More about the Silicon Valley legend

My friend Shane Greenstein this week wrote a book review of Moore’s Law, a biography of the famous Silicon Valley chemist, entrepreneur, billionaire, philanthropist and visionary. As Shane opened his Wall Street Journal commentary:

Fifty years ago, Gordon Moore formulated his famous “law,” typically summarized as “the number of transistors that can be placed on an integrated circuit will double every two years.” The accumulation of exponential improvements that he foresaw has indeed ushered in perpetual reductions in the cost of computing and the size of computers. And it’s at the core of the information technology revolution.

Even before his famous 1965 paper, Moore was a Silicon Valley pioneer, there when they put the silicon in Silicon Valley. After attending SJSU, Cal and Caltech, in 1956 Moore went to work for (later Nobelist) Bill Shockley, working on the technical challenges of adapting silicon (rather than germanium) to construct reliable semiconductor devices.

The next year, Moore and others of the Traitorous Eight set the pattern for the Valley when they Shockley Semiconductor to found Fairchild Semiconductor — the Valley’s first (unsanctioned) spinoff. In 1968, two of the eight — Moore and Robert Noyce — took Andy Grove to start Intel. (Four years later, fellow Traitor Gene Kleiner joined Tom Perkins to found the Valley’s legendary VC firm)..

Never fired and married to the same woman for 60+ years, Moore was neither a showman nor celebrity like the late Steve Jobs. Arguably, he and his compatriots did more than anyone else to set the pace of Silicon Valley (at least during the five decades when there was still silicon in Silicon Valley).

The authors of this latest biography — “a chemist, a historian and a journalist” — offer the definitive story of the 86-year-old chemist. As Shane concludes:

The authors have material on just about everything in Mr. Moore’s life—his relationships with his wife and children, the Porsche he drove, even his childhood adventures with nitroglycerin. The detail, on occasion, becomes overwhelming.

Yet the book brings an insider’s perspective into the discussion of Moore’s law. What became the law first emerged in 1965, in an article modestly titled “Cramming More Components Onto Integrated Circuits,” published in the journal Electronics. By the authors’ account, nobody paid much attention to it at the time. Why do we know it today? Mr. Moore revisited the idea in 1975, updated it and devoted public speeches to it. Others began to notice its deep foundations at the boundaries of science and production; Carver Mead, a longtime professor at Caltech, was the one who actually coined the term “Moore’s Law.”

Gordon Moore’s forecast was spectacularly right. Yet, as this compelling biography proves, even if he had never hazarded it, he would remain a legend in Silicon Valley.

Alas, it’s hard to deny the passing of this era. Moore is the only one of the Intel founders still alive, and one of only two (with Jay Last) of the Fairchild founders. Despite its continuing microprocessor monopoly, nearly 15 years later Intel stock is still only half its peak level of August 2000. Internet software — not electronics — is where Kleiner Perkins is making its money nowadays.

Microsoft gains powerful Windragon ally

Once the co-owner of the powerful “Wintel” monpoly, in the past decade, Microsoft has been suffering an increasing slide towards irrelevance. The onetime brash PC pioneer has looked more and more like a legacy software company protecting an installed base.

Nowhere is this trend more painful than in the mobile world. Since 2006, it has never held more than 15% of the global smartphone market and in fact — under the twin onslaught of iPhone and Android — its market share has been in a freefall, giving up more than three-fourths of that share. To add insult to injury, 2011 was the year when more smartphones shipped than PCs, a trend that’s only going in one direction.

To save its mobile strategy, Microsoft’s put all its money on its Nokia alliance in hopes that would save Windows Phone 7 from the same ignominy as its predecessors. Given that Nokia’s share has also been in freefall, this has the potential of extending Gary Hamel’s “two drunks” analogy from acquisitions to joint ventures.

On Tuesday, Microsoft landed a public endorsement from one of the mobile industry’s most powerful players: Qualcomm. To me, this has the potential to be the most powerful ally that Microsoft could attract, as a net positive for both firms and a negative for their respective rivals, Apple and Intel (and to some degree, Google).

At the company’s annual shareholders’ meeting, Qualcomm CEO Paul Jacobs bragged a little about the company’s successful shift of its semiconductor business from radio modems (“basebad processors”) to its Snapdragon all-in-one process — which include an ARM-compatible CPU, radio, graphics, multimedia and other features. Based on its Snapdragon success, Qualcomm has continued to gain market share, passing TI last year to become the leading supplier of cellphone CPUs (“application processors”), according to Strategy Analytics.

(By not making its own cellphones or infrastructure, Qualcomm also had a unique position in the mobile phone value chain. Instead, it devotes its $3 billion/year R&D budget to developing components and basic technologies that it sells to all comers among the various handset and platform providers).

Qualcomm has two reasons to throw its weight behind Microsoft. First — unlike Android — it’s only game in town, as the sole CPU supplier for WIndows Phone products thus far. Secondly, Qualcomm sees Windows-on-ARM as opening a whole new market — allowing it (and other ARM licensees) to take CPU market share away from Intel in the PC and tablet space.

In Jacobs’ view, the Windows-on-Snapdragon (or — as I put it, “Windragon”) combination will merge the office productivity options of Windows with the mobility, power saving and performance of the Snapdragon CPU line. Part of the shareholder meeting was devoted to demonstrating the Snapdragon S4 processor designed for the Windragon market.

