Why Intel Will Be a Mobile Loser

59 Comments

Intel (s INTC), as it’s wont to do, overnight made a splashy unveiling of a new family of processors: the Atom Z6xx series, whose chips are much more powerful than current versions but consume less power. Why the hoopla? Apparently these chips, which run at over 1.5 GHz, can be used to power not only smartphones but also tablets.

So as I usually do when a public company announces major news, I checked Intel’s stock price in pre-market trading: largely unchanged from its closing price the day before. If this was supposed to be such a big deal, why was the market ignoring it, and responding by shrugging its proverbial shoulders? Here’s why.

Future Shock

It reminded me of a piece of wisdom imparted to me by a friend of mine, a veteran of Wall Street. He says that investors pay for growth, and that anytime they express apathy towards a growth stock, they are pointing to a company’s inability to find new markets.   On the other hand, if they’re willing to overvalue a stock, then they believe that company is going to find newer markets to keep growing its revenues. Google (s GOOG) and Apple (s AAPL) fall into that category.

Chart forIntel Corporation (INTC).pngIntel’s stock performance over five years, as this chart shows, has been mostly flat to down. Much like Microsoft (s MSFT). Just as they have always been, the fates of the Wintel duopoly are still pretty intertwined. From the looks of it, both companies are going nowhere fast. Which is amazing considering that both companies still have a monopolistic control over the personal computer ecosystem.

Former Apple executive Jean-Louis Gassee, in a recent blog post titled “Very Personal Computing,” wrote:

The personal computer has reached the S-Curve’s shoulder while very personal computers are still at the S-Curve’s knee, poised for the type of growth the PC has enjoyed over the past 30 years…Apple, Google, and now HP have seen the past and the future: The PC business is mature and graying; the growth is with the new very personal computers. Relying on Microsoft (or even Google, unless you’re Google) for the operating system puts you in a fast race to the bottom, to meager margins, to having key decisions for your business made in Redmond or Mountain View.

The market has read the tea leaves as well, thus explaining the stock performance of Microsoft. Same goes for Intel. Despite its efforts to launch new chips or dabble in likely-to-fail OS efforts such as its joint venture with Nokia (s nok), the Mobilin, Intel resembles an elephant on top of quicksand.

Take its Atom Z6xx series announcement, for example. From a short-term tactical perspective, it will be months before these chips actually show up in phones on store shelves. It will be months before one of those phones gets any kind of traction, volume-wise. As for tablets — now even an optimist knows that would be a tough market to crack. In other words, don’t expect these new chips to have an impact on Intel’s bottom line anytime soon.

Mobile Chip Competition

Take a step back, and it becomes very clear that the company is facing competition from deep-pocketed, well-entrenched competitors — and they’re no pushovers. Because it had a monopoly on PC chips, and faced feeble competition in the x86 world, Intel managed to put its rivals out of business and maintain very fat margins, making it one of the most beloved stocks of the 1990s. Thanks to its overflowing coffers, it also experimented with different markets — networking and ultra-mobile processors (such as XScale) — by buying a lot of companies. All these attempts to expand beyond PCs and servers were schizophrenic at best.

Why? Because it couldn’t reconcile the future with these lower-margin products. As a company, Intel is addicted to the margins it gets from the PC and server markets. It needs those margins to keep its money machine going. But these are slow-growing markets. To wit, its own low-cost Atom processors are cannibalizing the core PC market, thanks to the growing popularity of netbooks at laptop replacements.

Anand Chandrashekhar, Intel’s senior VP and general manager of the Ultra Mobility Group, said: “Intel has delivered its first product that is opening the door for Intel Architecture (IA) in the smartphone market segment.” That is a shocking statement, and one that shows Intel’s long road ahead.

It’s an ARM’s (Mobile) World

IA is making its debut at a time when its rivals are firing on all cylinders. The ARM-based mobile application processor ecosystem is as crowded and vibrant as an Asian bazaar. From Qualcomm’s (s qcom) Snapdragon to Nvidia’s (s nvda) Tegra to Texas Instruments’ (s txn) OMAP, the smartphone and tablet markets are very competitive. ARM-based chips are faster and consume a lot less energy. It will be at least a year before Intel can match them in the power department, analysts say.

Unlike in the PC market, where Intel’s best competitor was an anemic AMD (s amd), its mobile industry rivals are pretty cash-rich. And none is stronger than Qualcomm, which in many ways is a proxy on the fast-growing Android smartphone market. Qualcomm at the start of this century was quite aware it was losing its CDMA dominance and as a result started looking at new markets, so now it has a major head start over pretty much all its rivals.

