Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
>> I think these lines are key to the discussion we had yesterday on the forum. Intel plans to promote both Centrino and Mobile Pentium 4 for the mobile market, the former being promoted more towards business, while the latter is positioned more towards the consumer. IBM's decision to introduce new Pentium 4 based laptops is in agreement with this. DewDiligence, take note. <<
wbmw: You just shot yourself in the foot. The ThinkPad line has always been marketed primarily to businesses. Hence by your logic the new ThinkPad should have been an excellent candidate for Centrino. That the new ThinkPad instead uses a P4 only strengthens the speculation that Centrino may indeed have some problems with quality or yield. Dew
The CNET article contains the usual marketing spin from INTC that users must have Centrino in order to do wireless computing.
Moreover, the article highlights two reasons why the “desknote” market continues to thrive in spite of INTC’s attempts to kill it off: 1) The mobile Pentiums have until now been seriously overpriced relative to their desktop counterparts; and 2) Many consumers, even in the middle-weight category, are evidently content to forgo SpeedStep.
Overall, the article indicates just how difficult it will be for INTC to present a clear marketing message about what each mobile processor line represents. INTC’s cornucopia of mobile CPU’s reminds me of the array of overlapping brands from General Motors. I never could figure out the difference between an Olds and a Buick, and maybe others couldn’t either -- which might be one reason Olds is being canceled. FWIW. Dew
>> What I was trying to say was that the Centrino platform is a specification for a collection of features, not just a wireless connection. This specification will be relied upon by the wireless ISPs to allow an easy connection. Over time if the specification becomes widespread, it will achieve the status of "standard" in the same sense that uses the term Industry Standard server to refer to an Intel platform. <<
Your comments could have come directly from INTC’s marketing department. You have articulated what INTC would like to see happen with WiFi but you have stated it as though it were a fait accompli.
In reality, several companies other than INTC will have an impact on the way wireless standards evolve in the realm of mobile computing.
>> See my post, IBM is actively working to get Hotspot vendors to sign on to the Centrino standard. <<
Did you mean the WiFi standard? Last time I checked, WiFi was not a proprietary Intel technology.
You are so right it’s almost funny. Imagine, if you can, that IBM had just unveiled a new ThinkPad, which used an established CPU, just a few weeks before the much-heralded launch of a brand new power-efficient chip from AMD. How do you think this message board would react?
Under such circumstances, there would no doubt be at least 65 posts proclaiming that IBM’s action proved: 1) AMD’s new chip must be thoroughly without merit for IBM to preempt its launch; 2) AMD’s chip was probably nothing but vapor; and 3) AMD’s ability to continue as an ongoing enterprise was fatally compromised as a result of 1) and 2). Dew
What’s your definition of “trollish”?
>> the fact remains that many in the investment community take careful note of large insider sales and the conditions surrounding them.
I certainly do (which is one reason I am bullish on Corning).
Alan Abelson of Barron’s (which, IMO, is a much better indicator than Business Week) has said that “among the many reasons for insiders to sell shares, a strong conviction that the stock will go up is not one of them.”
(Curses. I may have inadvertently encouraged Usuck to pay another visit.)
>> I meant LV and ULV versions of PIII and P4 [will be dropped] <<
If the LV & ULV models of the P3M/4M are not dropped, then you would really need to start worrying about Centrino.
>> Well I guess this proves Intel is headed for the moon! <<
Might be a good place for a fab, actually. Low gravity, plenty of silicon, and no pollution controls…
Re: TMTA in trouble?:
Over the years, I have found Business Week (especially Gene Marcial’s column) to be a pretty good contrary indicator. FWIW. Dew
I don’t often comment on intraday price movements, but I’m impressed with how well GLW is holding up on a day when investors are acting as though the world is about to end…
This time I have to agree with TheInquirer: IBM’s decision to unveil a non-Centrino ThinkPad when the Centrino launch is just a few weeks away is significant. It might be that IBM wanted dual-band WiFi, which Centrino won’t have until mid-year at the earliest. Or there might be another explanation which is more troubling for the Centrino launch…
Problem with Centrino?
http://www.theinquirer.net/?article=7440
Yes, consider the source. But, occasionally, even this source has the correct story.
>>
IBM IS JUST about to release the newest series of Thinkpad laptops, the R40, but there's something strange.
The machines use Celeron or Pentium 4 processors just as Intel is about to release its Banias series of processors, the mobile Pentiums that everyone has been waiting for.