This sort of tight alliance would seem to me to accelerate the estrangement of Apple and Qualcomm in the handset world. A shareholder asked when (or why) Qualcomm can’t provide more chips to Apple, which appears to be limited to its baseband processor (in the iPhone 4 and iPhone 4S) to complement Apple’s own A5 ARM cpu. If Qualcomm is going to help Microsoft gain share against the iPhone, iPad and perhaps reverse the OS X gains against Windows, then the Apple of Steve Jobs (and perhaps Tim Cook) is not likely to be too friendly or dependent on the San Diego chip giant.

Certainly success of Windragon would accelerate the shift from Mac to iPad sales and Apple’s expected phase-out of OS X.

The other interesting implication is what it does to Intel. There’s no love lost between the firms, particularly in the WiMax vs. LTE 4G wars (that Jacobs alluded to Tuesday). At the same time, Intel is suffering from the same shift away from PCs that threatens Microsoft.

Will Intel make ARM-compatible CPUs, as it once did? Will it try to more aggressively win Android allies to counter the Microsoft-Nokia-Qualcomm alliance?

Carly reconsidered (III): Craig Barrett's testimonial

The editorial page of this morning’s Merc had an interesting business/political tidbit: a glowing testimonial to former HP CEO Carly Fiorina from former Intel CEO Craig Barrett. The dead tree headline was

Fiorina was architect of Hewlett-Packard’s success

while the online version says “History straightens out facts: Carly Fiorina positioned HP for success.” A few excerpts:

Under Carly Fiorina, Hewlett-Packard recognized that the computer industry needed consolidation and engineered the largest high-tech merger in history, combining HP and Compaq. There were plenty of skeptics to the bold actions taken by HP. But history has a way of straightening out the facts and the noted opinions of outside experts.
…
Carly Fiorina, the architect of the HP-Compaq merger and now a candidate for U.S. Senate, deserves great credit for her actions while CEO of HP. She understood the challenges of the marketplace, the dangers of the status quo, and the need for companies to move forward with bold actions to ensure their success.

As CEO of Intel at the time, I remember watching those plans unfold with more than casual interest, as HP and Compaq were two of Intel's largest customers. It was a ringside seat to an industry-changing event. It wasn't always pretty, but it was carefully planned and well-executed, and the bottom-line result was exactly on target.

As one of those skeptical “outside experts,” I’ve freely admitted that Carly was right and I was wrong. Barrett’s comments also fit what current HP employees say about her role in setting up the implementation being overseen by Mark Hurd.

Barrett concludes with a political defense of Fiorina, in (futile) hopes of inoculating her against the mud that Barbara Boxer allies will dredge up if she survives the primary.

In the current political campaign, many accusations have been leveled against her and her tenure at HP. As someone who knows the industry well and was there watching all the details, I have to respectfully disagree.

Carly Fiorina, who started her career as a receptionist and rose to be the first woman to run a Fortune 20 company, tackled the most difficult issues and brought exactly the right approach to a company that would have faced a more uncertain future otherwise.

I found the column remarkable, because normally Intel CEOs have commented on things important to Intel, not its customers or partners. Apparently Barrett got to know Fiorina well during their overlapping tenures: Fiorina was HP CEO from 1999-2005, while Barrett was Intel CEO from 1998-2005, succeeding Andy Grove.

While Barrett is entitled to his opinion, I would take issue with one claim: “HP has always been and still is an innovative company bringing great products into the market.” The New HP (both Fiorina and Hurd) is about cost-cutting, including cutting materials costs (ala Toyota) in hopes of reducing costs without noticeably reducing quality.

News headlines also implied that Hurd cut the HP Labs from 700 employees in 2005 to 500 employees in 2009. This makes sense if (as I concluded in February), Fiorina-Hurd decided that HP’s future lay as an efficient commodity IT producer rather than a high margin differentiated one — but it’s hard to argue that HP is as innovative as it was 20 years ago.

MeeGo: Nokia and Intel learn to let go

For years, I was trying to figure out why the world needed two Linux-based tablet operating systems: Nokia’s Maemo and Intel’s Moblin. Almost three years ago, Intel created Moblin as a fork of Maemo and its Hildon UI (inherited from Symbian).

The reality was that the world didn’t need a forked niche mobile platform, but of course forking is the reality of open source. The changes may be available, but when not invented here and an unwillingness to share control get involved, corporate egos trump the nominal openness of an open source license.

The two factions have been flirting with cooperation but had been unwilling to consummate the deal. On Monday, they finally did.

In an announcement at the European cellphone industry’s global big tradeshow — Mobile World Congress — the two factions announced plans to merge the two code bases. As Nokia’s open source guru reported in his blog:

We’ve been busy with our friends @ Intel.

We decided to expand the relationship we started already last spring. We merge Maemo and Moblin projects into one single project called MeeGo. MeeGo is an open software platform – an operating system – for a wide range of devices. It’ll run on X86 and on Arm based hardware. It will be developed as an open project hosted by the Linux Foundation.

So what does it mean? Many things.

Joint development
We will merge Maemo and Moblin projects. Their architecture is already very similar. They share many components but sometimes use different versions. But they build and integrate releases independently. And while Maemo is for ARM, Moblin is for X86. Now we merge them to get the best of both. A good Moblin build and integration, Maemo’s mobile optimizations and ARM support, Qt etc. We can also now make the bright engineers of Intel and Nokia to work close together.

Of course, the announcement is less about saving R&D engineers and more about combining installed bases, APIs, ecosystems, and third party developers. Neither platform was very interesting on its own, but together the hope is Meego will be the future for both ARM and x86 devices. Both will use Nokia’s QT and Nokia will peddle applications via its (struggling) Ovi store.