The company has a minority stake in HTC, one of the largest smartphone makers in the world. It’s also been a major power player in the Android ecosystem and has very close relations with Google. More importantly, Qualcomm knows how to integrate application processors with mobile radios, thanks to a lot of in-house intellectual property and expertise — something Intel sorely lacks.

If Qualcomm is a fearsome competitor in the Android ecosystem, Intel is locked out of the Apple ecosystem. Apple has bet the farm on its internal chip technologies such as the ARM-based A4 currently being used inside the iPad. In the iPhone, ARM is the architecture of choice as well. Microsoft has started working closely with Nvidia’s Tegra and RIM’s (s rimm) devices, too, are ARM-based. Last month Hewlett-Packard (s HWP) agreed to buy Palm (s PALM) in a deal valued at $1.2 billion, and with that its own OS that runs on ARM — not Intel-based — chips.

History Repeating Itself

Many of us focus too much on the here and now, forgetting the lessons of history. Back in the day, IBM (s ibm) and Digital Equipment tried very hard to compete with PowerPC and Alpha chips, but couldn’t make a go of it against Intel’s dominant x86 platform. They failed because they were trying to reinvent the wheel.

The mobile ecosystem is pretty much the same — trying to go against the ARM ecosystem is trying to reinvent the wheel. ABI Research says that ARM-based ultra-mobile devices will surpass x86-based devices by 2013 because, as Stacey wrote, “ARM has always had an advantage in mobile because the chips based on the instruction set were designed to sip power rather than glug it. That translates into a longer battery life and presumably a smaller form factor for the battery and end device.”

Ashok Kumar, analyst with Rodman & Renshaw, put it best when he said:

They’ve publicly said that only in the next version of Atom–which is a 2011 event–will power consumption be low enough to truly address the smartphone market. That’s been the official positioning so I don’t know what’s changed. In terms of actually making a push into the ARM market, that’s likely to be a 2011 event.

To put it bluntly, it won’t be until next year that Intel will have a competitive offering on the market — an eternity in the mobile world. Too bad Intel sold its StrongARM technology to Marvell (s mrvl).

Related content from GigaOM Pro (sub req’d):

As Devices Converge, Chip Vendors Girding for a Fight

For Phones, the Future Is Multiple Cores

59 Comments

MSP

Intel isn’t behind technology wise, the problem is that the ix86 architecture carries a lot of unnecessary baggage in order to be able to run Windows. Basically, it has to be able to run code for previous generation Intel CPUs from the i386 upwards. This means it has a much higher hardware footprint than ARM CPUs which are about as lean and mean as you can get. This doesn’t matter too much on high end power hungry CPUs since you can simply add on more transistors, clock frequencies up, and compensate by fitting bigger heatsinks and cooling fans. On low power devices however, this is a real killer. Microsoft planned a Windows 7 based tablet based on Intel Atom chips, but there were no OEM takers, simply because battery life was too short. The Atom chips also underperform compared to ARM chips.

If Intel is to compete, they will have to dump the ix86 and develop a lean RISC architecture like the ARM from the ground up.

JazzCat

ARM will be bringing their dual-core chip with integrated gpu to market before intel are truly out of the starting gate. I don’t see them threatening the status quo. Go ARM! : (I worked at Acorn, eons ago. :)

Sam

I would argue that TI is a bigger competitor than Qualcomm in this space, and TI moved away from offering integrated radio solutions. Their reasoning is that radio is a commodity product, so Qualcomm’s advantage isn’t really much of an advantage. Want more proof – check the current smartphone offerings and you’ll see that almost all OEMs use discrete solutions. Even Apple uses different vendors for basedband, radio along with its own A4. Integrated radio expertise isn’t really a differentiating factor.

Danny

I think the stock didn’t react positively because it hasn’t received any media attention, at least not that I’ve seen, and I try to watch financials everyday.

Eventually we’ll have a bunch of little computers running around looking like smart phones. Intel processors would certainly foment that I think.

gp

well Mr Gigaom is known apple fan and worshiper ….how come when mighty Apple who choses ARM processor …new low voltage can be good choice …..stuff like coputing power , power consumption ability run majority of app with recompilation be good thing ?

Bob Peterson

“Why? Because it couldn’t reconcile the future with these lower-margin products.” Intel unwisely hitched its future to Microsoft. It built chips only for the operating systems Microsoft would make, so the fractured approach to mobile reflects exactly what Microsoft achieved in mobile.

