IDF There's never been a better time to re-examine the old pals act between Intel and Microsoft in light of news from the Intel Developer Forum this week. The relationship has never been smooth, but between them, the pair have for the last ten years, driven the PC industry.
One perspective that's often overlooked when weighing up the relationship is their respective experience. When a boyish Bill Gates first flew down to Santa Clara, his "Micro-Soft" start up had thirty staff. Intel already had 7,000 staff, and had not only invented the microprocessor, but designed and fabbed two successive generations: the 4004, the 8008, and the 808x series which would be the basis of their fortunes.
Intel was a hardcore, three-biros to the top pocket, industry pioneer, and had learned the business the hard way. By 1981, it had not only needed to perfect the processes for large volume microprocessor production - which in the 60s, many sages said would be impossible - but it had worked out where to sell the damn stuff. It imagined industries where chips could automate the work, and went ahead and built them.
We remind you of this not as a homage to Chipzilla, which is strangely coy about its history, but simply because it's valuable to understanding the immediate future. Companies are often accused of neglecting their "corporate memory", the residues of experience that helped make them great. Well, despite an often ruthless HR policy, Intel has never forgotten what marked it apart from the other silicon pioneers.
So to understand the Wintel relationship, you need to know how Intel thinks. And this is how it thinks.
In the past week we've heard two announcements that will reflect on the relationship in the coming months. One's all about a new market for the gruesome twosome: phones. And the other's about speeding up their aging offspring, the PC.
Let's take the second first.
Deep down Intel knows the mobile experience is truly horrible, and it wants to fix it. We want one inch thick laptops which last all day, but the technology doesn't get us there. Intel can't, like Apple, simply define some standards in a quick half-hour meeting, and email them to the hardware division for implementation. It needs to coerce and finesse its OEMs to arrive at the same destination. (Apple's power management is terrific, but then it doesn't have to play by the ACPI rules).
Likewise Intel appreciates that a laptop user roams between different networks, and doesn't want to faff around with Control Panel settings just to do their work. So it showed off the multihoming Skamania software today, which let you do just that, and it was mightily impressive.
It was an example of great Intel software doing something Microsoft should be doing; so rather obviously, we asked Intel if this wasn't really something Microsoft, the software company, should be doing.
Although the presenter mentioned hardware, and "end-to-end", just then a gigantic thought bubble appeared behind him spelling out the words: "WE KNOW". And the words vaporized to be replaced by the phrase: "BUT WE'RE GOING TO DO IT ANYWAY".
Which is good.
Microsoft has been so concerned with integration recently that the most obvious and incremental consumer benefits haven't had gotten a look-in. This happens when companies navel-gaze for too long. At Microsoft, you can get rapid promotion for suggesting wheezes such as tying the MSDN subscription to Passport. But you get nothing for adding a 'Location Manager' to Windows: something that allows you to create a profile to unify your TCP/IP settings, printer preferences, and choice of networked drives, depending on where you are. Macintoshes have had it for years, and even Linux has a Location Manager these days, for heaven's sake.
But Intel's new found confidence, and possibly (and we can only hope) the promise of continuing scrutiny at Redmond, mean that Chipzilla is happy to fill in the gaps left by dozey Windows product managers.
Intel is adding a humble utility, but if you take a look at the post-apocalyptic Mad-Max-style Windows landscape, you'll appreciate that this isn't easy to do. For ten years Microsoft has incorporated all kinds of neat utilities into the OS, which has left only the brave or the foolish to start Windows utility companies. Capital, and bright ideas, have fled the scene, to go and write free software, or Java, or anything where they won't be crushed overnight.
So we figure, Intel is both politically and commercially free to fill the gap. This may or may not change the relationship significantly: it depends on how hard Intel wants to push it.
Make a phone, if you think you're hard enough
The other new Wintel follow-on from last week's Cannes announcement, detailed here this week, is about the pair creating the baseline for smartphones. Microsoft wants to "open the phone business so anyone can make a smartphone." A Microsoft smartphone.
Of course Microsoft is opening the floor so generously to all comers because it's failed to get a serious established phone vendor to make a Microsoft smartphone. And you can ask yourselves, why is this?
A good question, especially since the business looks a lot more open, and "horizontal" than it did a year ago. You can pick up baseband chips, wireless stacks, and even whole platforms now. But what you won't be able to do is take these parts, and march up to a wireless carrier and demand that they certify it for their network by Friday. Integrating and testing a phone is a multi-million dollar exercise.
So while a basement PC assembler can acquire the parts cheaply, and quite possibly ship something that might not altogether work first time, no handset manufacturer can do this and expect to remain in business for very long. A dodgy phone can bring down the network, and even the biggest and most experienced manufacturers have had to suffer the humiliation of withdrawing models at this late stage.
So why is Intel lending itself to such a spiel? Well, partly because it's instantly good publicity. Some writers, who've only ever understood one thing, the PC industry, can't wait to see the model transplanted.
Intel itself, we suspect, is shrewder, and it knows that for it to win in phones, it doesn't need to rely on Microsoft. The smartphone hardware platform of choice - exclusively right now, outside Motorola - is Texas Instruments' OMAP. Intel needs to notch up XScale design wins, and it's happy to go along for the ride.
It's also happy to help PDAs sprout phone functionality, but such devices are always going to be at a disadvantage to phones-with-bolted-on PDA, because the model favours high volume phone manufacturers, not low volume PDA dabblers. The phone guys can amortize that expensive testing stage (which takes many months) over tens of millions of devices. A Compaq or an HP, or - and let's be generous - a bucketshop startup is only really going to sell hundreds of thousands.
But probably what militates against the phone business becoming a Wintel game is the health - or should we say sickness - of the Wintel PC business as it is today. Who, apart from Dell, is making proper money on PCs? If you discount the white box industry, the garage guys who can assemble twenty a week, the answer is nobody. And we've just explained how white boxers can't exist in any sensibly regulated cellphone market. So that leaves, exactly one company: Dell.
This isn't exactly a lure for a new generation of entrepreneurs. There's plenty of scope in the margins for strange fruit: for single-carrier deals, for single-purpose low-cost devices like today's pagers, but nothing that can hit the volumes of today's PC or handset businesses. We think Intel knows this, and knows how it can win. Microsoft doesn't have much more of a role in this equation than to pay for the drinks.
We began this amble with an aside about corporate memory. It was tougher for Intel to break into control systems then than is for it to get into phones now. It's a case of remembering what it's forgotten, back before when there was a Microsoft to dictate the terms. ®