More importantly, both Nokia and Intel let go, and turned control over to a neutral broker — the Linux Foundation. LF (like its predecessors OSDL and FSG) has plenty of experience contending with giant corporate egos.

The one thing that seems ambiguous is the positioning vs. the world’s most famous (quasi) Linux mobile platform, i.e. Android. Are sponsors being coy about avoiding comparisons to Android, or is this really an up-market alternative, between handsets and PCs? In an era of iPads and netbooks, will such distinctions remain a year or two from now?

Hat tip: tweeter David Wood, live from Mobile World Congress.

New York adopts an industrial policy

I didn’t get why the NY attorney general is suing Intel, other than it’s the same publicity-seeking path his predecessor to get elected governor. Antitrust enforcement is a Federal issue, whether for the US DoJ/FTC/FCC or the Eurocrats in Brussels.

However, blogger Geoffrey Manne suggests that NY has an interest in hurting Intel and helping AMD, given that AMD is talking about building a $3 billion plan in upstate New York. For some reason, I thought the idea of Federalism and the Constitution was to prevent inter-state trade wars, but I guess both have gone out the window along with original intent.

Like most lawyer-politicians, the AG’s understanding of business and economics is dubious at best. FT’s Lex aptly summarized the likely effect:

Share prices for Intel and competitor Advanced Micro Devices barely reacted. The problem for AMD, which is set to face Intel in a Delaware courtroom in March, is that legal victories offer only consolation, and perhaps the chance of a pay-out to help pay down debt. The period when the company had a clear technological advantage and opportunity to make a dent in Intel’s market share of about 70 per cent has passed.
…
[E]ven if, as alleged, Intel is shown to have forced customers to guarantee market share levels in return for cash rebates, the structure of the industry will probably remain unchanged.

The benefits of high supplier power

From today’s Merc:

At the heart of the European antitrust case against Intel are a series of once-confidential memos that provide a rare glimpse into the intense and at times acrimonious negotiations between the world's biggest chip maker and Hewlett-Packard, the world's biggest computer company.

Even a company as powerful as HP, the documents suggest, had to make business decisions based on how Intel would react. The Palo Alto computer company reportedly was so worried about threats from Intel that it dramatically scaled back its plans earlier this decade to sell computers powered by chips from Intel competitor Advanced Micro Devices of Sunnyvale.

And when AMD offered HP 1 million free microprocessors, saying it was a gift "no reasonable business partner could refuse," HP took only 160,000 to avoid Intel retaliation, according to the report unveiled last week by the European Commission.
…
HP declined a Mercury News request to be interviewed and Intel insists it did nothing illegal, calling the commission's findings "wrong as a matter of fact, law, economics and elementary fairness."

Of course, HP is not the world’s largest computer company — which has remain unchanged since the System 360 in 1963 (if not before). But it is the largest PC maker, and the rest of the story is pretty convincing.

Still, the allegations suggest that our local chip designer (no longer making chips here) is as predatory as Microsoft under Bill Gates.

Nokia platform proliferation

Nokia made two major extensions this week to its platform strategies — one in handsets, one in infrastructure. Both are about finding growth in the face of increasing commodization.

For more than a year, I’ve wondered when Nokia’s Maemo and Intel’s Moblin were going to tie the knot. With Tuesday’s announcement, it’s clear that neither side has made an exclusive commitment, but both parties seem inclined towards something more than a dalliance.

Both Maemo and Moblin are Linux-based platforms for mobile devices such as tablets or netbooks. For Nokia this more about netbooks than being about phones. For Intel, it’s a chance to break into the phone segment, which it once tried (and failed) with XScale (which 3 years ago Intel sold to Marvell).

Nokia still has both the highest volume and broadest product line of any cellphone maker, as well as unparalleled global distribution through telecoms operators. Nokia is not going to share control with Google in the commodity (i.e. Open) Handset Alliance, so combining the two similar Linux platforms is a way for it to attract at least a few allies going forward.

However, this muddies the Nokia platform strategy considerably. Unlike its rivals, Nokia had a fairly focused platform strategy with the S40 and S60. Going beyond the limited Maemo experiments could muddy its platform waters considerably. What does it mean for Symbian, the technology it bought last year and is now open sourcing? LinuxDevices is pro-Linux, anti-Symbian, and now assumes Linux will displace Symbian:

The partnership news further suggests that the rumors that Nokia is moving forward with Linux -- and not, it seems, Symbian -- devices that combine MID and smartphone characteristics, are true. It also appears that Nokia will likely focus on Linux for its future high-end smartphones, while leaving a soon to be open-sourced Symbian to handle less powerful smartphones and feature phones.

I think this is wishful thinking on the part of a pro-Linux analyst. Unless Intel and Nokia plan on joining LiMo or Android, the world isn’t ready for yet another Linux-based handset platform.

Instead, the Nintel (Innokia?) alliance will be growing the segment in between smartphones and laptops, which will be distinct devices from either one. Will they make phone calls? Yes, but using earphones and not by putting it to your ear.

In this regard, they are aimed more at the low end of the laptop segment — and thus at Microsoft — rather than the high end of the smartphone segment. Presumably an alliance with Oracle’s OpenOffice.org will be necessary to support Microsoft Office documents needed by users of these devices. (This may also be aimed at Apple, which has both smartphone and laptop products and is rumored to be working on a tablet-sized iPhone).