Dennis Forbes

I don’t think anyone expects Intel to instantly become a major force in the smartphone market over the coming month or even year. They only recently gained any interest in the field, obviously reading the tea leaves and seeing the changing focus of technology.

Competition is great. Your undermining of the challenges Intel has faced in the traditional CPU market is quite inaccurate: over the years countless challengers have come and been defeated. I fondly recall Motorola coming and going. SPARC coming and going. POWER coming and going. Cyrix and a million other upstarts coming and going. Transmeta coming and going.

The whole “RISC Revolution” of the ~1990 that came…and went.

And here we are today with the Xeon 7500 being head and shoulders above everything else available, where power consumption isn’t a concern.

I love my Nexus One, but if Intel can spur more competition and innovation in the field I am all for it. As an Android developer I have the luxury of knowing that my creations are platform agnostic.

Steve 'Chippy' Paine

You use a long-term title with a short-term article Om. 2011 is not an eternity.

I don’t take any notice of stock characteristics but found this to be an interesting read.
http://carrypad.com/2010/05/05/morgan-stanley-report-moorestown-launches-intelnokia-smartphone-expected-in-mid-late-2011/

PLUS

You need to add Android into the equation (ready for Moorestown at launch, demo’d already) along with Nokia (IP, voice stack, marketing, partner for devices, radio technology) and also, as time goes on, the appstore element. One appstore across netbooks, phones, TVs, tablets. SUre, UI will need to be changed for each, but not core code.

From Medfield on, a customer can go to either Intel or ARM for an Android solution with an app store. That includes Nokia.
The technology and OS components drop out of the equation in 2011 and you’re left with price, marketing and partners. Intel could, over time, dominate all three of those.

Bill

Apple appears to be setting a trend with the iPhones and now iPad. These will indeed benefit from more processing power and Intel is the primary producer of processors. I think it logically follows that Intel will get into, and be successfull, in the mobile world.

snolan

While I agree that Intel faces a huge struggle competing with ARM for mobile devices, we should all be glad they are willing to try.

1) competition is good, and that will keep everyone innovating

2) it will make for some incredibly low power consuming Atom devices for the power sippers of the world to turn into “green” x86 servers and netbooks… not as big a market as mobile devices, but running full blown Linux on a low-power device does have many advantages

david Robins

As the prices go down and the functionality goes up, Intel will face major problems. eventually the same chip family will be in everything from big servers to smart phones with very low margin. Customers win!

shreshtha

Apple repeated the history, reinvented the wheel (OS) and lately quite successful.
PowerPC is a good arch.

Marco A.

Intel keeps promising to do better than ARM in power consumption, but what they mean is that they’re trying to get a generation ahead in fab process. ARM as a CPU architecture takes half the power per computation as x86, maybe less when not pushed for performance, and there’s nothing Intel can do about it (aside from taking an ARM license). Intel can get ahead only by fabbing with transistors that are enough smaller to make up the difference.

Globalfoundries will be producing Cortex A9s in 28nm by late 2010. Intel hopes to be producing 22nm x86 chips by late 2011. Will it be enough? Only if Globalfoundries stands still. With the size of the smartphone market set to outsize that for PCs, it seems unlikely.

Intel might be able to hold off ARM smartbooks for a few more years, if only because Microsoft won’t port Windows 7 to ARM, nor will Adobe put CS5 on Ubuntu. Multicore ARM CPUs for servers running Linux will be coming soon though. That leaves Intel squeezed from both below and above. A few years down the road and Intel will be all but forced to take an ARM license (or buy ARM). They’re too smart to simply sit there and watch the business cease to exist.

Ken Jackson

Marco, that’s only partially true. And with their shrink process they’ve been able to do things like give up a little perfomance for a LOT of power efficiency.

Intel is doing other things as well. They’ve moved to in-order processors for Atom. They’ve done instruction fusing. SMT is a given. New power states and power gating. A far more power-efficient cache. And to go back to instruction fusing, with Intel’s uop architecture, they have a fair bit of tweaking they can still do.

Intel has faced stiff competition before when their architecture wasn’t supposed to be capable of cutting it and they showed that x86 was a lot more flexible than people thought. I wouldn’t count them out yet, and Moorestown is just darn impressive. And if the numbers hold, I think it is a leap over the current competition.

Marco A.

