Since I wrote the last post, a few others have weighed in with their thoughts.
The most interesting thing is that many people believe that Microsoft will be delivering their ARM Tablet OS in time for the back to school sales season in 2012, despite Microsoft having made no statements to date. Instead we get treated to Bloomberg quoting anonymous sources. Bloomberg originally seemed to be the sole source of all of the articles claiming a summer 2012 delivery.
I was very skeptical about the possibility of Microsoft being able to pull this off. If they based the tablet OS on Windows CE, which already ran on ARM, yes, it would be possible, but Windows 7 (or 8)? I just couldn’t see how they could port Windows 7 or 8 to ARM in that time frame. Windows 7 was specifically designed to run on X86/AMD64 processors. Porting it to another processor architecture would be a non-trivial exercise even for a company as large as Microsoft. And the indications I’d seen up until that point were that Microsoft hadn’t started work until relatively recently.
Well, research is fun. I got bouncing around, found this and that, and had to change my mind. It looks like it is quite possible that Microsoft could have Windows 8 in production next year, and ready for use on ARM based processors at the same time.
Part of the problem is that I just haven’t been paying as much attention to Microsoft as I used to. I don’t use their technology anymore, so to a large extent I don’t watch them day to day – most of my concerns have been with financials, not with technology. And financials are a quarterly thing. So a couple of things happened that I missed.
Another problem is that the reporter who had the details is someone I don’t like, and don’t usually read. Mary Jane Foley of ZDNet often reports things that are demonstrably wrong, as reported in my articles Research, Research, And More Research and Research, Research, and More Research Addendum. Now it’s possible she had access to information that we didn’t, but if she did, she didn’t say so. When someone talks about a press release, and uses numbers that AREN’T IN THE PRESS RELEASE, well, I get suspicious.
At the same time, I recognize that she has good sources inside Microsoft. Really good sources. And the story she tells, over several years, is an interesting one.
It starts back on March 3, 2009, when she wrote Will Microsoft port Windows to an Arm-based OLPC laptop? The question was raised by the OLPC Project planning to move to an ARM chip for their next model.
On June 3, 2009 she published Microsoft: No Windows 7 for ARM-based netbooks (for now) which covers an official statement by Microsoft that they wouldn’t be producing an ARM version of Windows 7. Windows 7 wasn’t released to manufacturing until a month later, and wasn’t released to the stores until October.
On January 4, 2010 she wrote 2010: The year Microsoft could give the tablet another shot? At this point rumors about the IPad were everywhere, and since Microsoft had been one of the early innovators in the tablet market, it was an obvious question.
On January 22, 2010 she wrote Are slates going hurt Microsoft’s bottom line any more than netbooks? Remember my prediction that Microsoft could see bankruptcy? Pressure on margins is one of the problems that Microsoft faces. To quote Mary Jo:
Are all slates tablets? Are all tablets netbooks? Does the distinction between slates, tablets, netbooks, smartbooks and e-readers matter to anyone — or do anything beyond confuse customers more than they are already)?
The problem from Microsoft’s point of view is that if it walks like a duck, and quacks like a duck, the customer is going to use it for a duck, even if it’s a moose. Slates, tablets, netbooks, notebooks, and smartbooks have the same capabilities as far as consumers are concerned. All can do email, surf the web, watch videos, etc. E-Readers are more limited, and consumers recognize that, as are smart phones, though the problem with smart phones is due to screen size more than capabilities.
Microsoft can’t sell a $500.00 copy of Windows to install on a $300.00 piece of hardware, consumers don’t see the value. And if Microsoft pulls too many features, consumers migrate to Apple’s IPad, which isn’t feature limited, or when they hit the shelves the Android powered tablets, which again aren’t feature limited.
On January 27, 2010 Andrew Nusca wrote Why Apple will sell millions of iPads in 2010, where he compared the IPad to the Netbook, another device that many pundits had thought was an ‘extra’ device that people didn’t really need, but which fulfilled a need for a basic, portable, inexpensive computer. A year later, when we know that Apple sold 15 million IPads in 10 months, we know he was right.
On January 28, 2010 Mary Jo asked Does Windows need more padding to fend off Apple’s iPad? Her initial response to the IPad was somewhat dismissive, however like everyone, she hadn’t actually seen one at this point.
Now we’ll take a side trip. This is another part of why I’ve predicted Microsoft’s demise, On February 4, 2010 Mary Jo wrote Microsoft’s challenge: Innovation, innovation, innovation which covers some of the problems that a large corporation like Microsoft can have in developing new products. I’ve worked for large companies (though never one of Microsoft’s size) and seen the sort of infighting that she talks about. It isn’t fun, if you are on the loosing end.
