clearly mistaken.
Sure, apple is no saint and some of their product decisions are crazy but frankly they are in far better shape than Microsoft who keep on getting their footgun out.
A laughable comment. MS has been in business for 40 years successfully selling a very wide range of software, services and hardware. They have made many mistakes but have never been in any danger of failing. Microsoft would probably run hot dog stands if they thought the business was feasible.
I am going to keep your entire comment and print it out and when I feel a bit low I will take it out a start to laugh and fell much more cheerful
BTW given that “commoditisation of mobile hardware” is in your opinion inevitable, and given that apparently Apple lost the PC wars decades ago and that apparently hardware commodification has also operating as a powerful force in the PC markets for at least a couple of decades perhaps you could use your special insight into the tech markets to answer a couple of questions about Apple’s PC business.
Why given commoditisation is Apple’s PC business so financially successful, taking the largest share by far of the profits of all PC OEMs added together?
Why is Apple still able to sell higher priced PCs into the top end of the PC market and why, indeed, is there still a top end of the PC market in existence after decades of commoditisation has run its course?
Why is Apple’s PC business growing while the Wintel PC market is shrinking?
Why is Apple’s proprietary PC platform so very healthy?
Its all terribly odd isn’t, if one didn’t know better it might make one question one’s over simplified view of the global tech markets
The only simplistic view is by people who think the tech market is exempt from basic economics.
I do enjoy your views. I really do.
PC stats are about as honest as a Politician’s manifesto.
We did a quick show of hands at my camera club recently. Again an unscientific poll but it will gove some idea of what is going on.
Of the 94 members, 32 were already Apple Mac users. so approx 1/3rd. Of the rest, three had tried W10 (So were probably included in those stats you quoted). None of them had kept it and had rolled back to W7.
12 were actively contemplating switching to an iMac within the next three months. 7 were going the MacBook route. The rest of the windows 7 users were going to stay on it for as long as possible and go nowhere near W10.
To me, that looks like a disaster for Microsoft and not Apple.
We are all entitled to our opinions so YMMV does indeed apply to you.
Personally, I am glad to be able to become an MS free house. If Apple does go TITSUP then the move to Linux will be easy for me as I’m an RHCA/RHCE. See, I don’t have the Apple ‘Rules’ blinkers on like you might think. 40+ years in the IT industry has made me very sceptical of any company.
I’ve been exclusively using Linux since 2001. I have no need for windows or Mac.
I am not in the USA (Thankfuly).
Apple is not irrelevant here. Far from it.
I travel the world on business and see lots of iPhones devices alongside lots of Androids.
It is not longer a strange event when someone pulls out a MacBook instead of a Windows Laptop. 5+ years ago, it was.
I admitted my survey was not scientific but it does give an idea as to the soer of things people start to consider when face with Windows 10 and all that it means especially the snooping.
Browesr Surveys are just as meaningless. More and more people (on all platforms) are blocking sites like this.
I do at my router. At work they are done the same.
Just a sensible thing to do.
My org is like IBM in that it offers a MacBook instead of a totally locked down Windows Laptop. The takeup is more than 60%. Staff are making an informed choice.
In some places Windows rules. Mostly pirated XP or W7 (especially India).
I could go to the Souk in Riyadh and buy just about any bit of software I want for a few $$$$. But If I want to connect any Windows PC to the Banks network, I have t oget it certified as veing virus and malware free. Don’t have to do that with my MacBook.
Nothing is black and white but I do think you are going way over the top with your predictions of Apple going to Chapter 11 in the near future. That ain’t gonna happen my friend unless there is a repeat of the 1929 Wall St Crash or the South Sea Bubble bursting.
I have seen many things in my 60 odd years on this planet. That is why your predictions make me smile.
You are using anecdotes to support your views. The fact is the statistics show Apple is going backwards very quickly.
Nearly half of Apple’s revenue comes from carrier subsidies for the iPhone. That cash will disappear within three years halving Apples profit margins and destroying hundreds of billions of dollars of shareholder value.
BTW 2025 is ten years away. It isn’t ‘soon’.
Carrier Subsidy?
How USA centric is that then?
In other markets people are more savvy and use unlocked devices and the best rates on short term/rolling contracts.
I pay $16/month for 500mins/500SMS and 1GB data (No 4G charges btw). One month notice and I can take my number with me should I move carrier. Oh, and no roaming charges when I visit the USA and Europe (+ NZ & AU ).