Intel keeps promising to do better than ARM in power consumption, but what they mean is that they’re trying to get a generation ahead in fab process. ARM as a CPU architecture takes half the power per computation as x86, maybe less when not pushed for performance, and there’s nothing Intel can do about it (aside from taking an ARM license). Intel can get ahead only by fabbing with transistors that are enough smaller to make up the difference.

Globalfoundries will be producing Cortex A9s in 28nm by late 2010. Intel hopes to be producing 22nm chips by late 2011. Will it be enough? Only if Globalfoundries stands still. With the size of the smartphone market set to outsize that for PCs within a year or two, that seems unlikely.

James Katt

Intel can always quickly catch up by licensing the ARM architecture again.
After all, they are pretty much all the same. Then by having the ARM blueprints, Intel can add it’s own expertise to the mix. I’m sure this would give it a leg up on other ARM lincensees.

Duh.

David Emery

Interesting analysis, but how do you account for Apple’s ability to pop out of nowhere and accomplish the kinds of system integration* that you’re implying Intel can’t pull off? I think one difference is that Intel doesn’t really compete in ‘vertically integrated markets’ like Apple or Qualcomm do, and the few times they’ve tried, Microsoft cut off their body parts.

Or am I mis-reading your arguments???

  • I know how hard this is from my day job.
Russ

Hey Om,

This definitely reminds me of the old PowerPC vs x86 arguments of years past. People could spend night and day talking about technical fine points about one architecture over the other without looking at the actual market dynamics. “Trying to reinvent the wheel” is way too simplistic a reason for why PowerPC didn’t take off. Intel just flat-out won that war – against IBM, Motorola, Digital, Apple and the rest.

The lesson learned from that round in the 90s is this: Don’t bet against Intel. They’ve been behind the technological curve a few times over the past few decades, and have come out on top for various reasons again and again – nothing you wrote in your article makes me think that this time will be any different.

I agree with you that it’s not good their “real” mobile competitor is still another year out, but so what? Mobile phones are swapped out every 16 months on average, so there’s plenty of churn for them to capitalize on later. And capitalize they will. Atom is a good architecture – they’re still playing catchup, sure, but at least they’re backing the right horse.

Honestly, it really doesn’t matter how dominant ARM is now. First, more and more devices are being run on VMs anyways (WebOS, Android’s Dalvik), and secondly (and more importantly) mobile manufacturers are insanely price sensitive. They’ll happily switch away from ARM if they can shave off a few dollars per device. We’ve seen before in Intel’s OEM war against AMD, they has ways of making that economic incentive happen.

As soon as Intel has a chip within decent spitting distance in terms of power/performance of the ARM chipset, they’re going to put on the full press (if they haven’t already) to the device manufacturers. They’ll use their economies of scale – which QCOMM, TI, Marvell and the rest still can’t match to my knowledge – to drive down prices. And then they’ll start kicking in the famous Intel marketing dollars. We’ll be hearing that familiar “Buh-bum-bum-bumm” at the end of those Verizon Droid commercials, and seeing “Intel Inside” stickers lining the windows of your nearby AT&T store.

Which brings us to the Intel brand in general. The first phones that sport an Intel-branded CPU inside are going to be consumer gold for no other reason than your average buyer knows that Intel is the company that makes them “good” CPUs. They’ve been told that for 20 years of Pentium commercials featuring Bunny Suits and Technology Rock Stars. Intel is a massive brand and a marketing powerhouse – discounting this in a category as consumer-centric as mobiles is pure folly.

Add to top of all this that Intel knows that dominating the mobile market (including the burgeoning tablet market) is a must. PC sales are going to do nothing but remain flat or even start falling over the next decade. Intel has its sites set, and I wouldn’t want to be in the way of that steam roller.

BTW, I was a big believer in QCOMM a few years ago – so much as to have my Mom buy some shares in it when she asked me. Since then I’ve seen just how well they’re really not doing. (I don’t make stock recommendations to my Mom any more). The rest of the ARM manufacturers are in the same boat – i.e they’re all little versions of AMD really. Once Intel shows them real competition, it’s going to be trouble.

Oh, as disclosure for anyone else who reads this – I work for Nokia, but have ZERO knowledge of anything they’re doing with Intel on MeeGo. No one can predict what will happen with that platform, but if Intel starts supplying chips to the manufacturer of 40% of the world’s phones in any substantial way? That’s could be a pretty big blow to ARM’s dominance as well.

Anyways, Om! I thought I’d share my thoughts. I think Apple’s A4 has you all hot and bothered, and given you some tunnel vision. Betting against Intel? Please come back from Apple land!