Along the same line, on February 8, 2010, John Carroll wrote Microsoft’s innovation disconnect, which covers some of the same issues, from a different viewpoint. When you put both articles together, they explain Windows Vista, and several other Microsoft products which under-delivered.
On March 25, 2010 Mary Jo wrote Microsoft’s Courier: May the best OS win, which covered the various OS possibilities for use in the Microsoft Courier tablet project. She covers the logic behind each. It’s too bad that the project never saw the light of day, it would have been interesting seeing which Microsoft would have picked.
On April 28, 2010, David Morgenstern wrote Palm-HP: Microsoft bites bigtime, covering the HP purchase of Palm, and it’s implications for Microsoft. As Morgenstern says:
Certainly, such a solid technology partner as HP has been deeply involved in evaluating Microsoft’s OS road maps. For example, it was Steve Ballmer himself who took the wraps off of HP’s Slate tablet computer at CES in January. At the demonstration, the device ran Windows 7.
In fact HP’s buy of Palm was a slap in the face for Microsoft. The largest PC manufacturer showed that it had no confidence in it’s major technology partner. This also sent a message to all of the other PC manufacturers. They’ve got to be considering their options as well.
On April 29, 2010 Mary Jo wrote Microsoft nixes plans for its dual-screen Courier tablet. The timing is really curious taking into account the HP-Palm deal just days before, and the retail release of the IPad on April 3, 2010.
On April 30, 2010 Mary Jo followed up with Microsoft and HP: Love on the rocks? This covered both Microsoft’s dropping of Itanium support, as well as HP’s purchase of Palm.
On the same date David Morgenstern wrote Apple iPad’s tablet competition drop like flies, e-book readers next, in which he covers the early changes to the market as it becomes clear that the IPad is selling very well, HP has cancelled the Windows 7 Slate, Microsoft has cancelled the Courier, and he questions whether E-Readers can survive.
On May 25, 2010 David Morgenstern wrote Prediction hell: How dumb do Bill Gates & Microsoft’s iPad, Surface predictions look now? I think the title covers it quite nicely.
On June 1, 2010 Mary Jo wrote Microsoft delivers public test build of Embedded Compact 7 operating system. Don’t forget that development on this OS must have started some time in 2007.
On June 3, 2010 Mary Jo wrote At Microsoft, there’s a thin line between a tablet and a slate, where she spends some time trying to determine what Microsoft’s definition of Slate is, and ends up confused. The problem is that Microsoft appeared to be trying to force an artificial difference between devices, which would have allowed them to charge more for the OS for one device than another. While I can understand why they’d want to make more money, this is against the best interests of their customers, who want the least expensive device that will suit their needs, and to whom the Operating System brings no real value.
On June 17, 2010 Mary Jo wrote Demystifying Microsoft’s mobile operating system roadmap. She claims that Microsoft actually has half a dozen mobile operating systems.
On June 28 2010 Mary Jo wrote Microsoft starts sharing Windows 8 plans with PC partners. From here on in is where things get interesting, and a bit confusing.
On July 12, 2010 Mary Jo wrote Microsoft’s Ballmer: Windows 7 slates are coming this year – curiously I haven’t seen one yet, and it’s March 2011.
On July 22, 2010, Mary Jo shocked everyone, she wrote that I confess: I bought an iPad (and so far I love it). Her reasons for her purchase:
But none of these has the amazing battery life, fast on/off, ultra-handy form factor, a made-for-touch interface (rather than touch bolted on to a keyboard/mouse-centric interface) or a built-in app store.
These reasons helped Apple sell 15 million IPads in 10 months, shaking up the market. These reasons are why Hard Disk Drive shipments are predicted to drop in the first quarter of 2011 by over 3%. And these reasons are putting a lot of pressure on the other PC and consumer electronics OEMS to come up with competing kit, before Apple owns the market.
Of course Apple won’t be able to own the market, because Foxcon isn’t capable of producing that much product. But Apple is going to make a hell of a lot of money out of the IPad, and it will probably help them sell more Macs.
On July 23, 2010 Mary Ho wrote New Microsoft, ARM licensing agreement; Could a Windows Phone tablet be coming? In it she first mentions the rumors about LongARM. LongARM was supposed to be an ARM version of Longhorn, which eventually became Windows Vista after many delays. As she mentions Microsoft had an existing agreement with ARM. This was an extension of that agreement, and neither side was willing to admit what the extension covered, which is interesting, considering Microsoft used to conduct most of it’s product development in public.