In all other respects, the R40 seems to be just the thing to keep IBM competitive, most particularly for busy office users. The upper end of the R40 series comes with dual-band wireless LAN capability built in. It means the machines can use both 802.11b, the normal Wi-Fi standard, and 802.11a, the new faster version of Wi-Fi.
IBM has built in several features to make systems administrators happy with the new machines. There's an 'Access IBM' button that links the user to diagnostic and online help pages. There's also an interesting feature called RapidRestore which lets you erase the primary partition and recover from another partition. A boon for those "I got a virus by looking at a work-related website" type moments that keep cropping up. It should also make redistribution of the laptop to a new user much easier.
But back to Banias. The fact that IBM has chosen to release a new series of laptops without the new Intel processor is a strange one. Has IBM fumbled the ball and produced a simple upgrade to the R32 series just when Banias is about to make its big hit? Or has Intel fumbled the ball and created a new mobile processor that just doesn't cut it? Early Banias prototypes have been seen in the wild and seemed to do well but could Intel be having problems making enough of them? We think we should be told.
<<
>> Some think TA (technical analyst) is bunk. <<
I must admit I am one of those people.
I would expect GLW’s near-term price movement to be dominated by speculation about what will be said at the meeting a week from Friday. Dew
P.S. Good article in on GLW Business 2.0. Based on that, I decided to sign up for a trial subscription.
[OT] Business 2.0 article on Corning’s non-telecom businesses:
http://www.business2.com/articles/mag/0,1640,46301,00.html
This is the part I thought might be of interest to the semiconductor gurus on this thread:
>> To etch designs onto the silicon wafers that become memory chips or microprocessors, chipmakers use a machine called a stepper. Inside are lenses and mirrors that help guide the laser beams burning the design onto the silicon. Corning has long made fused silica, the same purified glass used in the Hubble telescope, for those mirrors and lenses. But now that chipmakers are squeezing ever-smaller circuits onto their chips, the next generation of steppers will require an even purer kind of glass. Corning expects to supply it by adapting a technique that its researchers have been working on for five years, which involves growing a perfectly clear crystal of the mineral calcium fluoride. Corning anticipates that as the semiconductor industry emerges from its latest cyclical rut, [calcium fluoride]sales will climb 20 percent annually between now and 2008.<<
>> OK, so to round out the argument does anyone remember when AMT went into effect? <<
Let me step in here because I think the issue is not when AMT went into effect but rather when it became consequential for middle-class and upper-middle-class taxpayers.
Back in the Reagan administration, the brackets for ordinary (non-AMT) income taxes were indexed for inflation. However, the threshold for bumping into the AMT (i.e. where the AMT calculation produced a higher tax than the regular calculation) was not adjusted for inflation and it never has been. Hence, over the past 20 years, the AMT became consequential for a larger and larger percentage of taxpayers.
Originally enacted as a way to assure tax collection on the very rich, the AMT has evolved through legislative inaction into an onerous tax on many members of the middle class. Dew
Agree with your comments. Excellent post. Regards, Dew
Re: Corning Inc. may drop photonics:
CFO Flaws stated in the CC that, regardless of what the company decides to do with photonics, there will almost certainly not be a significant asset writedown (goodwill or otherwise) this year. Hence there is no danger of bumping against the 60% debt/capital covenant on the untapped $2B credit line.
Thanks for the update, Chart.
Today, I bought back some of the shares sold on 1/14 @4.70 (see message #37). It would have been instructive to see how GLW would have performed today if it hadn’t been such a wipeout for the broad market. I thought the CC was quite positive overall.
As I posted earlier, I’m impressed with the actions the company has taken to shore up the balance sheet. Now that liquidity is a non-issue in the short-to-medium term, and the operations are already free-cash-flow-positive, GLW’s businesses will essentially take care of themselves. I think the likelihood is now high that the company will be profitable for the full year on a GAAP basis. Regards, Dew
Q4 results: mostly positive:
GLW’s press release ran 35 pages (!) on my printer (using medium-sized fonts), so there may be a lot of stuff I missed. Here are what I consider some of the highlights:
Overall, the 4Q02 operations were weak, but the company more than compensated for the operational shortcomings by making some great financial maneuvers: i.e., selling the precision lens division to 3M and buying back deeply-discounted debt on the open market.