;-)

-Russ

Conquistador

Paul Otellini? Are you trolling tech blogs again?

charbax

“They’ll happily switch away from ARM if they can shave off a few dollars per device”

ARM prices are extremely low today. You really think Intel is interested in providing components that allow for $100 smartphones, tablets, e-readers and laptops? When I say $100, I mean $100 unlocked, out of contracts, the bill of material and manufacturing already being below $100.

charbax

If anything, Intel is going to try to convince manufacturers like Nokia to keep the prices artificially high. But once cheap competition is available to all consumers, I don’t think anyone is going to be tricked for expensive stuff anymore.

Ben Rockwookd

“Anyways, Om! I thought I’d share my thoughts. I think Apple’s A4 has you all hot and bothered, and given you some tunnel vision. Betting against Intel? Please come back from Apple land!”

Sigh… i think you should come out of Nokia land!

:-)

Ben

David Sanabria

OM,

I am continually fascinated by polymorphic nature of technology companies: today’s goat was yesterday’s prize bull.

Intel may have trouble penetrating the mobile market for the next 18 months, but its entrance seems like a foregone conclusion. The give and take between Intel, Qualcomm, Apple, TI, et al will make winners of us all in the end as both consumers and members of the species. The drive toward smaller and more mobile is changing the way we use technology and how we integrate it into our lives. It will also reduce the immediate financial cost of our habits as well as the long-term environmental cost.

Today’s futurists try to understand how current trends will produce tomorrow’s, and historians will look back and try to figure out why we made the choices that led us there. I am an enthusiast: I am loving the ride!

Om Malik

David

I agree with your assessment regarding goat vs prize ram. I would say, having observed two shifts in my life (PC and Internet), I can say new computing paradigms bring new leaders.

Mobile is a whole new ball of wax and it would bring its own set of winners and losers. You can see signs of who will be what in months/years to come.

sajal dogra

Om,

While Intel has certainly lagged in round 1, I think round 2 in 2011 will make things interesting. Here’s an interesting writeup by Anandtech: http://www.anandtech.com/show/3696/intel-unveils-moorestown-and-the-atom-z600-series-the-fastest-smartphone-processor/1

Power numbers sound competetive, and performance is great. I’d say Intel doesn’t need to do better. As long as performance/power is competetive, Intel’s process technology lead will enable it to push those lower price points, and squeeze out the higher cost providers. I won’t be surprised to see a pricing battle sometime in 2011-12, like we saw between Intel-AMD a few years back. Too many players in the mobile chips market right now..

And why should Intel be a growth stock? Given the 3% dividend yield and a forward P/E of early teens, I would call it an attractive value stock proposition.

Om Malik

Sajal

I think there are two key questions for Intel:

  1. However great that platform (Moorestown) be, it is still not as good as some of the others and lacks the important integration with mobile radio technologies.
  2. Where are its partners for this product? Most of the major handset makers (volume players) are all tied to other chip makers.

I made some of my other arguments above so I won’t repeat myself.

That said, here is a question for you regarding growth vs value stock. It is trading at 11 times 2010 earnings. Does that sound like a value stock? Love to get your thoughts here.

Ken Jackson

There’s a difference between stock performance and market performance.

That is Intel could end up dominating this new mobile space, but the stock price still trend down if the mobile space doesn’t have the margins that the desktop space had. And this is almost certainly true.

So if investors believe, “Wow Intel may actually take over this mobile space”, but they believe they’ll do it at the cost of the higher margin, near equal volume desktop space, they still may value the stock lower.

This says nothing about the technology of the company.

For Apple things are fundamentally different. They’re moving from a space where they’re a niche player to one where they may dominate. This justifies growth in the stock. But if it was MS who did the iPhone and had sales equal of Apple, I think their stock would still be flat. It’s a market that will likely cannibalize their current dominant market. The only way that MS or Intel become growth again is to find a new HUGE area that doesn’t cannibalize their existing one. For example, imagine home robotics, or something like that.

Ken Jackson

Good article, although I think you underestimate how good Moorestown is. And the fact that Intel actually still has headroom with Atom in both power and performance. Probably a fair bit more than ARM does, and no one, not TI, Samsung, AMD, NVidia, has shrink process technology that Intel has.

And given that Android and WP7 are both easy to move to x86 (the WP7 emulator is the actual WP7 OS running on x86), if Intel can show on real designs that Moorestown has the numbers they showed yesterday then I think we’ll start to see Android and WP7 devices on it (although I know Intel apparently isn’t happy with how MS architected WP7).

