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.
Because analysts’ ratings are a lagging indicator
>> Exactly! So what use are they? <<
If nothing else, they are a source a comic relief.
Corning (GLW) is a pretty extreme case. The analysts who hated GLW at $1.10 in early Oct. upgraded it when it went over $4. (It’s now at about $5.)
Some things never change…
>> Seems to me those computers were more than 4 years old a couple of weeks ago... Why wait until now to upgrade Intel? <<
Because analysts’ ratings are a lagging indicator
>> I have, on several occasions, tried to start a discussion about strained silicon and how the Intel transition might go. You guys seem to have no interest in discussing it. I am not sure if that is because you have your heads in the sand, or because it has nothing to do with bashing AMD. <<
Please count me in for any discussions on SOI, strained silicon, and related process technologies. Regards, Dew
>> SOI has never been done in high volume manufacturing. You won't find anything specific to [the SOI process at] AMD, other than rumors… <<
EP: the rumors are what I am most interested in. t.i.a. Dew
>> Hammer is a poor design and AMD's SOI process doesn't work. <<
I would appreciate any elaboration about AMD’s SOI process that led you to your conclusion. T.i.a. Dew
There are some worthwhile responses on the Yahoo MB. The thread starts here:
http://messages.yahoo.com/bbs?.mm=FN&action=m&board=1602585284&tid=tmta&sid=16025852...
Major omission in ad for Compaq Tablet PC:
I sent the following letter to TMTA’s Investor Relations. I believe the letter is self-explanatory:
>>
I am writing to express my displeasure with the wording in the advertisement for the Compaq Tablet PC in the 9 Dec 2002 issue of Business Week. That issue contains an 8-page ad spread for the Compaq Tablet PC beginning after page 60. On the last page of the spread is the following text:
“The Compaq Tablet PC. Work the way that works for you. With a 1GHz processor, 256-768MB of memory and up to a 60GB hard drive, The Compaq Tablet PC does what you need…”
--
Although the overall content of the ad is excellent, it is disturbing to see that HP has chosen not to mention the name of the processor used in the Compaq Tablet PC. Instead of saying, “With a 1GHz processor,” the ad could have – and should have – said: “With a 1GHz Crusoe™ processor from Transmeta…”
One gets the impression that HP is ashamed to mention the name of the processor used in this very impressive machine. And that strikes me (and perhaps other readers) as shortsighted. Sooner or later, prospective buyers of the Compaq Tablet PC are going to learn that the machine is powered by a 1GHz Crusoe. When these people investigate further, they will undoubtedly grasp that the Crusoe is the ideal processor for this device and a shrewd selection by HP.
So why not tell people straight away that the Compaq Tablet PC is powered by a Crusoe and turn this into a selling point, rather than something to be concealed?
I look forward to your response as well as the possibility of improved ad copy in future versions of this advertising campaign.
<<
>> Orders of computers and electronic products posted their largest increase in a year, rising 6.2 percent…<<
Not complaining about a 6.2% increase. But the base period (Oct 2001) was a pretty slow period, making for a relatively soft comparison.
wbmw: Is this the same old SpeedStep?
http://www.theinquirer.net/?article=6396
This article by our friend, Mike Magee, gives specs for the various Banias and Dothan chips in “performance mode” and “battery mode.” Magee makes it sound as though Banias/Dothan have only two voltage states a la SpeedStep.
When we discussed this at some length a few months ago on the TMTA MB, you posted that Banias will have more than two voltage states and will operate in much the same way as Crusoe’s LongRun, which has multiple voltage states and dynamically adjusts the voltage and power according to the CPU load.
A clarification would be appreciated. T.i.a. Dew
>> I thought that statement was particularly FUD-like, especially considering that actually premium was right smack in the middle of what you said was "typical".<<
spokeshave: Perhaps I am at fault for overemphasizing the conversion premium as a percentage rather than the raw conversion price of $7.37 -- which is the number that ultimately affects the dilution of common shareholders from conversion of the notes.
