Are we about to see the end of the Wintel hegemony? Or, worse for Microsoft and Intel, have we embarked on a revolution that will change the whole industry?
Ian Fogg, an analyst with Forrester Research, certainly thinks so. “Clearly what’s happening is a revolution, and it started in the smartphone market,” he said. “We’re living through a revolution in the way we use digital products. It’s broader than the phone market, broader than the PC market, and affects all of our connected devices, including eBook readers, tablets, TV sets and even cameras.”
It’s happening very quickly, too. “How does the phone you use today compare with the one you had ten years ago?” asked Ian Drew, ARM’s executive vice president of marketing. You might have gone from a text-only Nokia 8110 “banana phone” – considered cool when it was featured in The Matrix – to an Apple iPhone 4 or an HTC Desire. Your PC, by contrast, might still be running the same ten-year-old operating system, Windows XP, launched in 2001.
This isn’t to knock the PC business, which has been dominated by Microsoft and Intel since 1981, when IBM chose their products – Intel’s 8088 processor and Microsoft’s MS-DOS operating system – for the IBM Personal Computer. Today’s Windows 7 and second-generation Core processors are dramatically more powerful than the versions that were available a decade ago; but unlike mobile phones, they still perform the same basic functions.
So where does Wintel go from here? In ten years’ time, will we still be running versions of Windows NT/XP on variants of today’s Intel x86-compatible PCs?
The question took on a new urgency at January’s Consumer Electronics Show in Las Vegas. At a press conference on 5 January, Windows boss Steven Sinofsky announced support for system-on-a-chip (SoC) architectures, not only from Intel and AMD, but also from ARM.
ARM dominates the smartphone market, and looks set to dominate the rapidly emerging market for tablets too. Following the success of Apple’s iPad in 2010, around 100 tablets were shown at CES, many of them running Google’s Android operating system on ARM chips.
By showing Windows and Office running on ARM, Microsoft clearly signalled that it wants to compete for a slice of this market, and that it doesn’t think that supporting Atom – the x86-compatible chip that Intel designed to compete against ARM – is enough.
Microsoft’s move to ARM might look like it’s stabbing its old partner in the back, but this is by no means the first time Microsoft has supported alternatives to Intel’s x86 architecture. It’s tried several times before (see Microsoft’s Mixed Bag of Chips on page 37) but the difference this time is that Microsoft isn’t merely hedging its bets by backing a processor that may be successful, as it did with DEC’s Alpha or IBM’s PowerPC.
More likely, Microsoft is simply following its customers: many Wintel PC manufacturers, including Dell, Samsung, Sony and Toshiba, are already using ARM chips such as Nvidia’s Tegra and Qualcomm’s Snapdragon in their products. Microsoft may simply have concluded that, at this point, ARM is unstoppable.
Lenovo’s chairman Liu Chuanzhi puts the threat even more strongly. “We have an extreme focus on the innovation of LePad and LePhone [Lenovo’s ARM-based tablet and smartphone] because these products will dominate the future market,” he said in an interview with Bloomberg. “Anyone who loses this battle will be phased out from the history of this industry.”
Intel’s Graham Palmer, concedes it’s a challenge for his company. “The plus side of this for Intel is that, for the first time in more than 30 years of Windows/Intel compatibility, we’ll have a version of Windows designed for low-power, touch-enabled tablets,” he said.
“This will let us scale down to match our Atom processor family, bringing our leading processor capability to a multitude of smaller devices. On the downside, there is the potential for other companies to develop products for the PC market segment.”
Flirting with other operating systems
Intel has always shown a similar willingness to hedge its bets. It has dallied with Linux, culminating in the deal with Nokia that created MeeGo, a merger of Intel’s Moblin and Nokia’s Maemo. Intel is also believed to be porting Google Android to run on its Atom processor, and Acer has already implemented a port.
The big questions for the future are how long ARM and Atom/x86 will continue to be adjacent worlds with different strengths and weaknesses, how far they’ll overlap, and whether they’ll ultimately merge.
Flint Pulskamp, research director at analyst IDC believes that some cases “will naturally favour ARM and some will favour Intel. But there others – gaming and entertainment devices, networking products – that aren’t particularly orientated to one architecture rather than another. Those are up for grabs,” he added.
“There are also areas such as healthcare and automation that are small markets now, but they’re going to grow explosively. Atom is already pitching for this ‘long tail’ of apps.”
In this brave new world, Intel and Microsoft no longer control more than 90% of the market. Indeed, they might be lucky to get even a 10% share of the sales of new products. The days when Wintel had a monopoly are effectively over, and a growing number of people is starting to define the market in a very different way from the antitrust mavens at the US Department of Justice and the European Commission.