I also wouldn’t count MS out either. They’re being pretty smart not cannibalizing Windows, but I think they’re getting closer to realizing that they’ll need to (although no use losing $10B in profit any earlier than one needs to).

pk de cville

If the link is credible, I wouldn’t be surprised if someone bought them.

Ken Jackson

Global is good, but I think virtually everyone considers their foundries as less refined than Intels. One of Intels strengths is moving more on-chip with good shrink and yield. Moorestown is a perfect example of the strategy.

And based on what I’ve seen they have a fair bit more they can do… which is pretty exciting for everyone I think.

ARM vs x86 is a lot more fun than the old x86 vs x86 world.

Tom Foremski

The smartphone/mobile phone market doesn’t want to be commoditized as happened in the PC market with Intel and Microsoft aggregating all the value, reporting 60% plus profit margins, while the PC makers and everyone else had to make do with single digit profit margins.

Also, Intel has won the war in that it has 80% of global microprocessor markets. Even if it did well in smartphones and tablets, it wouldn’t mean a whole lot of new revenue.

Intel’s future growth is tied to global economic growth rather than fairly niche markets. That’s why the stock market yawns when Intel makes these type of announcements.

Om Malik

Tom

The future of technology is tied into the mobile cycle and not as much as the PC cycle, The global economic growth is more evenly represented in the mobile cycle and that is where all the future is.

Bill Kircos

Om – this is an intriguing read, but has some flawed logic. Virtually no other industry grew in 2009 – many shrunk significantly. PCs bucked that trend and grew, and are projected to grow this year – led by mobile laptops – by 20%. Most envy that kind of growth, and the ‘death of the PC’ has been proclaimed hundreds of times the last 2+ decades. Seems to me everything is becoming more PC-like. Are people using i-phones as cell phones or more like mini-computers? Also, Intel has ambitions in multiple adjacent-PC areas, not just phones: 100M+ netbooks to date; TVs, cars, tablets, even fishing and golf cart companies are calling us. Sure, many of these areas are still a WIP, and yes, you can argue we’re behind. But on yesterday’s Atom news, we went out of our way to emphasize an historic and eye-opening drop in power – up to 50X – something most said was impossible. I don’t think we did a bunch of handwaving about winning or why INTC should skyrocket. (Disclosure – I work for Intel, and doth protest too much).

Om Malik

Bill

At no point in my piece am I making a case that PC is dead and INtel looses domination over the PC ecosystem. I can debate that PC-like and computing everywhere are two distinct phenomenon, but let’s leave it for another day.

The point I am making:

  1. Intel isn’t going anywhere in fast growing new markets.
  2. Growth investors bet on the future.
  3. Hence Intel stock is stagnating.

Beyond that, my contention is that the way Intel is structured, the company isn’t capable of living with thinner margins that come with a competitive marketplace. That is one of the reasons why it hasn’t quite succeeded in its expansion into newer markets.

Now to your remark about “power saving.” I don’t take anything away from you guys, except that the ARM ecosystem is still miles ahead of your employer.

Jay Garantz

Looking to the past year, Intel’s stock has almost doubled. Looking at it that way, the market has spoken and seems to disagree with you. Also, many would say Google is a growth stock, and yet it is trading 30% below the all time high it hit in 2007. Is the market saying that growth investors shouldn’t bet on Google?

With much respect, my point is:

  1. Intel isn’t really a “growth” stock and hasn’t been for some time. It’s in the Dow 30 – not exactly a growth stock hub.
  2. Computers larger than smartphones (netbooks, laptops, tablets, servers, etc…) will still be sold in massive quantities (40 million+ netbooks this year in the U.S. alone – Intel powered) for at least a few more years, if not decades.
  3. Intel will be fine and at some point will have a decent share of the smartphone chip market, which will augment the dominance they have in the netbook and laptop market.
Zac Schellhardt

Good read Om, but it seems you’re using a conclusion as your first point. There are two real observations you can make (with different confidences):

  1. Intel stock is (relative to growth markets) stagnant; this is an objective observation as much as it can be, but is still a bit of an apples/oranges comparison.
  2. Growth investors bet on the future; while this is true, you seem to be inferring that investors are always correct.

Based on those two observations, you can conclude that maybe Intel isn’t going anywhere in growth markets. There are obviously other factors–comparing growth between Intel and ARM doesn’t really work since Intel’s portfolio is much larger than ARM’s and subject to many different market volatilities.