Considering the precipitous decline in AMD’s stock price this year, the conversion price of $7.37 does strike me as shockingly low. After all, AMD was above $20 in early 2002 and was above $10 as recently as August. For AMD shareholders, it is unfortunate that AMD waited until now to raise capital on terms that are far less favorable than what would have been available fairly recently. Regards, Dew
spokeshave: Agree that the pricing date was the 20th. (Not sure why you mentioned the 19th.) To analyze AMD’s convertible offering even further, perhaps we should look at the price on the 18th – before the announcement that AMD was raising capital. AMD closed on the 18th at 6.52 and then fell about 15% in the next two days; the sharp drop was due in part to the fact that AMD felt the need to raise capital.
I certainly did not mean to mislead anyone with my original post that AMD’s convertible has what I consider a low conversion price. I apologize to anyone who may have interpreted my post as some kind of FUD. I stand my statement that AMD’s convertible shows a considerable degree of desperation on the company’s part.
As for the discussion on insider trading, it goes without saying that we were talking about fiduciaries. Lower-level insiders are generally bound by trading restrictions only as imposed by the company itself. Regards, Dew
It doesn’t matter that the notes were first offered on 11/19 (when AMD’s price was lower). What matters is the date when the terms of the converts became fixed (the so-called “pricing date.”).
Re Sanders’ purchase: I should have left out the word “bad.” Insider buying or selling immediately before the release of material news – whether good or bad – is a violation. This might seem counter-intuitive; i.e. one might expect it would be ok to buy before releasing bad news or sell before releasing good news.
In practice, however, “bad” news is often treated by the market as positive (if it is not as bad as feared) and, similarly, good news is often treated as negative (if the market was expecting even better news). Hence the SEC’s policy of prohibiting any insider trading based on non-public material information is a sensible one. Regards, Dew
AMD odds and ends:
>> The notes are convertible into the company's common stock at an initial conversion price of $7.37 per share.<<
This shows just how close AMD was to being locked out of the capital markets entirely. Typically, 5-year convertible notes paying 4.5% interest from a company like AMD have a conversion price of 25% to 50% above the current stock price. But AMD’s notes have a conversion price which is below the current price.
What does this mean? That the buyers of these notes don’t think much of the chances for substantial appreciation in AMD’s stock price over the next 5 years.
--
Re Sanders’ stock purchase: note that Sanders waited until after AMD announced the layoff and special charges before buying. This protects him from an insider-trading violation. Had he bought just before the release of bad news, it would have been an actionable offense. Dew
>> Sanders bought over 300,000 AMD shares on Friday!!!!!!!!! <<
He knew the purchase would warrant a story in the newswires. That guy just can’t stay out of the limelight.
Cover story in The Economist: “Computing’s new shape”
Here is an excerpt from the article’s lead paragraph:
>>
“A computer on every desk and in every home.” This was Microsoft’s mission statement for many years, and it once sounded visionary and daring. But today it seems lacking in ambition. What about a computer in every pocket? Sure enough, Microsoft has recently amended its statement: its goal is now to “empower people through great software, anytime, any place on any device.” Being chained to your desktop is out: mobility is in.
<<
--
Sounds like a rather considerable tail wind for TMTA.
wbmw: as I said in an earlier post, company-sponsored benchmarks should not be taken too seriously. I think Astro will more than hold its own versus Banias and the rest of the competition based on a composite of pertinent metrics: price, application performance, form factor, power consumption, and heat.
Launch in 3Q03 is what TMTA has been saying officially since Astro was first announced. However, Astro may be slightly ahead of schedule because it was not originally expected to be showcased as early as this year’s Comdex. FWIW. Dew
Astro ASP = $120?
In a recent press interview, Matt Perry stated that the CPU for a mobile-computing device should sell for no more than $120 to stay competitive in the end-user price range for such devices.
Perry was supposedly lambasting the high prices for Banias, but he may also have been giving us a clue about the anticipated ASP for Astro.