Let’s start with the debt repurchases. In 4Q02, the company bought back $215M face value of debt at an average of 58 cents on the dollar!! For the full year, they bought back $493M face value at an average of 62 cents on the dollar. These kinds of opportunistic debt repurchases effectively neutralize some of the sins committed during the tech bubble!
Thanks to these debt repurchases, the convertible offering last August, and the sale to 3M, the balance sheet is now quite solid. The 12/31/02 debt/capital ratio was 46.7%, which is well below the 60% covenant on the still-untapped $2B credit line. In other words, GLW has unfettered access to $2B of credit – but will probably never need it.
(Solomon Smith Barney’s warnings from last summer of GLW’s eventual demise from its inability to access the credit line can now be seen for the B.S. that they were. And this is hardly surprising because SSB was, after all, the home of our dear friend, Jack Grubman.)
It gets even better: since January 1, 2003, GLW has bought back an *additional* $158M face value of debt on the open market. (We will know what they paid for this debt when Q1 results are reported.)
Segment info: GLW has changed the way business segments are reported, combining all non-telecom businesses into the new “technologies” segment. This will make it somewhat harder to determine the profitability of the LCD operation and the specialty-materials business. I believe that GLW made this reporting change for competitive reasons. (This is a more credible explanation than the one given in the PR.) Sales, but not profits, continue to be reported by detailed business lines.
Share count: As of 12/31/02, all but $155M of the $575M of convertibles issued last August had already been converted into common stock. Assuming that the remainder is eventually converted at $1.97/sh, the converts will produce 79M additional shares. Adding these to the 1.267M basic shares outstanding at 12/31/02 gives an effective basic share count of 1.35B shares.
Guidance: I posted earlier about the positive signal management is sending by re-establishing financial guidance on Feb. 7. Although I personally would be content to have the company forgo guidance, I am impressed by the confidence they are showing and I believe that guidance *does* matter to a large cross-section of the investing public. Hence, the re-establishment of guidance can be expected to boost the share price to a non-trivial degree.
That’s it for now. More comments after the CC in about eight hours. Dew
GLW posts 4Q02 results and re-establishes financial guidance:
http://biz.yahoo.com/bw/030123/232504_1.html
>>
We will continue to strengthen our balance sheet by reducing debt. We intend to invest in research and development in support of existing growth businesses such as LCD, as well as emerging opportunities including chemical processing and diesel emissions control.
Flaws said the company will provide first-quarter 2003 financial guidance at its annual investor conference at the Pierre Hotel in New York City on Friday, February 7, 2003 and simultaneously via audio webcast.
<<
--
This is quite positive, IMO. If Q1 were expected to be bad, it would be odd to re-establish guidance in a public forum among all of the analysts and institutional investors. So my hunch is that Q1 will be surprisingly strong. Dew
>> Hector is definitely becoming a Jerry clone, <<
Could it be that AMD’s forays into the microprocessor business are just an elaborate ruse, and the company’s true mission is in… biotechnology?
>> does iHub now have a problem with links? <<
Yes, evidently. iHUB truncated some of my very long links (without any warning of a problem). Shorter links seem to work OK.
Tellabs CEO: “glimmer of hope” peeking through the “nuclear winter”:
Like GLW, TLAB is not providing any forward guidance.
http://biz.yahoo.com/djus/030122/1855001198_2.html
“A more favorable regulatory climate that may let local phone companies expand their long-distance services outside home markets may also stimulate sales…”
Unfortunately, that’s about the most positive line in the story…
>> there is a 3 year statute of limitations on the IRS collecting back taxes.<<
For fraud, the limit generally extends to six years. (One of my regular backgammon partners was an IRS auditor.) Dew
Re: ISO vs. NQ options:
Good summary by Alan of the distinction between the two kinds of options from the options holder’s viewpoint.
For investors, it is perhaps more important to examine the distinction between the two kinds of options from the viewpoint of the issuing company.
In general, any income that is currently-taxable to an option exerciser is considered a tax-deductible expense to the issuing company – even though the company does not expend any cash on the transaction. The company’s book expense on employees’ option exercises reduces the company’s current profits, which results in a lower tax bill and actually saves money for the company (provided that it is paying taxes in the first place). During the tech bubble, a large % of reported “pro forma” earnings for many companies consisted of these tax benefits.
When the tech bubble burst and stock prices fell, many options went out-of-the-money, and hence the corresponding tax benefits to the issuing companies disappeared. This further reduced profits (or increased losses) at many of these companies, which exacerbated the impact of falling revenues during the post-bubble period. Dew
I saw Possum’s post. If GLW were to guide toward break-even in Q1 (before non-recurring items), the stock would surely fly.
