Electronics Weekly Electronics Design & Components Tech News
For a $40 billion revenue company in a $300 billion industry, Intel is sounding remarkably defensive these days.
Asked to defend its absence from the tablet and smartphone markets, Intel’s CEO Paul Otellini says the company makes money from the growth of smartphones and tablets because it sells processors to server manufacturers and one new server has to be installed for every 600 smartphones or 122 tablets.
Tortuous or what?
There again Otellini says Intel is working on re-inventing the PC as if it hasn’t already been done and is now called iPad.
Intel’s Damascene conversion to low-power is now official with Intel aiming for power consumption of 15 watts as a mid-point for its laptop processors. Currently, Intel’s mid-point for laptop ICs is 35 watts.
For the archetypal chaser of speed, this transformation is not unlike a Frenchman declaring he’s a one-woman guy.
Otellini slaps down any suggestions Intel will make ARM architecture processors. It is relying for low-power on its move to Finfets at 22nm.
22nm Finfet-based Atoms called Silvermont will be out next year, and 14nm versions of Atom called Airmont will be out in three years.
Intel is underplaying the port of Windows 8 to ARM – defending its status as Microsoft’s best and oldest friend by claiming that Windows 8 will run better on x86 because Intel has a team at Microsoft ‘working deep within the OS’ to facilitate new features.
Intel also argues that x86 will deliver backward compatibility.
So there we are. The Gorilla is getting edgy because there’s a mouse in his cage.
Where would we be without galloping megalomania… it’s what the semi business does best.
Andy Grove might call this “a classic product inflection point” and as he rightly says, the future of companies is determined by their execution, at these critical time points.
Make no mistake about it, the recent Pad and smartphone developments are as critical to Intel now, as the emergence of the PC was in 1980. Intel can choose to project itself as a winner, and than follow through with a methodical plan to dominate this space OR they can brace themselves for a reaming of a lifetime, their choice!
At the moment it is all posturing with very little immediate financial downside. The real problem comes when Intel is forced to choose between delivering the best Pad chipset they can OR protecting Laptop revenue/margins. That’s when we will all see Intel’s true colors.
Scunnerous, the powers-that-be have been working on this Typepad problem and reckon they have fixed it. If you’d like to have another crack at it, it should work, all the best, David
That’s a very interesting way to look at it, Scunnerous, certainly all the other process tricks – high k, metal gate etc have been done for that very reason, so why not Finfets? Sorry about the log-in problem I’ll alert the powers-that-be.
Agree with previous comments and I also find it amusing that Intel presented ARM, in the mobile space, as a “competitive threat”… an upstart, while the truth is that ARM is dominant in that market.
I haven’t seen a single report or article which saw the Intel presentation as a “confession” that they cannot compete with their current process technology and need FinFET and 22nm to get there. One then has to wonder how much ARM will improve when the foundries move to FinFET at 22nm, as they apparently intend to do.
BTW I cannot login here: when I try through Typepad, I get:
“Site is not registered: http://www.electronicsweekly.com/cgi-bin/mt/mt-comments.cgi?__mode=handle_sign_in&key=TypeKey&static=http%253A%252F%252Fwww.electronicsweekly.com%252Fblogs%252Fdavid%25252Dmanners%25252Dsemiconductor%25252Dblog%252F&blog_id=20”
Brilliant, RupertG, gave me a good laugh. Ha Ha.
It’s OK. Intel’s always got Itanium. Er, WiMAX. Oh, ok, UWB. LCOS? SoC? Radio Free Intel (ah, remember that…)? Ovonyx? Larrabee? Er…
Hm. No, no idea why Intel is a bit thoughtful.
Excellent point Stooriefit, no wonder Intel is over-using this solitary club
Of course, one of the attractions of supporting ARM in Windows is that Microsoft get to sell you that software you already own all over again.
It’s just like CD coming in when LP was in it’s pomp. All the yuppies went out in their Golf GTIs and bought Transformer and Brothers In Arms on CD, even though they already had them on vinyl and “compact” cassette.
Great revenue generation for Philips and Sony.
They both tried it again with minidisc and DCC, Sony having qualified success but Philips’ DCC bombed.
I suspect Windows ARM could add a lot to Microsoft’s bottom line even if it only takes off in a minor way.
Intel can’t do this forced obsolescence trick ‘cos backwards compatibility is the only club x86 has in its bag!
Wow. Yes John I see, for MS to say Intel’s comments were ‘factually inaccurate and unfortunately misleading” is strong stuff indeed. MS must have been furious. Yes it seems odd for Intel to be over-stating its case that much. Makes you think it may have over-stated the performance claims for Finfets – just as it over-stated the claim that it invented Finfets. V. fishy. Something’s not quite right at Intel.
Yes, David, exactly – and thanks for tying my point nicely back into your article. I hadn’t even spotted that Intel were also going very close to appropriating the FinFET as well….
And of course, now Microsoft has rebutted this ridiculous FUD for what it is.
Do you think Intel might really be panicking? What month is it – are we close to their next quarterly results? Sometime in June perhaps?
I assumed, Dick, that on ARM/Windows if you want Office etc you can have it. Don’t see why not. The Intel marketing bollox about this is, I assume, simply FUD.
Spot on, John, that sounds right on the ball. The hype on the Intel Finfet announcement implied they’d been invented at Intel and were unique to Intel. Very strange for a company in a science-based field – but showing how rattled they are.
It does appear as though Intel are on the back foot a bit, I’d agree. It always astounds me how they always tried to pretend on the one hand that mobility was not that important – and why would we want good battery life anyway? – and yet on the other hand are now appropriating every mobile semiconductor innovation as if they were the first to have done it! They talk about SoC design as if they were the first to ever integrate a graphics core on the same die as a CPU (and AMD fall foul of this as well with this ridiculous “APU” nonsense – it’s just a cut-down SoC). They’ve made their snazzy announcement of FinFETs, claiming that they will be accelerating their mobile roadmap “at twice the speed of moores law”. The only reason this 2xmoores thing has come about is because they have been putting off manufacturing Atom on 32nm for so long, of course its going to get faster now that their bothering to place Atom even close to the bleeding edge so late in the game.
Their hyperbole is infuriating, and sinking so low as to leak information about Windows on ARM just shows how much of a mess they are in. The fact of the matter is that there are a vast amount of developers that know how to develop on ARM, and are familiar with it – does Otellini really believe there will be no software on Windows ARM? I’ll bet once the ARM PCs are hitting the shelves, they’ll be FUDing the world for years to come….
My favourite so far this week is “Legacy applications won’t run on Windows on ARM.”
So what? Does it matter? No – most ARM/Windows hardware will not want Microsoft Office or Adobe CS5, or whatever, will they?
Dunno, Dr Bob, planar ARMs are better bets than planar Atoms at the same geometries. We know Intel’s claims for the power/performance benefits of Finfets.If Finfet volume production proves the claims true – without increased cost – then, presumably, Yes.
So an ARM built on Finfet 22nm technology would be a better bet then?