On July 29, 2010 Mary Jo wrote Microsoft’s Ballmer: Windows 7 slates are ‘job number one’ – the question is whether they will sell.
On August 2, 2010 Mary Jo wrote What Microsoft isn’t saying about its iPad compete strategy and I think she hit it on the head. I don’t think anyone, Apple included, expected the IPad to sell as well as it did. I think that everyone expected it to sell well, but Apple was still having trouble keeping stock on the shelves past Christmas, a sign that the company had underestimated demand. And if the manufacturer had underestimated demand, you can bet that their competitors, like Microsoft, didn’t have any idea how well it would sell.
So Microsoft was playing it by ear at this point, and planting misinformation. They hadn’t decided on what their final plan to combat the IPad was, in fact they probably had only recently become certain that they needed to directly address the IPad. Remember that this was written just four months after the release of the Wifi IPad, and just three months after the 3G IPad, and at this time for all everyone knew the IPad was just a fad.
On August 9, 2010 Mary Jo wrote Microsoft’s RearType: Physical keys to the iPad, Kindle and tablet kingdoms? The importance of this is it shows that Microsoft, like Apple, is trying different concepts. After all, what is an IPad but a Netbook with a virtual keyboard?
On August 24, 2010 Mary Jo wrote Another Windows 7 slate dropped from this year’s Christmas list about MSI’s decision to hold off production until a more advanced Intel chip was available. This points to the problem with Windows 7 on Intel processors – battery life. Which is why Apple used an ARM processor, and option not available for use with Windows 7.
On September 8, 2010 Mary Jo wrote Office for the iPad? Or are Office Web Apps good enough? What Mary Jo doesn’t consider is that Apple has already set the acceptable prices, and Microsoft isn’t going to want to sell a full featured copy of Word for $10.00!
On September 15, 2010 Mary Jo wrote Microsoft’s Surface tablet: Three years too late? in which she comments that if Microsoft takes three years, it will be three years too long. She also comments that her while she’d prefer a WinPad her IPad is doing what she needs, and doing it well.
On October 5, 2010 Mary Jo wrote This year’s Windows 7 slates: Not under my Christmas tree in which she expresses skepticism about Microsoft’s Christmas 2010 Slates for technical reasons. And she was right – there was no way that Microsoft could match the IPad in that time frame. Google’s Android was in slightly better shape, but it couldn’t match the IPad until Honeycomb was released in 2011.
On October 15, 2010 Mary Jo wrote Are Windows Phone 7 slates waiting in the wings? Don’t get your hopes up in which she covers how Microsoft is artificially limiting the hardware that the OS is installed on, and that therefore there won’t be any WP7 tablets.
On October 20, 2010 Mary Jo wrote Don’t forget: Microsoft is (supposedly) working on a Windows app store, too, in which she totally forgets that the first App Store was on Linux (like everyone else). Of course App Stores are just another part of the wave of disruption that is overtaking the industry. Brick and mortar software stores will be wiped out, like record stores. Prices for software will be driven downwards towards zero, putting pressure on Microsoft’s margins.
On October 28, 2010 Mary Jo wrote Microsoft: iPad? What iPad? in which she covers Microsoft’s Q1 earnings, which looked unaffected by the IPad. As I pointed out in my own evaluation, it was too early to see any effect from the IPad on Microsoft at that time.
On December 13, 2010 Mary Jo wrote Microsoft to show off true iPad competitors at CES? She questions a New York Times article which claims that Microsoft is going to show off a slate running Windows 8 at CES. The problem as she well knows is that Windows 7 isn’t capable of being a true IPad competitor, since it isn’t designed as a Touch operating system. It has Touch features yes, but it’s mainly designed for traditional mouse and keyboard input. In fact she’s starting to sound pretty depressed. She’s a Microsoft fan, and they are letting her down.
On December 21, 2010 Mary Jo wrote How will Microsoft address the ARM-based tablet space? She again considered the possibilities of Windows 8, but thought that it was likely to be 2013 at the earliest.