But that probably won’t happen; GLW discontinued forward guidance a few quarters ago, and I think they will wait for a good quarter or two before they resume. Dew
FWIW: I posted highlights of the article on the graphics/NVDA MB:
http://www.investorshub.com/boards/read_msg.asp?message_id=698627
Most of the article should be old hat to readers of this MB.
Re VIA:
VIA’s best bet may be continuing to supply the CPU for the $199 PC’s sold at Wal-Mart, where VIA’s $45 ASP comes in handy. My guess is that no other microprocessor company is going to fight real hard for that account.
Article on semiconductor fabs:
http://news.com.com/2102-1001-981418.html
Mostly old news. A few excerpts which are germane to the foundries such as TSM:
>>
In this new era, Intel could gain an even greater opportunity to undercut opponents or enter new markets because of its unique self-sufficiency. Asian foundries such as TSMC, meanwhile, will grow in power, and IBM may become the de facto R&D department for an array of semiconductor companies.
…foundries are rapidly growing in sophistication and importance. Collectively, they will grow by 31 percent annually and likely account for 40 percent to 50 percent of the world's semiconductor output by 2010, said Dan Byers, director of worldwide marketing for Taiwan Semiconductor. He said TSMC will start making 90-nanometer chips around the same time as Intel…
<<
Not having a "reco” meter removes some of the fun, but it might actually be a good thing because it forces people to post for the information value rather than just to cheerlead.
Are you looking for any surprises in tomorrow’s earnings release or Friday’s CC?
EP: looks like there’s an iHUB bug which truncates very long links. Try this link to the TMTA message board itself:
http://messages.yahoo.com/bbs?action=topics&board=1602585284&sid=1602585284&type=r
then go to message #51075. Dew
spokeshave: no slight intended
I may have misinterpreted that poster’s questions, but I thought he mainly wanted to understand what limits the clock frequency on a given piece of silicon.
The determination of clock frequency:
I was going to answer this post on the Yahoo TMTA board:
http://messages.yahoo.com/bbs?.mm=FN&action=m&board=1602585284&tid=tmta&sid=16025852....
but then I figured some of the people here, like wbmw, EP, or Alan, could do a better job. T.i.a. Dew
>> TMTA may be selling 1GHz product because the market demands it but that doesn't mean the binsplits are good enough to allow any profits at that speed. <<
It’s clear that TMTA is earning a positive and increasing gross margin on the TM5800. However, your comment is nevertheless apropos because TMTA has sacrificed some gross margin in order to meet HP’s demand for the 1GHZ chips. At the most recent CC, TMTA stated that yields from TSM have improved and, consequently, I infer that gross margin will benefit commensurately.
>> Is market share more important now [to TMTA] than profit? <<
Yes, evidently. Is this a reasonable approach? Possibly. If TMTA’s second-half financial numbers turn out as robust as management’s guidance, I will conclude that they did the right thing by sacrificing short-term margins. FWIW. Dew
>> …you can not infer anything whatsoever about speed distribution based on maximum frequency rating because market forces dictate which speeds can be sold and at what price. <<
EP: I understand that the nominal clock speed of a given chip may not be the same as its true capability. But I was looking only for a rule of thumb – something crude like Alan’s “+/-10%.” Are you disagreeing with Alan that this rule of thumb has any value for semiconductor-industry investors and analysts?
>> TMTA may simply be doing an "AMD". Announcing products it has no ability to manufacture but must give the impression it can. If however you want to assume that 1GHz is at the high end of the frequency distribution then you need to know something about who is fabing their parts…<<
I do not see how TMTA is pulling an “AMD.” We know that the Crusoe TM5800 can be produced at 1GHz because that is the clock speed which is used in HP’s Tablet PC (which has been shipping for a couple of months) as well as other announced designs. From your argument, I conclude that TSM’s production line might be yielding even faster chips which are labeled 1G because there is currently no product using a >1G version of the TM5800.
>> A foundry is optimized for yield and not speed. I would expect to see a greater frequency distribution from a foundry than a dedicated CPU house like Intel or AMD. <<
Assuming other variables are kept constant, by what magnitude would you expect the frequency distribution to widen at a foundry relative to a dedicated house? 50%? In other words, if Alan’s “+/-10%” is a reasonable rule of thumb for dedicated houses, would “+/-15%” be a comparable rule of thumb for foundries? T.i.a. Dew
Question re binsplit-distribution curve:
I understand a little about the randomness which affects binsplits on a production line, and I know enough college math to expect the binsplit distribution curve to approximate the bell curve of the so-called normal distribution.