For example, in Deloitte’s tenth annual Technology, Media & Telecommunications Predictions, the global analyst predicts that more than half of the computing devices sold globally this year won’t be PCs. “While PC sales are likely to reach almost 400 million units, Deloitte’s estimate for combined sales of smartphones, tablets and non-PC netbooks is well over 400 million,” the report states.
“Unlike the netbook phenomenon in 2009, where buyers chose machines that were essentially less powerful versions of traditional PCs, the 2011 computing market will be dominated by devices that use different processing chips and operating systems than those used for PCs over the past 30 years.”
Canalys, a UK-based research company, has adopted a similar tack. Unlike other research companies that track IT shipments, it’s decided to count “media tablets” as PCs.
“Any argument that a pad isn’t a PC is simply out of sync,” explained senior analyst Daryl Chiam. “With screen sizes of 7in or above, ample processing power, and a growing number of applications, pads offer a computing experience comparable to netbooks. They compete for the same customers and will happily coexist.”
The question is, where does it stop? If you’re going to lump tablets in with PCs, why not also include games consoles, internet TV sets, set-top boxes, eBook readers and handheld media players? Many of these are now used for playing games, as media centres and for surfing the web – all of which previously required a PC.
However, this isn’t the end of the PC market. “Traditional PCs will still be the workhorse computing platform for most of the globe in 2011,” as Deloitte noted. The world’s installed base of PCs is more than 1.5 billion units and should hit 2 billion in 2014. “However,” the report added, “when looking at the future of computing devices, 2011 may well mark the tipping point as we move from a world of mostly standardised PC-like devices, containing standardised chips and software, to a far more heterogeneous environment.”
Right vision, wrong execution?
Curiously enough, Microsoft saw this coming at least 15 years ago, and tried to do something about it. Recognising that the x86-compatible desktop version of Windows wasn’t going to run on low-powered mobile devices any time soon, Microsoft produced Windows CE to play that role. CE was the base operating system for ARM-powered handhelds such as the Compaq iPaq and other Pocket PCs, games machines such as the Sega Dreamcast and for Windows Mobile phones. It still underpins Windows Phone 7.
Even back then, Microsoft recognised the potential of lightweight devices for mobile internet access, and at the Comdex trade show in Las Vegas in 1999, it said: “Microsoft and leading hardware manufacturers plan to deliver a new category of internet access devices that promise to give non-PC users a fun way to experience the best of the web.”
“Moreover,” added Microsoft spokesman Jon DeVaan, “these devices will not become obsolete; users will always experience the best internet technology as the devices are updated via the web automatically.”
These so-called web companions turned on instantly, were easy to use, cost significantly less than PCs and ran Windows CE on low-power MIPS and Hitachi SH3 RISC processors – but excluded Intel x86 chips.
Microsoft also shared the consensus that smartphones would be important, and at Comdex one year later the company was showing off a prototype smartphone, codenamed Stinger. In 2002, HTC (which made the iPaq for Compaq) announced that it had designed and delivered “the world’s first Microsoft Windows-powered smartphone”, the Orange SPV. The Apple iPhone didn’t appear until 2007.
More than most people, Microsoft co-founder Bill Gates was keen on tablet computing. The company’s efforts go right back to the Pen extensions to MS-DOS, used on the GRiDPad in 1989, then run through web companion tablets, PocketPCs and smartphones with built-in handwriting recognition, Windows XP Tablet PC Edition, and the ill-fated Origami and Courier projects.
In his Comdex 2001 keynote, Gates said: “I’m using a tablet PC as my everyday computer. It’s a PC that’s virtually without limits, and within five years I predict that it will be the most popular form of PC sold in America.”
If Microsoft’s ideas were right, what went wrong? “All those were extensions of the PC market, and they all essentially died,” said Forrester’s Fogg. “Windows CE wasn’t a failure, but it kept the same design cues and APIs as Windows; it didn’t revolutionise the market.”
Microsoft had done well by expanding the Windows market with its Windows Everywhere strategy in the 1990s, extending its range from desktops to laptops and up to servers and even supercomputers. But it never managed to do anything disruptive, unless you count the Xbox. It was evolution rather than revolution.
The Kin phone, developed following the purchase of Sidekick, “was actually very innovative”, said Fogg, “but it didn’t fit the Windows architecture or the Windows branding. They had something that could have been very disruptive, but they killed it in favour of something developed in-house. It’s left Microsoft weak and exposed, because the disruption is happening outside its control”.
It’s the same old story: if you don’t disrupt your own product lines, somebody else will do it for you. Google is simply doing to Microsoft what Microsoft did to IBM. With Microsoft’s decades of guaranteed profits now diminishing, so must its share price.