On January 3, 2011 Larry Dignan wrote So long Wintel, hello Google, ARM? He talked about the possible death of the Windows-Intel Alliance and its replacement with a Google-ARM Alliance, quoting a research note which regarded mobile ‘Touch” technology as “highly disruptive.” Based on sales of the IPad, I quite agree. Even if Microsoft does produce a “Windows for ARM” it’s going to have to accept lower margins, as Google’s Android is free. In fact Charlie over at SemiAccurate claims that Windows Phone 7 is free. That’s one hell of a business model…
On January 4, 2011 Mary Jo wrote Forrester: Line between apps and browsers blurring with tablets in which she covered a Forrester survey that expects tablet sales to reach 44 million per year in North American by 2015. I think they may be low on that estimate myself. The study also mentions that a lot of average users don’t know the difference between and App and a Browser, and that website owners will have to optimize their websites for tablet users.
Also on January 4, 2011 Mary Jo wrote CES: The long and winding road to Windows on ARM about speculation that Windows 8 will run on ARM, and that there will be some sort of announcement at CES.
On January 5, 2011, while at CES Mary Jo wrote CES: Will ‘Jupiter’ be key to Microsoft’s Windows 8 app store’s future? Jupiter is apparently a new application packaging technology that Microsoft may be using for Windows 8.
On January 5, 2011, while at CES Mary Jo wrote CES: Microsoft shows off Windows 8 on ARM. They also demoed it on an X86 SOC at the same time, and the demos were apparently the same (they wouldn’t allow the demos to be videoed).
This is the big one. To have a demo up and running, Microsoft had to have been working on this for well over a year. This totally throws off all of my original calculations. My original assumption was that Microsoft didn’t start working on an ARM port until after they saw that they IPad was a success, but this demo makes it clear that they must have started work on the port before the IPad was announced.
Now there would have been good reason for Microsoft to have done this. By the middle of 2009 it was quite clear that Intel was failing totally on producing low power X86 processors. Intel was also a failure at producing decent graphics units.
At the same time ARM Holdings has been producing more and more powerful chips. As an ARM licensee, Microsoft would be aware of the ARM roadmap, and the fact that ARM chips would reach approximate performance parity with Intel chips in 2012, while maintaining exceptionally low power requirements.
Add all of this together with the Windows 7 RTM of July 2009 and what do you get? A 2009 development start date for Windows 8 with both ARM and X86/X86-64 versions. The ARM version could have been started slightly later, but it was probably started close to the same time.
Compare the secrecy around Windows 8 with the lack of secrecy around Longhorn. With Longhorn Microsoft talked up features such as WinFS which ended up never being delivered. With Windows 8, Microsoft has been working quietly. There’s a saying in sales – Under Promise and Over Deliver. Microsoft appears to have learned that lesson after Over Promising and Under Delivering with Longhorn/Vista.
On January 5, 2011, while at CES Mary Jo wrote CES: What Microsoft’s Ballmer didn’t say. I’ll skip most of what she said and quote one paragraph:
Most alarmingly, he didn’t have anything to say about how Microsoft plans to address the slate market beyond what company officials have said already — namely, that Windows 7 makes a darn good slate/tablet operating system and will be the operating system that Microsoft makes available to its partners for the next two-plus years. I strongly disagree, as even the nicest looking Windows slates hitting the market are either 1. super pricey; 2. horrible re: battery life; 3. heavy/bulky; and/or 4. not touch-centric.
Having checked the prices and specifications on the Windows 7, she’s right.
On January 7, 2011 Mary Jo wrote With Windows coming to ARM, what happens to Windows Embedded Compact? And that’s a good question. There are a lot of companies that depend upon the product, so what is Microsoft going to do with it?
On January 24, 2011 Mary Jo wrote How Microsoft plans to market against the iPad in which she covers a Microsoft presentation she’s seen about how to sell against the IPad. In her final paragraph she says:
Microsoft and its partners cannot afford to stand by idly as the iPad gains more traction. During Apple’s most recent earnings call, officials there said Apple sold 7.3 million iPads in its most recent quarter. According to Apple, 80 percent of the Fortune 500 are piloting or deploying iPads. If iPads are categorized as “computers,” Apple — as a result of iPad sales — is the No. 2 worldwide PC vendor.
Um, pardon me, but the IPad is a computer. An odd one in some ways, but you can do anything on it that you can do on your desktop (except multi-task).
On January 25, 2011 Adrian Kingsley-Hughes replied to Mary Jo with Microsoft’s secret weapon against the iPad is very, very familiar. What Adrian is ignoring, and what Mary Jo has known all along, is that until Windows 8 arrives, Microsoft doesn’t have anything to compete with the IPad. That’s why Mary Jo owns an IPad.