Intel hopes this will head off the predicted shift of netbooks from Atom to Netbook. I am not sure that I agree that it will work, since Qualcomm is working hard with Android to extend its ARM-based processors into the Netbook territory, and TI is also working hard to move upmarket.

The other major announcement was that Nokia Siemens Networks is buying Nortel’s CDMA and LTE infrastructure unit for $650 million, less than the $850m it offered five months earlier for a slightly larger product portfolio. This is a consolidation of a fragmented industry that delivers North American market share to NSN, mainly Nortel’s relationships with Verizon and Sprint

It’s not clear how NSN really benefits from the Nortel acquisition, given that the CDMA business is facing end of life (even if existing customers like Sprint desperately want an support path). Although LTE is the growth path for Verizon (not Sprint), Nortel’s LTE efforts never got very far before Nortel gave up. NSN gets some sales contacts as well as 2,500+ Nortel workers, but with it new platforms that it needs to maintain — not necessarily conducive to achieving scale economies.

It appears as though this is aimed at outlasting Alcatel (which now owns Lucent’s CDMA business) and Ericsson in the war of commoditization against the Huawei and ZTE. Based on an earlier interview, telecoms.com reported:

In a recent interview with telecoms.com, Tarek A. Robbiati, the chief executive officer of CSL, Hong Kong’s first-placed mobile carrier, predicted that Chinese vendors will come to dominate the global mobile infrastructure market. “Further consolidation will come in the next three to five years. In the end there will be only three [infrastructure vendors] left, and two of them will be Chinese. The European vendors are just too slow,” he said.

The chauvinism of a Hong Kong operator against European manufacturers would be expected, but Robbiati had a Euro-centric career since graduating from London Business School in 1996 — including work as an equity analyst for Lehman and a finance VP at Orange — before moving to Australia in 2005 and HK in 2007. So if he’s frustrated with European vendors, it’s the voice of experience.

Another bad week for WiMax

Lots of WiMax news today:

• The only national WiMax carrier, Clearwire, has doubled its national footprint, going from one city (Baltimore) to two (add Portland). Clearwire has ambitious (I’d say unrealistic) goals to win 425,000 subscribers, more than half of about 750,000 households in the Portland metropolitan area.
• Intel wrote off $950 million of its $1.6 billion investment in ClearWire. CLWR shares dropped 74% during calendar year 2008. This also suggests that Intel’s $300 million in WiMax advertising is probably worthless, too.
• Nokia dropped its only WiMax device, a WiMax version of its N810 tablet, to shift all of its 4G focus to LTE.

When ClearWire was preparing its IPO near two years ago I concluded:

As with other such businesses, I can see three outcomes:

1. Clearwire gets enough money from the IPO to grow big;
2. Clearwire does well enough to get gobbled up by one of the big boys (a successful strategy recommended by my old mentor Charlie Jackson); or
3. Clearwire runs out of money and dies.

I think we can rule out option #2 — Sprint got out of the WiMax business by handing it to Clearwire, and it’s got big enough problems of its own that it can’t bail anyone else out. The other US carriers aren’t interested in WiMax because they’re all going with LTE. Foreign carriers are even less interested in WiMax.

With two of its key strategic investors pulling back, a few analysts speculate Google that will add to its existing $500 million investment in Clearwire (but no one is predicting a complete bailout). I can’t see why Google would want to do this.

The chances of #1 don’t look very good either. As of its Sept. 30 balance sheet, ClearWire has less than $400m in liquid assets on its books. Verizon Wireless is on track for capital spending of about $10 billion just in 2008 — and (for those of you who don’t watch American TV) it already has a large nationwide network.

So absent a tremendous positive cashflow from its existing (or near-term) markets — or a new round of debt or equity — I don’t see how ClearWire is going to build a nationwide footprint. And if there are doubts about the prospects of a nationwide network, business travelers aren’t going to sign up for the existing and forthcoming cities.

Eight months ago, I said: “WiMax as technology is done, finish, finito — stick a fork in it.” After this week, it seems as though Nokia and maybe even Intel agree.

In those eight months, the availability of capital and the willingness of consumers and businesses to spend money on technology has dropped dramatically. So I’m even more pessimistic about WiMax than I was before.

Update Friday 7pm: Clearwire contacted me to point out that it received a $3+ billion investment in December, which I discuss in my Friday posting.

Netbooks: the future of PCs

In my in-laws’ Sunday paper this morning, Fry’s advertised a Windows notebook computer for $300, their “everyday low price” albeit with “no rainchecks.” (Perhaps they now have better prices after suing a former exec for defaulting on $10m in loans to cover gambling losses). The same computer is $327 at Amazon.

While this particular model of the Acer Aspire One has only an 8gb flash memory drive, it otherwise has the recommended features: 8.9" screen, 1gb of RAM and Windows XP. It certainly makes me feel like a dolt for spending $150 to (as long planned) put Windows XP on my MacBook Air instead of buying a new $300 netbook.

There is no denying that the netbook boomlet — started in Fall 2007 by the Asus Eee PC — was one of the biggest computer stories of 2008. The 1 million units sold in 2007 have reached 13 million this year as both Taiwanese ODMs (like Acer) and major branded PC makers (Dell, HP, Lenovo) have jumped into the fray.

As it turns out, the netbook was also the subject of one of the better student projects earlier this month from my MBA technology strategy class. The four students started out asking whether netbooks are betwixt and between smartphones on the low end, and laptops on the high end. Their conclusion was that the netbook looks a lot like a Clayton Christensen “disruptive innovation” for the existing laptop market. I am inclined to agree (which is why they got a good grade).