Given the recent disclosures about Astro’s impressive application throughput, an ASP of $120 seems entirely reasonable. This would be almost double the $64 ASP of the TM5800 during 3Q02.
The second half of 2003 could see some serious ramping in revenues based on the multiplicative effect of higher chip volume and a much higher ASP. Profitability on or ahead of schedule is looking more and more likely with each passing day. Dew
Thanks for the additional info. In your month-by-month chart of semi equipment, BTB rose during Q4 in five of the seven years shown. I guess you could call that a “weak” seasonal effect in the sense that seasonality seems to be far less significant than the position in the economic cycle.
Does anyone produce semiconductor bookings and billings which have been seasonally adjusted? If not, would it be helpful to present the data that way?
Does anyone produce semiconductor bookings and billings which have been seasonally adjusted? If not, would it be helpful to present the data that way?
>> AMD (or INTC) should have bought TMTA when it was under $1.<<
bababouie: yes, it would have been a heck of a deal for an acquirer to pick up TMTA for less than $1. But there is absolutely no way that such a deal would have been acceptable to TMTA’s B.O.D.
TMTA founder Dave Ditzel and his core engineering team have spent nearly a decade developing TMTA’s technology. (Ditzel also happens to be the Vice-Chairman of the Board and a substantial shareholder.) The code-morphing software used in Crusoe and Astro is probably the most complex engineering project these developers have ever worked on.
Trust me: these folks believe strongly in TMTA’s prospects and they have no intention of selling out cheaply. Even at a much higher price, I doubt that TMTA would seriously entertain a sale of the entire company. At the right price, however, TMTA might be willing to consider a substantial minority investment by a corporate partner.
>> AMD could have done as a stock deal and they could now have a good competitor to Banias and they could even fab Astro @ Dresden and probably get more Mhz out of it than the foundries do. <<
Since I’ve already addressed the buyout issue in general, let me speak to an AMD buyout, specifically. I believe that, for most TMTA shareholders, AMD paper would be about as appealing a proposition as the Quilted Northern stuff hanging in the bathroom. Regards, Dew
Does this mean that you will return to the TMTA MB?
>> Dew, good find. I didn't know that the demo was running at 1GHz at the time. <<
Actually, mmattheych on the Yahoo MB deserves the credit for finding this. Agree with you on two scores: 1) It’s impressive that the Astro was running at 1.0MHz rather than 1.4 as some of us had inferred; 2) Company-sponsored benchmarks should not be taken too seriously.
>> Still - it may not hurt to have some TMTA as a hedge, given the new information.... <<
Right on, bro…
More on Astro vs Banias:
[Thanks to mmattheych on Yahoo for catching the info from Prudential]:
http://www.investorshub.com/boards/read_msg.asp?message_id=591697
More on TMTA's Astro:
wbmw: Maybe the performance penalty of TMTA’s CMS is not so great after all in the latest version of the CMS used with Astro.
Prudential Securities has evidently seen the same demo of TMTA’s next-gen Astro chip as tomshardware. Here is an excerpt from Prudential’s current rating on INTC:
http://investor.cnet.com/investor/brokeragecenter/newsitem-broker/0-9910-1082-20686826-0.html
>>
· Transmeta unexpectedly unveiled working silicon of its next generation ultra low power Astro x86 processor for the notebook market. This chip appears to be much more ambitious in terms of the addressable market for Transmeta.
· Astro's performance appears to be quite good with a 1GHz clocked chip easily beating out a P4 mobile at 1.8 GHz in a demo Transmeta showed to us.
· The issue with Intel's new Banias mobile solution is price performance that may not cut it given the competitive environment and Intel's own P4/PIII mobile products. The other issue is product positioning that is inconsistent with Intel's historical focus on clock speed rather than overall "performance".