Why would I need carrier subsidy again?
Paying upfront for a device and with this sort of deal makes economic sense. Freedom of choice. no lockin.
Plus I slip in a SIM for Dubai or Jordan or HK or … when I travel there on Business.
Apple is doomed.
Apple has always been doomed.
Apple will always be doomed.
These are eternal truths.
They must be eternal truths because they have repeated over and over again for the last forty years.
This is an interesting article about the relative balance and strengths of different user bases and ecosystems (MS, Google/Android, Facebook, Apple etc).
Just talking about market share is a pretty crude (and frankly clueless) way to try to understand the complexity of what is happening.
https://techpinions.com/user-bases-and-their-potential/42292
Whew! I was starting to worry that you forgot your usual “Apple: an awesome company or the awesome-EST company” proxy-bragging.
I have to ask, though: what exactly is a “successful platform dynamic” – does it enable you to proactively grow synergy? Perhaps you could un-toss that word salad for those of us who aren’t buzzword-obsessed MBAs (or former political lobbyists)?
Its pretty straightforward really. A platform is called a platform because it allows other things to stand upon it. Those other things encompass a wide range of well known third party activities and endeavours as well as user activity. So a mobile computing platform supports a developer community of third party apps, it supports specialist and dedicated software systems for specific professional and corporate use, it supports accessing content (films, books, music, etc), it supports generalised web browsing as well as ecommerce activity, it supports various systems for monetisation activity on the platform, and so on.
If one wants to measure or compare the success of a platform then measuring and comparing platform activity (which can be called platform dynamics for the sake of brevity) is the obvious way to do so. If a platform is called a platform because it supports other things then measuring its success at hosting and supporting those other things would seem an obvious and common sense approach to measuring the relative or absolute success or failure of a platform. Of course of those things that stand upon a platform may make no difference to the experience of any one end user of that platform (they may have no interest in any of them) but at the level of the entire ecosystem based upon that platform they do matter, and they matter in terms of what potential or limitations any given platform offers to potential new end users.
If one measures any metric of platform use or activity or any aspect of the platform economics then iOS comes out above Android. iOS always comes out above Android at the per capita level and it pretty much always comes out better than Android on a total platform usage level – for any metric.
What does that mean? Well for starters it means that total market share in mobile devices does not dictate total platform dynamics (using the definition explained above). Even though it has a minority market share iOS offers the end user a richer, wider and better ecosystem compared to Android. It may be true that if a platform is reduced to a tiny number of users (such is the position of Windows phones) then platform dynamics really do suffer but with an installed and growing base of several hundred million that’s an unlikely fate for iOS.
It may also be true that if the Android ecosystem grows a big enough market share compared to iOS then some of those platform metres may start to favour Android. But the key point is that to the end iOS user, the iOS developer, the iOS peripheral maker, the purveyor of third party content and to Apple itself it wouldn’t matter if Android grew so large that its platform metrics were better than iOS.
One of the reasons it wouldn’t matter is because with probably close on a billion users sometime in the next year to eighteen months (and certainly with several hundred million) and with such a well developed third party/developer community the platform is going to have the sort of installed base, weight and momentum that means that the iOS user is never going to feel left out or ill served. And its clear that Apple can make an astonishing amount of money (far, far more than any other OEM) serving a minority market share.
Finally one needs to consider market segment and customer demographics when considering platform dynamics and relative platform strength. Apple attracts precisely those users that are going to use the platform facility the most. All surveys pretty much show the same thing: iOS users do more on their devices than average Android user, they surf more, buy more, access content more, buy more third party peripherals. The Android system is a useful platform for reaching out to the next billion users of mobile devices (and as a result Android may well be the agent of some startling social change) but by definition those are poorer users. Every hundred million users that Android adds will add far less to platform activity (and hence to supporting the third party platform ecosystem) than the next new hundred million iOS users.
So sorry for a long winded reply but in terms of understanding what it happening in the tech world, and above all in the mobile device tech world, it is absolutely crucial to realise the severe limitations of market share as an tool for understanding platform or business success.
Presuming Google also is going to write their own BIOS. Should be classical ROMs.
I really doubt this rumour is true.
ARM is pretty generic tech, readily available from many foundries, and I don’t see how Google could add much value to it.
Google can make a huge change by offering a royalty-free (by paying licence fees) reference platform to small OEMs and providing oper source drivers. This would allow OTA updates from Google servers bypassing vendors and telecoms.