I draw four conclusions from the rapid success of the netbooks.

First, Intel’s low-cost, power-saving Atom microprocessor has been both a success and a failure. It’s a success in that it’s becoming the popular choice for the various netbooks. However, with the exception of a (numerically) few Linux servers, Intel is still basically a one trick pony: selling microprocessors to run Windows. Their efforts to use the Atom to create an entire new category of device, the “mobile Internet device,” is instead commoditizing its main source of product growth, laptops.

Second, the fact that netbooks are Windows machines raises doubts about whether any new computing platform can be established, at least in the next decade. Sure, a wide ranges of vendors offer netbooks with Linux but 90% of buyers prefer XP. In retrospect, Palm never stood a chance in its hopes to establish the Linux-based Foleo on its own and so was right to pull the plug.

Third, within 18 months — if not Fall 2009 then Fall 2010 — netbooks will take over the US college market, becoming the standard computer for most entering freshmen. (In the rest of the world, it will be netbooks vs smartphones and I can’t predict the winner yet). For students, netbooks have no disadvantages and many advantages. Students don't buy CDs or software; except for PC gamers, they don’t need high performance machines. They walk all day across campuses, long distances, and thus need to lighten their backpacks as much as possible. They are cash poor, and some carry laptops in urban campuses where robbery is a risk. And unlike us geezers, they have good eyes, and thus can read smaller but higher resolution LCD screens.

Given all this, Apple needs to respond to the netbook threat. Yes, my MacBook Air is small, light, with a low-power processor, small hard disk and no DVD drive — but it’s $1600, not $300-400. There have been rumors of an Apple netbook in October and November, but they are so vague to make it clear that no announcement is coming at next month’s Macworld Expo.

While I think Apple will respond (and probably in 2009), I think the response will be different than expected; I believe today’s predictions will prove as inaccurate as the predictions of the “Mac tablet” (which proved to be the iPod Touch). For 20 years, Apple has been loathe to cannibalize its main franchise, and I don’t think it will cannibalize its 13" MacBook (or MacBook Air) sales to respond to netbooks.

Instead of a stripped MacBook, I think it more likely that Apple will offer a scaled up iPod Touch or iPhone: adding a keyboard and increasing from the 3.5" 480x320 screen to either netbook resolution (9", 1024x600) or perhaps even an HDTV-compatible 720p (1280x720). Unlike the Foleo, such an iPod Touchbook would start life with more than 10,000 applications.

The only question in my mind is whether the device will be a GSM phone or merely a Wi-Fi-connected mobile Internet device. I think requiring an AT&T contract would put Apple at a huge disadvantage, so my money would be on the Wi-Fi device (or both).

ARM'd to the teeth

Last month, Intel said it wants to own mobile phones the way it does PCs. The only obstacle along the way is dominant position of ARM-licensed CPUs in the mobile phone market.

In tomorrow’s Financial Times, the CEO of ARM “sounded a note of defiance”:

“Who is to say it will be Intel taking market share from ARM and not the other way?” said Mr East.

“Intel will probably always be able to make a microprocessor that runs faster, but ARM can do one that uses less power. We are still a long way ahead in that.”

As part of its counter-attack, ARM claims it can leverage the efficiency of the ARM chips to save power in server farms.

However, the FT article makes the words of the British CEO sound like bluster:

Mr East insists that ARM, whose chip designs power 98 per cent of the world’s mobile phones, has little to fear. However, the company’s heavy focus on the “Intel issue” at its recent analyst day in London is seen as a sign of growing concern.

“ARM is clearly paranoid about Intel,” said Robert Sanders, an analyst at Dresdner Kleinwort. “It is a very well resourced company and could easily catch up with ARM.” Mr Sanders said most mobile phone manufacturers, who had invested heavily in using ARM-based chips, were highly unlikely to switch to Intel’s products.

Intel has made some missteps before, including its last foray into mobile phones (that was jettisoned in the Marvel spinoff). But thinking that it can scare Intel (or convince the world that it would threaten Intel) seems like a losing strategy.

Intel wants to own mobile phones, too

This morning I saw an interesting tidbit about Sharp developing an Intel-based cell phone for the Japanese market, to be sold by operator Japan Willcom. With 4.6 million subscribers, Japan’s 4th largest mobile carrier has less than 10% of the share of market leader DoCoMo.

The D4 phone uses Intel’s Atom CPU of the Centrino family, and will run Windows Vista (and for some reason, not XP). It has a 1024x600 screen and a 40gb hard disk. UMPC Portal describes the keyboard as “quite useable in a Psion-5 kind of way.” As such, it fits into the UMPC/MID category rather than a true cell phone, and in fact Willcom doesn’t expect to sell more than 100,000 a year.

Will this be a beachhead for Intel’s diversification into dominating (and displacing ARM) as a mobile phone platform? Intel’s prior diversification efforts have had their share of missteps. Intel’s previous attempts at the low power chips Xscale were spun off to Marvel, while its vertical integration into circuit boards also failed to displace the Taiwanese makers.

Frankly, I think Linux (not Windows Vista) is the growth opportunity Intel in the mobile space, as with its Mobile Internet Devices and its (Nokia-derived) Moblin technology. With web browsers, e-mail and IM clients available for Linux-based tablets, about the only thing an ultraportable Vista has going for it is native support for Microsoft Office documents. And if OOXML is really a standard, then that will eventually be available on Linux too.