The argument’s not definitive, but you’re right in pointing it out, since it’s certainly a potential (maybe probable) outcome for Intel.

WaltFrench

@Bill Kircos says, “an … eye-opening drop in power – up to 50X…”

As a non-engineer (mostly, a mathematician), I’m in the dark about how you could use fifty times less power. Fifty times what?

Do you mean 98% less? (I.e., that in nap mode the predecessor processor uses 50 times more than the latest?)

Could you throw out real data— a number dropping by 50 times to another number?

“Inquiring Minds Want to Know!” ®

IC Engineer

If I interpret correctly, 50X is the power reduction by using power switches in their chip. If Intel left their chip powered-on all the time it would consume 50-times more power (in idle mode) compared to using power switches that can be turned off (in idle mode). This does not mean that battery-life will improve by 50X because battery life is still dominated by active power consumption (video play back/record, web browsing, emails, WiFi etc.). The 50X number is a “marketing” strategy — it does not mean anything from a battery life point of view. The overall improvement in battery life is going to be a tiny single digit percentage!

pk de cville

Om,

I’ve seen Intel’s predicament your way for some time now. And I have a question for you…

Is there anything stopping Intel from licensing ARM architecture? I believe Intel could use its lead (12 to 24 months?) in fab tech and massive fab resources to deliver the fastest and most power efficient ARM based chips.

Wouldn’t it eventually dominate the market under this strategy?

When will we hear Intel’s intro of Intel fabbed ARM chips?

Thanks,

Om Malik

Nothing stopping them from licensing ARM except their own hubris and frankly even if they do, it would take them time to compete in the market place.

I think it is interesting because in a competitive marketplace, Intel can’t compete. It has the same issues as Microsoft — I guess muscle memory from being a monopoly does that to you.

Curtis

@Om,

You’re overlooking a very critical area in your response to pk de cville, that is Qualcomm’s patent portfolio which principally covers cpu’s, soc’s, and radios for the ARM architecture. I do believe that this same patent portfolio accounts for the majority of Qualcomm’s revenues. That said, Intel can certainly license ARM technology, but all that does is enable Intel to feed $$$ to their new “frienemy”.

My $.02,

Best.

jst

Intel does NOT have the same issues as Microsoft on any scale. You sound off like the sheep that are being herded off the cliff.
All ANTI-Microsoft, Anti-Intel. Regurgitating the same nonsense that people have been throwing up for the last twenty years. If you believe that Intel can’t compete on any scale now or in the future you truly are biased and shouldn’t even have a forum to spew your garbage.

gbp

Good analysis. PowerPC and Alpha were expensive compared to Intel during that time. Intel was not a big company back then. Intel can still comeback if they can buy Marvell which they sold their ARM based processors to.

Om Malik

Marvell-buy would be intriguing but something tells me that it would be a whole new can of worms, especially in this era of overactive FTC.

Yuvamani

Intel’s XScale sale looks like total crap right now, doesnt it :)

The reason they sold xscale (which btw was one of the best arm implementations of that day) was that:
1) Margins were low
2) CE guys did not want an expensive ARM. And their devices did not need the processing power to justify the cost intel wanted to charge
3) Intel tried to innovate by adding instructions to arm which nobody wanted to adopt because well the ARM instruction set was brilliantly universal. Thus no differentiation. The XScale did find itself in a few Treos which were the cutting edge of that generation.

Charles

Great article. Intel is trying to shove a square peg into a round hole. . . and they keep trying this over and over figuring that if they throw enough money and resources at it they will win.

ARM chips are making there way into servers which will bring a great deal of R&D into ARM chips pushing them further along at a faster pace, thus reaching speeds that most users will find perfectly acceptable across a wide range of produce, phones, notebooks, and even desktops. The one hurdle they have left is getting MS on board since they still hold the dominant OS in the world that works only on x86 architecture.

However, GPUs are bridging the gap in every direction. That is a very interesting area especially when considering most users needs of high demand cpu usage comes at the expense of media. . .

Personally I can’t wait for high performance tablets, notebooks & desktops that don’t need a fan and run the same OS :D

Hamranhansenhansen

You don’t have to wait. Apple systems either don’t have a fan or the fan runs very rarely, and even then are tuned to be quiet, and they all run the same OS. Apple’s systems also have H.264 decoders in hardware, so video doesn’t run on the CPU or cause the fans to start.

layman_user

This is a wrong assumption !!! All apple notebooks including the Macbook Air/pro have a fan. And the Apple systems are one of the hottest (in terms of temperature) out there in the market. And if ARM is able to enter Notebooks and Servers, there is no way they can acheive that without bumping up the power.