On February 1, 2011 Mary Jo wrote Why IE 9 is key to Microsoft’s tablet push where she covers some cosmetic advantages IE 9 brings to Windows 7 on tablets over IE 8. She also admits she’s a Chrome user
On February 16, 2011 Mary Jo wrote A Windows 8 tablet in Q1 2012? Hmmm… Based on my revised calculations on when Microsoft started working on Windows 8 for ARM, I don’t see how this could be possible. Major Windows releases usually take about three years to produce (Vista was an exception, Windows 7 is really Vista Service Pack 1). But…
On February 21, 2011 Mary Jo wrote Windows 8 roadmap: A picture is worth a thousand build numbers. Now this would make sense. Thee years from the Windows 7 RTM would be the summer of 2012. No doubt Microsoft and the hardware OEMS would like the release to be early enough to get machines on the shelves for the back to school shopping season. And by keeping quiet, and not making any promises, Microsoft avoids the problems that they had with Longhorn, where Murphy’s Law happened with great regularity.
On March 1, 2011 Mary Jo wrote Microsoft delivers its ‘other’ tablet operating system: Windows Embedded Compact 7 – in this article she once again highlights Microsoft’s artificial attempt to divide tablets into ‘consumption’ and ‘consumption-creation’ devices. This plan is going to hurt them. I know cartoonists using the IPad, novelists using the IPad, songwriters using the IPad (one of whom wrote a songwriting App called Scansion – consider this a free ad for Phil Mills, who is a really nice guy). People consider tablets to be computers. They know what IPads can do, if they buy a Windows Embedded tablet and it can’t do those things because Microsoft won’t let it, they aren’t going to be happy, and Microsoft will get the blame.
On March 2, 2011 Mary Jo wrote Does Apple’s iPad 2 further dent Microsoft’s iPad compete plans? Her conclusion is that Apple hasn’t take the Enterprise seriously with the IPad 2, since they didn’t include remote wipe.
What does this all mean? Well, Microsoft has officially announced Windows for ARM. This indicates that Microsoft thinks that Windows CE based product
- is not capable of competing with the IPad’s IOS
- is not suitable for a tablet computer
- some combination of the two
We can now be relatively certain that Microsoft was already working on an ARM version of Windows when the IPad was announced. The timing of certain things, i.e. Microsoft’s new agreement with ARM Holdings, the RTM of Windows 7 and the start of work on Windows 8, the point when Intel’s failure to deliver on their low power chip promises became readily apparent, the point when ARM’s processing power parity with Intel became visible. Add this all together, and unless senior staff at Microsoft are total fools (and they aren’t) they would have been making alternative plans, with the alternative plan being an ARM port. I didn’t see this coming, because parts of this I didn’t know until after New Years day this year, and this is the first time I’ve actually sat down to try to put it all together.
Microsoft’s biggest problem is that while they have a plan, and it’s a solid one, it’s late. Apple apparently saw where things were going earlier, as did Google. In part that could be because Apple hasn’t been tied to the X86 platform as long as Microsoft, and Google is new to the operating system market, with it’s first operating system, Android, being for ARM.
Another problem is that effectively IOS (the IPad OS) is free, and so is Android. Apple came in at a very aggressive price point with the IPad, a price point that they Android tablet manufacturers are having a hard time meeting. If Windows for ARM has any cost at all, manufacturers won’t use it. They won’t be able to use it, because it will make them uncompetitive.
Microsoft’s cash cow has always been Office. About half of Microsoft’s profits come from Office. Microsoft has a problem here as well. Apple sells the individual parts of IWork like Pages for $9.99 each. Microsoft can’t afford to sell Word for $9.99, and it can’t afford to give away Outlook, like Apple gives away Mail. It’s just not a sustainable business model.
Microsoft has 89,000 employees as compared to Apple’s 49,400 employees. Apple has marginally greater sales, with less employees ($1.32 million per employee compared to Microsoft’s $0.70 million per employee). Microsoft has actually slimmed down, at it’s peak it employed 92,000 people, at a time sales were actually lower.
The point of the numbers is that Microsoft can only continue to function in this way by keeping the price of software artificially high. Competition drives prices downwards, and Microsoft has been successful in the past in limiting competition. In the tablet market it can’t limit competition because it doesn’t have a credible contender.
By re-writing the rules governing portable devices (including notebooks) Apple has broken Microsoft’s strangle hold on the industry. We have HP, who is one of Microsoft’s biggest partners openly stating that they are going to install WebOS on Notebooks as well as tablets. We have companies fighting to have the first Honeycomb tablet on the market. And while there are Windows 7 devices, they aren’t priced to compete.
I think that we will see a dip in Microsoft’s year end financials, which will be released at the end of July.
Sunday March 6, 2011