"Silicon" Valley gone after 50 years

In 1957, the Traitorous Eight left Shockley Semiconductor to form Fairchild. This is the first Silicon Valley spinoff, as well as the real start of the careers for Robert Noyce and Gordon Moore (Intel) and Gene Kleiner (Kleiner Perkins). The next year, Fairchild sold its first product, the 2N697 silicon transistor, and the rest as they say is history.

Saturday, the Merc reported that Intel is closing its last fab in Q3 2008.

The D2 line has been used to pilot production for fabs elsewhere in the US and the world. This is the last major semiconductor fab in Silicon Valley.

The end of the story has been obvious for a long time, even if the timing has not. Given housing costs, salaries, regulation, water and energy issues, manufacturing in the Bay Area has been heading for the exit for more than a decade.

The question raised by Merc readers (in the comment section) is whether Intel plans to eventually move all the R&D away as well. On the one hand, the Bay Area is the most expensive tech cluster in the country (perhaps second only to Tokyo in the world). There is a lot of unique software expertise here, but semiconductor expertise is more broadly dispersed.

Of course, Intel — like HP, IBM, Toyota, Ford, Sony and Nokia — will have R&D dispersed around the world. Some R&D is going to remain in the US, even if not in the valley. Could there be an Intel without R&D at its headquarters? I don’t know.

Mobile LinuxWorld: Moblin

At LinuxWorld Expo on Thursday, I had a chance to meet Andrew Wilson of Intel. Andrew is formerly the head of the carrier grade Linux initiative of the OSDL (now Linux Foundation). Wilson is now heading up evangelizing Intel’s Moblin efforts.

Intel’s Moblin, like Nokia’s Maemo, is about making some sort of mobile Internet device that’s somewhere today’s (postage stamp screen) cell phones and a full-fledged laptop. (The two experiments started long before the intro of the iPhone, which tries to address that demand using a standard phone form factor).

With Maemo, Nokia ported various Linux components to run on a mobile device. But, most importantly of all, they created a new Linux user interface they call “Hildon” that solves many of the Linux/Unix UI usability problems. Ironically, Hildon was initially developed by Psion (the forerunner of Symbian) for Nokia’s Communicator family of Symbian-based cellphones. No word on how (or if) the Psion Hildon code ended up in a Nokia product that competes with Symbian.

Nokia has so far shipped two products based on Maemo, both of which used Wi-Fi but had no GSM connection. The initial Nokia 770 Internet Tablet (released in May 2005) had many problems, but this January it released the Nokia N800 Internet Tablet corrected many problems.

Mobilin adapted the Maemo code (including taking Hildon), and also added other other code such as a Firefox-based mobile browser. Some say the Moblin UI is iPhone-like, but that may just be pretty icons. Both want to commoditize operating systems, but Nokia wants to sell devices and Intel wants to replace ARM chips with Intel chips.

Many had speculated whether Intel was stealing from Nokia, cooperating with Nokia or even helping Nokia’s efforts, and similarly how Nokia would react. Wilson said that Moblin has attracted several Nokia contributors, and is hoping to have many firms besides Intel actively involved in developing code at Moblin. There seems to be no public support by Nokia (i.e., at the press release level) but that does not preclude Nokia engineers from joining and monitoring the technical effort.

Nokia is offering Hildon to the GNOME Mobile initiative, née GMAE (GNOME Mobile & Embedded initiative), which seeks to make GNOME and GTK relevant for mobile devices. Both Intel and Nokia are sponsors of the GNOME Foundation. So at both the strategic and technical level, the two companies are working to achieve some of the same goals.

Technorati Tags: Intel, mobile phones, Nokia, open source, GNOME

AMD celebrates gift from Brussels

Catching up during a brief vacation, after a busy week of getting ready for the new semester.

Tuesday’s papers had a full page ad from AMD celebrating last Friday’s EC action against Intel. I saw the ad in the Wall Street Journal and the San José Mercury News, but I’d bet it also ran in the NYT, USA Today, and perhaps the Washington Post.

With a letter by CEO Hector Ruiz, the AMD ad trumpets

The European Commission reveals the truth about Intel.

Obviously AMD is grateful for a powerful ally in its bitter rivalry with Intel. AMD has been counting for a long time on antitrust efforts to help its fight, with (among other things) funding of a PBS documentary praising the value of antitrust enforcement.

According to the AMD summary of the European Commission allegations, Intel:

• Maintained its monopoly by paying substantial rebates to PC manufacturers on the condition that they do little or no business with AMD,
• Paid manufacturers to delay or cancel the launch of AMD-based products, and,
• Provided below-cost microprocessors for servers to prevent rival solutions from being chosen by strategic customers.

As with any lobbying or other public policy-oriented message by an interest group, AMD’s characterizations could be charitably construed as a selective presentation of the truth.

It’s not surprising that Intel’s woes sound like the Microsoft case — ongoing investigations and occasional allegations of monopolistic practices. In both cases, the European threat of action seems more severe than in the US. And of course, both tie to the Wintel duopoly.

However, as measured by market share Microsoft’s grip on customers over the past decade has been tighter than Intel’s. If anything, Intel’s problem is that it has a richer and more determined rival than Microsoft has had since AOL Netscape threw in the towel. Meanwhile, for the past decade, Apple has pulled its regulatory but not marketplace punches in a textbook case of co-opetition.