My question concerns a post on the Yahoo TMTA MB:
http://messages.yahoo.com/bbs?action=m&board=1602585284&tid=tmta&sid=1602585284&mid=...
This post claims that the midpoint of the TM5800 binsplit curve can be assumed to be near 900MHz since the maximum rated frequency is (currently) 1GHz. What I do not understand is how a strong inference can be drawn about the midpoint of the curve from merely knowing the maximum (unless you are privy to data from the actual production line).
T.i.a. to anyone who can help clarify this. Dew
Boston Globe article on Intel’s communications business:
http://www.boston.com/dailyglobe2/020/business/Inside_IntelP.shtml
[FWIW: I question the comment halfway through the article that a network-processing chip is necessarily more complex than, say, a Pentium – Dew]
>>
Inside Intel
Seeking new markets
By Peter J. Howe, Globe Staff, 1/20/2003
HUDSON - Three decades after inventing some of the world's first microprocessors, Intel Corp. has built a commanding share of about 85 percent of the market for the chips that power personal computers.
But with the PC market maturing, Intel is aggressively looking for new markets for its semiconductors and components, and a team based at its huge 2,500-person hilltop plant here is at the center of a major initiative to bring the ''Intel Inside'' concept to a range of telecommunications hardware.
Later this year, Intel plans to roll out a network processor chip called the IXP 2800, roughly 20 times as powerful as the current version, that it hopes will become a building block for a range of switches, routers, and wireless and ethernet broadband delivery devices.
Much of the design work for the new chip - which is barely a half-inch square but can move up to 10 billion bits per second - was undertaken at the company's Massachusetts Development Center, located inside the former Digital Equipment Corp. chip plant and office park that Intel bought five years ago. The final design was completed last month.
With a deep slump in global telecom spending depressing revenues for equipment makers like Lucent Technologies, Nortel Networks, and Sycamore Networks, this is a terrible time to be pushing into the telecom market, right?
Not exactly, Intel figures. Its strategic theory is that cash-strapped telecom equipment makers now devoting considerable resources to designing specialized chips called application-specific integrated circuits will welcome a chance to buy, off the shelf, an extremely powerful, programmable microprocessor for future generations of switching devices.
''Because of the downturn in the telcom market, we think [equipment makers] are more likely to utilize this technology,'' said Nick Finamore, the Hudson-based general manager of Intel's network processor division. ''They're really looking to reduce their development costs and to be able to converge multiple product lines onto one technology platform,'' buying a single model of microprocessor that is like ''one engine that goes into cars and trucks and motorboats.''
Intel's move into the telecommunications market is far from a bet-the-company gamble. But it is still an important and heavily financed thrust into a market estimated in the billions of dollars by mid-decade, and it could affect Intel's future growth. Over the last five years, Intel has thrown an estimated $10 billion or more at the communications market, between acquisitions of telecom-focused start-ups and research and development.
Intel's fourth-quarter earnings, reported last week, showed that the core PC and computer semiconductor business remains its biggest - and by far its most profitable - business. Out of total revenues of $7.2 billion and net income of $1 billion, the communications group, which makes chips for wireless WiFi systems, optical switches, and ethernet switches as well as network processors, accounted for $544 million in sales, but posted a net loss of $168 million.
While the division was down 20 percent last year from 2001, it showed growth in the fourth quarter, up from the prior quarter's $482 million in sales.
Linley Gwennap, principal analyst for the Linley Group, a Mountain View, Calif., firm that consults for chip makers and buyers, said he estimates Intel currently has about 11 percent of the global market for network processors, ranking it in fourth place after Applied Micro Circuits Corp., IBM, and Motorola, and ahead of Lucent Technologies spinoff Agere Systems. However, ranked by ''design wins,'' or announcements from equipment makers about whose chips they will use in their gear, Intel is winning about 25 percent [emphasis added].
''They certainly don't dominate the market the way they do in the PC processor market,'' Gwennap said. ''But it is a market that we expect has some pretty significant growth potential in the next several years. The nice thing about being Intel is that you can invest a lot of money in it even if you are losing money today. And by and large, the Intel network processors are competitive'' with other companies' offerings.