The best way to reverse that trend is to introduce new products that increase investors’ expectations of future profits – which is why Apple’s share price keeps going up.
Like Microsoft, Intel has successfully moved up the food chain, from calculators (for which it developed its first microprocessor), through desktop PCs to servers and supercomputers. Like Microsoft, the company is now threatened with disruption from below, by ARM processors originally developed in Cambridge for the Acorn Archimedes educational computer. As Microsoft attempts to extend downwards into the new markets for smartphones and tablet computers, Intel is trying to do the same thing with Atom.
As it happens, Intel once had its own ARM business, based on the StrongARM design lab it picked up from DEC after it was taken over by Compaq. But Intel sold that business to Marvell, targeting the market with x86-compatible Atom chips instead.
However, in 2009 Intel bought Wind River for $884 million, a company with huge expertise in ARM, and in January it spent $1.4 billion picking up another ARM business, Infineon.
Flint Pulskamp, at analyst IDC, argued that selling the ARM business might not necessarily have been a mistake.
“With Infineon, it has more of a pure ARM play, rather than having to support the branched-off StrongARM/XScale version. Managing your own architecture costs an awful lot of money!”
But whether ARM or Atom will win the middle ground between phones and PCs remains to be seen. Pulskamp points out that ARM is the natural choice for roles demanding low power consumption and long battery life, while Atom offers better performance and runs Windows.
However, ARM is introducing more powerful dual-core chips, and Pulskamp expects that the Atom will have ten times better power conservation in two years: “that’s much more comparable to ARM”.
If ARM runs Windows as well, “what will be the differentiators? [Other] operating systems, ecosystems, and cost,” he concludes.
Microsoft could design (or co-design) its own ARM chip the way it produced a stunning Xbox 360 processor with IBM; Microsoft’s ARM architecture licence allows it do just that. Equally, Intel could design and manufacture new ARM processors, or put an ARM core on an x86.
Those looking forward to the end of the Wintel alliance should consider the idea that Microsoft and Intel could just as easily work together on WinARM as well. They still have similar aims and common enemies.
We know that Microsoft has shown Windows and Office running on ARM-based devices, and has said the next version of Windows will support ARM. However, it hasn’t said what the user interface will be like, or how it will support touch. Nor has it said whether it expects manufacturers to use the new version of Windows on netbooks, tablets, smartphones or any other type of device. ARM-based servers could potentially hold huge advantages because of the low power requirements and lack of heat, for example.
The key point, according to Sarah Rotman Epps, a US-based Forrester analyst who covers the PC market, is that Microsoft isn’t out of the game.
“I was looking at some new data about what consumers want,” she said, “and the number one operating system they want on tablets is Microsoft Windows. That’s interesting because even though Apple gets a lot of love from the press, people have quite a strong affinity for Microsoft products.”
In this new Forrester survey, 46% of US consumers said they wanted Windows on a tablet, only 16% said Apple, 9% said they’d prefer Android, and 25% said they didn’t know.
“I was surprised to see how low the preference was for Google’s OS,” said Rotman Epps. “So we see this huge gap between what’s available on the supply side, and what’s wanted on the client side.”
The problem, according to Rotman Epps, is that Windows 7 is “not the sort of touch-first experience that Apple is delivering on the iPad”.
“But, on the other hand, consumers want something different. I don’t know if it will be mainly tablets. It could manifest in many different form factors, some of which haven’t been invented yet.”
Both Fogg and Pulskamp point to the Motorola Atrix – an Android smartphone that comes with an optional laptop dock – as an example of the innovation that’s starting to appear in form factors. “I think it’s a niche product,” said Fogg, “but it’s fascinating that people are trying to do this sort of thing.”
Microsoft could develop a new touchscreen interface for the next Windows [Update: since time of writing, Microsoft has showed previews of a new Windows 8 interface designed for touchscreens, as seen in the photo above], as it did for Microsoft Surface (which originally ran on top of Vista). Alternatively, it could do what Fogg suggests and adapt the “glanceable” Metro user interface from Windows Phone 7 to run on tablets.
“That would get Microsoft to market quicker, and be competitive with Android and iOS… or if it had the courage, it would do both.
“The challenge isn’t to go over the ashes of past decisions, but to ask ‘Where do we need to be, and how do we get there?’” he added.
In the early stages, markets often have many competitors trying different approaches before things settle down; the Apple II was an early leader in the PC market just as Nokia was in smartphones.
“The fact that something is first doesn’t mean it’s going to be the last operating system for any particular form factor,” said Rotman Epps. “But it’s incumbent on Microsoft not only to make it work, but to make it great.”
That’s a quote Microsoft boss Steve Ballmer should print out and stick on his wall. It’s not something you’d have to tell Steve Jobs.