It’s also not surprising that those that defend Microsoft also defend Intel, including the very same WSJ that gladly pocketed AMD’s advertising dollars. On the same page that blasted EC attacks on Microsoft in March, an unsigned editorial entitled “Intel in Euro-Land” proclaimed:

Computer chips are getting faster and smaller, and prices are dropping amid fierce competition. So naturally the European Commission thinks this is the ideal time to lodge another antitrust suit against another American technology titan.

There are several problems with the logic of this sentence, but finding them is left to the readers as an exercise.

Still, there are two problems that the WSJ editorial raises that are not so easily dismissed away.

First, the potential sanctions of 10% of Intel’s global revenues would be unconstitutional in a country that had a rule of law. Intel is not headquartered in the EU, and so the EU does not (or should not) have jurisdiction over anything but its operations there. (What would happen if every country in the world decided to fine Intel 10% of global revenues? It would be a great way to print money until Intel went broke.) Sure, an antitrust decision by the EC not to allow a merger in the EU is equivalent to veto power for a global company, but the EC should have no more ability to fine Intel for non-EU operations than it can tax them.

Secondly, while the WSJ is predictably glib in dismissing sanctions over a fierce business rivalry, seriously engaging its arguments does raise a fundamental question: where does normal competition end and anti-competitive behavior begin? The WSJ seems to be in denial that predatory competition exists, but certainly there are historical existence proofs of firms trying to drive out competitors to raise their pricing power. If, as alleged, Intel sold chips below its costs, that is the standard used by the (US) International Trade Commission for sanctioning such competition.

On the other hand, firms are expected to compete for customer loyalty and (legally) maximize shareholder returns. (New York to Paris: we call it capitalism.) Two once-great American innovators — IBM and AT&T — atrophied during the 1960s and 1970s in the face of US government antitrust actions that caused firms to internalize such limits and, in effect, “pull their punches.” Eventually IBM came back, but it looks like AT&T never will. The US (and world) economy would be the poorer if Intel did not continue to invest heavily in fulfilling the 1965 prophecy of Moore’s Law.

Technorati Tags: antitrust, EU, Intel

Moblins run loose

For its Mobile Internet Device initiative, Intel has decided to brand its sponsored open source project “Moblin.” [YALBNdtCS]†

Most of the code is GPLv2 or LGPLv2.1. Intel admits its complete debt to Nokia’s maemo project: “Porting applications from maemo.org to mobile Internet devices and vice versa will be straightforward.”

In my own research with Siobhán O’Mahony, we have found, however, that there’s more to open source openness than just a license, i.e. control over development and governance. Intel has historically had tight control of standards consortia that it sponsored (such as USB 2), so that pattern seems likely to continue here. As with other tightly-controlled sponsored projects, this means hackers can use the code but rivals will have little or no influence over its direction, let alone ability to use it for their own strategic goals. (Yes, the case of controlling GPL licensed Linux code is a little more complex than say MySQL or Sleepycat).

Intel’s goal is to sell more chips and crush ARM. Open source is a means to that end, not a goal in itself. Commodity software is bad for a company that sells software but good for a company that sells chips.

Just an aside: does “moblin” rhyme with “goblin”? I know bugs are often “gremlins,” but are they also “goblins”? What does this say about code quality?

† This is coming up so often that from now on I’ll just use the YALBNdtCS acronym for “Yet Another Lousy Brand Name due to CyberSquatters”

Technorati Tags: Intel, Linux, mobile phones, Nokia

Intel the one trick pony

Normally I think about the battle for total world domination as Microsoft vs. Google. An occasional wise guy might suggest Apple, Nokia or Yahoo, but they are definitely 2nd-tier contenders.

The end of the tech bubble made it easier to forget about the other half of the Wintel duoopoly. Yes, Intel’s stock has been falling over the past three years, but they’re still a $40 billion/year company with at net profit margin in excess of 20%.

Somehow I’d missed Intel’s declaration of war against ARM and its chokehold on the mobile phone processor market. Once upon a time, they were going to work with ARM via their XScale processor, but they dumped that off almost a year ago to Marvel, a SV startup that hopes to succeed where Intel failed.

In April in China, they unveiled the Linux-based “Mobile Internet Device,” in some ways an extension to their Ultramobile PC and in other ways a knock-off (as well as building upon) Nokia’s open source work in its 770 and N800 tablets. (If you’ve been living under a rock, these are non-GSM WiFi devices). The Intel products (to be made by various ODMs — see also the Japanese report) are due early next year.

There’s no way for Intel to succeed without besting ARM, who (according to a May shareholder presentation) have better than an 80% market share for its architecture. Intel has the sole PC architecture (since Apple switched) and the vast majority of chip sales in that architecture, and of course the 25+ year old ecosystem that goes with the x86. So this is really a war not only between two companies, and two architectures, but two ecosystems: given Intel’s resources, it at best will be a WWE-style SmackDown, and at worst, a fight to the death.

This harkens back to Intel’s decision in the 1980s to kill its RISC architecture, the iAPX 432. There are two ways to interpret this.

On the one hand, as Andy Grove recounted in his 1996 book Only the Paranoid Survive, Intel’s support for the i432 was diverting resources and sending a mixed message to the market. Instead, Intel put all its market and R&D power behind the x86, and they did pretty well by that until the bubble burst.