Intel competitors are hardly lying down waiting to be flattened. San Diego-based AMCC later this year expects to roll out what it calls its fifth generation of network processors, enabling equipment makers to put multiple voice and data services on a single line card. Although its total sales are running at about $120 million a year, it has been estimated to control over one third of the network processor market.
''No other vendor has the breadth and depth of network processing experience and proven track record that we do,'' despite Intel's size, said Jeff Cashen, vice president and general manager of AMCC's switching and network processing division.
While a Pentium chip might power a piece of word-processing or spreadsheet software, a network processor is a more complex chip that is helping direct millions of bits of data every second being sent from a Web page or e-mail server or the digitized bits of a phone call. Performance demands on network processors are intense because they directly affect how well Internet and phone connections work.
As computers do more communication, including handling voice calls over the Internet, and communications devices such as cellphones and wireless handheld devices get more computing capacity, the division between what is a computer chip and what is a communications chip is blurring - but Intel wants to be a giant in both areas.
Telecom equipment makers, clustered around Interstate 495, indicated they might be receptive to the Intel pitch to supply a greater share of the processors they use in their switches, although they have the usual concerns about their cost and reliability.
John St. Amand, chief executive of Telica Inc. in Marlborough, which makes a high-capacity voice data switch that has been purchased by Verizon Communications and several small carriers, said Telica's switch uses roughly 20 percent ASICs, designed by Telica engineers and hired consultants, and 80 percent chips made by Agere and Motorola. The refrigerator-sized Telica Plexus switch can handle the equivalent of 250,000 simultaneous phone calls.
''The goal for companies like Telica and larger companies would be going to more off-the-shelf processors for two big reasons: time to market and saving money, St. Amand said. ''It's good for us box builders to get out of the chip business.''
But, St. Amand said, ''A lot of companies like us don't want to hop on the latest and greatest chip to come out of a company like Intel or Agere. We try to stay clear of first-generation products'' until they have some track record.
Kevin Klapper, senior director of hardware development for Sonus Networks in Westford, another voice data switch maker with 22 customers, said Sonus made a decision from its founding five years ago to use programmable network processors exclusively and avoid ASICs.
''I think the battle has been waged and has been won by the network processor camp,'' Klapper said. ''They have reached a level of performance and flexibility that meets the needs of the telecom and datacom space.''
Klapper said Sonus is evaluating the new Intel processors, including a 2850 version that has security and cryptography features built into the processing chip, which avoids the need for using a second chip to secure data packets that move through the main processor. ''I'd characterize that as a unique approach that certainly has some future potential,'' Klapper said, particularly for a Sonus whose core business is routing phone calls as Internet data packets rather than traditional circuit-switched calls.
Gwennap said that a typical telecom start-up developing a design for a specialized ASIC chip may have to devote 10 or 20 engineers to the job and put up $1 million just to get a prototype from an ASIC maker such as IBM, Fujitsu, or smaller Asian chip foundries. In contrast, Intel expects to sell the 2800 chips for $485 to $695 apiece, and the security-incorporating 2850 for $725, plus fees for the software needed to program them. Specifications vary widely, but a typical switch or router that sells for $100,000 to $500,000 might use 40 to 100 of the chips in its circuit boards, representing a significant fraction of the final cost.
Whether a successful bid for market share with the network processors translates to new jobs in Hudson is unclear, although it would bolster the importance of the Hudson-based research unit within the multinational chip giant. Intel has seven large chip fabrication plants around the world, all designed to make multiple types of chips.
The Hudson site has seen $1.5 billion in investment from Intel since it was purchased from Digital in May 1998. Its centerpiece is a 96,500-square-foot ''clean room'' where production workers wear astronaut-like suits with filters to capture their exhalation, and walk around on a grated floor suspended 26 feet over the bottom of the room. designed to have no more than 1.3 microns worth of particles per cubic foot of air, to reduce impurities and defects in wafers that are turned into processor chips. (By way of comparison, the period at the end of this sentence is about 300 microns in diameter).
''We're still in the shadow of the business that's oriented around PCs,'' Finamore said. ''But we like to think this has been a quiet success story that's helped build a $2 billion business, and there's a lot more to come.''
Peter J. Howe can be reached at howe@globe.com.
<<
[OT] PCVA: The ebnews article from November was a good (bearish) reference on competition in the DSP IP space. Good find. Dew
That’s PCVA. Sorry about typo.