WaltFrench

@Charles — Strict literalist answer: iFixit couldn’t find a fan in the iPad; waiting not required.

@Hamranhansenhanser— New MacBook Pro here. In ~ 2 weeks, I still haven’t noticed the fan running. Been using it about 10 hrs/day in a private office, apartment with no TV and occasional flights (where I wouldn’t hear it).

Nor does it get actually hot for me, in usage patterns that made its predecessor too hot for my lap. Warm, yes.

@layman_user— go back to that hit job done by “PC Authority,” & you’ll see they compared it by (a) maxing out running a GPU-intensive test, (b) versus a laptop with a weak GPU (on which nobody would run Cinema4D) that burns 30 watts less power than the one in the MBP, then (c) using Windows to bypass the MacOS power management.

I imagine the ARM advantage is not just the instruction set, rather, that Intel has put its engineering resources into pushing the envelope, where they can command top dollar, versus low-power management, irrelevant in a gamer’s box with a 300W power supply. Now, it’s easier to shrink the die and turn up the clock for throughput at same power levels, than to shed functionality while somehow keeping X86 compatibility and fine-tune for a market where you don’t currently have any volume.

Too bad about that XScale choice. But hey, IBM is still making a fair amount of money on mainframes and consulting. Just lousy growth.

Brian S Hall

Excellent analysis, Om.
One of my core strategies for success going forward is boiled down to:

the smartphone is the computer!
The PC is a dinosaur.

This is why I have such a low opinion of Microsoft. They want to maintain their PC-gatekeeper status and maintain those fat Office packaged software margins. Given the cost structure facing Intel, however, and their level of competition as you’ve listed, it looks like they could be potentially even worse off.

Kevin C. Tofel

No popular mobile OS currently runs on x86 architecture and I don’t foresee MeeGo leap-frogging past iPhone, BlackBerry, et al.
Building effective hardware for a smartphone is great, but if there isn’t an OS that’s popular with devs and consumers, it won’t have a chance. Put another way: the benefit of x86 compatibility is lost on smartphones.

That means the Android port to x86 allegedly in progress is Intel’s best shot to get into smartphones. I think the only company that can save Intel in this space is Google as a result.

Hamranhansenhansen

No popular mobile OS currently runs on x86 architecture

OS X does, the core operating system under iPhone OS. And iPhone OS apps are all developed on Intel and compiled for Intel to run in the desktop iPhone and iPad simulators.

If Intel gave Apple a reason to use Intel chips in iPad, iPhone, and iPod, then Apple could do it in less than a year, just as they jumped from PowerPC to Intel in less than a year.

But can Intel give them a reason?

tofino

would that be one more reason for apple to disallow adobe’s cs flash exporter on the iphone? if they ever switched from arm, they would have to wait for adobe to get around to updating it. and wait. and wait.

MySchizoBuddy

Apple has already made up it’s mind and going full steam ahead with ARM by buying P.A.Semi and Intrisity. Apple wants complete control of its roadmap, design chips specially for its own products and being able to differentiate from competition. The competition is buying of the rack parts while Apple wants to custom design them.

Kevin C. Tofel

Good point on OS X as the core of iPhone OS. But taking your argument one step ahead, why would Apple move iPhone OS to x86 when they can design their own ARM chips and further control the hardware / software / user experience package?

MB

@kevin You are absolutely right. Apple realizes that mobile computing is the future. They also realize that being able to build their own cpu that is customized and optimized to work with iPhone OS gives them the advantage they need against the competition.

The A4 processor has been an integral part of the iPad’s success. The competition can’t build a slate device that can match the iPad’s dimensions, weight, lightening fast processing and still achieve 10+ hours of battery life without a customized cpu optimized to work with their mobile os.

snappy

tofino, that is exactly the reason. Sure there are other benefits from using xcode, apple will do all the heavy lifting is they need to do the switch, that way they are not as dependent on adobe or other large shops to get their butt’s in gear. Irony here is that the only ones really loosing here is adobe. if apple has to change for some reason then all those dev’s have to wait for adobe to rebuild / test and optimize. least not wait for any of then enhancement apple brings out every now and then.. how come people not see the irony in all the ramblings of the adobe devs of this. reminds me of an old saying my mom use to say when i complained about things as a kid. “you know you are cutting your knows to spite your face right?”

Comments are closed.