On the other hand, Intel is out of DRAMs and probably going to bail out of flash memory too. It tried motherboards and gave up. Its other comm chips went to Marvel. In short, it’s failed at everything other than follow-ons to the 8086. So in the 29 years since the x86 was born, it hasn’t really created any new lines of business. Apple went from Apple II to Mac to iPod (and maybe even iPhone). Microsoft went from Basic interpreters to operating systems to business productivity to server software — and created a whole new videogame platform while it was at it. IBM of course has the broadest revenue stream of anyone.

Oracle is a one trick pony, and Larry Ellison is still worth billions. WordPerfect and Lotus were one trick ponies (OK, Lotus came up with a 2nd trick) and they’re gone, just as Novell is in the process of disappearing. Intel is a lot bigger than those companies, but right now it seems to be saying that its growth will come from repackaging the same old same old, rather than creating some new product family or market.

Technorati Tags: Intel, Linux, mobile phones

Clearwire 1, Shareholders 0

In response to my earlier posting on WiMax startup Clearwire, the company sold about 15% of its shares, netting $559 million in its Thursday IPO, by pushing 24 million shares instead of 20 million. The stock fell both Thursday and again Friday, meaning those who paid $25 at the IPO have already lost 11.4% of their initial investment.

Interestingly, 20 months ago the company got $900 million in strategic venture capital, including $600 million from Intel. One thing I had missed was the lopsided voting rights of the founder and strategic investors, as reported by Bloomberg:

Clearwire will have two classes of stock, a format used by some companies to ensure that founders keep control. Class A shares, which will be offered to the public, will have one vote each. Class B shares will have 10 votes. McCaw controls 52 percent of the votes, according to the filing.

Wireless Week reports that the net voting shares are McCaw 49%, Intel 30% and Motorola 4%. Other shareholders are clearly just along for the ride, and so far it hasn’t been pretty.

Technorati Tags: broadband, Intel, IPO, WiMax

Why Professors Shouldn’t Make Predictions

Back in 1992, John Sculley (Apple CEO 1983-1993) made an amazing appearance before a group of executives at Harvard Business School. It was amazing not because of his brilliance, but his utter cluelessness. One comment was:

I wish we had started moving our technology to the Intel processor years ago so we had more options. Because most of the industry is taking advantage of the tremendous price drops that are going on the Intel world, and we can’t because we’re not on it, and it’s very difficult to move our technology over to that processor very quickly. So I wish we had that option, which we don’t.

As I noted in my analysis of Apple’s many failures during that decade, later in 1992 (according to Jim Carlton’s excellent book), a small team of engineers got a working prototype of Mac OS 7 working on Intel chips, but the project was abandoned when its sponsor left Apple. So if Sculley (nominally CTO) knew what was going on, he could have started the project years earlier and protected it once it launched.

Sculley’s host for the Harvard session was David Yoffie, a much more successful, famous and distinguished strategy professor than yours truly. Yoffie has written the various HBS cases on Apple, as well as a Qualcomm case that (I’m told) competes with my five cases on the company. I consider Yoffie one of the leaders in IT industry strategy, with his cases, journal articles and books. He has a Ph.D. from Stanford, has been an HBS professor since 1981, and also is a director of Intel and Charles Schwab.

However, today I found a 2004 interview of Yoffie that was very skeptical about Apple’s future. Among other things, he noted their revenues had shrunk from $11 billion to $6 billion (true). He then predicted long-term failure for Apple’s niche strategy, for its competition with Windows Vista (then called Longhorn), and Jobs’ ability to continue to succeed:

The interesting thing about teaching Apple is that when things are going bad, everybody says, "Of course." And when Apple is doing well and you begin to teach the same basic ideas and explain why it's not sustainable, there's enormous skepticism. I've seen the cycle now several times. When I first started teaching the case, things were going great and nobody believed me that Apple had problems. And then when Apple was going down, everybody said, "Yes, of course." We see these wonderful cycles in teaching where people tend to be swayed by the latest fad in the business press about Apple.

Let’s see what’s happened since Yoffie’s commentary was published:

Company	Market cap§	Stock appreciation	Net income (2006)	Return on sales
Apple	$77b	+301%	$2.0b	10.3%
Dell	$54b	-27%	$3.6b	6.4%
Intel	$121b	-31%	$8.7b†	22.3%
Microsoft	$288b	+7%	$12.6b	28.5%

§ As of market close Feb. 22, 2007
† 2005 is latest available

I think what Yoffie is missing is that CEOs matter. Even if Bob Sutton is right that they often don’t, in this case Apple really had an awful string of CEOs from 1983-1997: they wasted more than $2 billion on R&D and inventory-related losses, and the opportunity costs of neither improving the Macintosh nor making a winning PDA were huge. Yoffie’s first Apple case was published in 1992, which means that all the down cycles came in the past 15 years under the bozo CEOs. Meanwhile, Jobs has done the best job of the PC companies of breaking away from the PC paradigm and finding something truly new. (Can it be sustained? Who knows?)

So one take-away is that professors shouldn’t make predictions. Unlike stock analysts, we don’t make a lot of predictions, so with small numbers it’s hard for even the best to yield a Ted Williams batting average.

I also wonder whether Yoffie’s perspective has been colored by sitting inside Intel board meetings since 1989. In 2004, Apple was still outside the Intel ecosystem and Andy Grove had been predicting for a decade that Apple would fail without Intel chips. Thus my second take-away is that professors (like journalists) are human, and we should be more up front about our biases. As a result, I’ve updated my web page to list consulting clients.

Graphic credit: O’Grady’s PowerPage

Technorati Tags: Apple, Intel, John Sculley, Microsoft, Steve Jobs