<<
Regards, Dew
>> I believe I addressed the innate level of parallelism found in Transmeta's ISA, which is exploitable through CMS…
[from prior post]: the unfortunate end result is that ILP gains from ISA optimizations are limited. While no doubt greatly improved over an archaic legacy architecture such as x86, newer ISAs still have not been able to achieve 2x performance gains from doubling the execution width<<
It’s a matter of degree, of course. My point is that TMTA’s code-morphing software has more flexibility than an all-hardware solution when it comes to actually achieving the performance gains inherent in parallelism. Regards, Dew
wbmw: thanks for your detailed response. Your analysis is thorough except in one area: you overlooked what TMTA’s code-morphing (CMS) software is capable of adding to the mix.
It is the CMS which has the responsibility of assuring that the VLIW is able to exploit parallelism.
Although I admit that the 2.8X figure I cited (Astro throughput relative to TM5800) is merely a theoretical bound, the throughout advantage of Astro vs the TM5800 in practice should be better than your analysis suggests. Regards, Dew
>> Astro is going to clock 40% faster than the TM5800, so the overall power dissipation will probably be about the same (~7w). <<
Just to be clear: Astro has double the VLIW as the TM5800; hence, a 40% faster clock means that Astro has 2.8X the throughput of the 5800. Now, 2.8X the throughput with the same power dissipation is a mighty big overall improvement if you ask me. Dew
>> WiFi is not being integrated directly on the Banias CPU. It is being supported by a separate chip called Celexico that will be bundled with Banias <<
Thanks for the correction Am I the only one who thought WiFi was going to be on the CPU? I recall reading that Banias will have WiFi “built in” and I interpreted that literally. (I’m a software guy rather than a hardware guy by training, so I’ll believe almost anything when it comes to chip design.)
Let me rephrase the question: does Celexico offer any cost advantages versus the way others (e.g. TXN) will implement WiFi?
What is the cost advantage of putting Wi-Fi support directly in the CPU as Banias is supposedly doing?
Astro faster than 1.8GHZ Pentium 4M!:
http://messages.yahoo.com/bbs?.mm=FN&action=m&board=1602585284&tid=tmta&sid=16025852...
wbmw: Thanks for the heads-up on TMTA’s Astro. I posted the tom’s article on the TMTA MB on Yahoo. Let’s see how many reco’s it gets. (Anyone want to guess?)
On a related subject, you might find this amusing:
http://messages.yahoo.com/bbs?.mm=FN&action=m&board=1602585284&tid=tmta&sid=16025852...
Regards, Dew
wbmw: thanks for the PM. (I cannot respond via PM because that feature is available only to paid subs.)
Surprised that no one else commented on my post about AMD’s “allowance” against deferred tax assets. Although the allowance is a non-cash accounting action, it is effectively an announcement that AMD does not expect to be profitable in the foreseeable future. That would seem to be something worthy of comment by followers of INTC. Dew
>> and the second piece covered the establishment of an [AMD] allowance against deferred tax assets of $263 million. <<
I’m going out on a limb because, until we see AMD’s detailed financial statements, we won’t know for sure. But taking an “allowance against deferred tax assets” essentially means that the assets in question are being wiped off the balance sheet. Such an action is usually an ominous signal.
Here’s why: “deferred tax assets” arise from tax loss carryforwards, i.e. past operating losses which can be used to offset income taxes on future profits. However, a company must actually generate future profits in order for such tax loss carryforwards to be worth anything. Hence, a decision by a company to write down such deferred tax assets is in essence a statement that profits in the foreseeable future are unlikely to be large enough to fully utilize the tax loss carryforwards.
So you can see why this is indeed an ominous sign for AMD. FWIW. Dew
Great business model if they can pull it off. Much cheaper to sell old stuff than design and build brand new products.
Perhaps TMTA will re-christen these ol’ TM5600’s with a spiffy new tag like “TM5600E” (for “embedded”). Dew
I had assumed that the TM5600 was defunct. It hasn’t been mentioned at all in the last several conference CC’s. Anyone have a read on this?
>> NEC using the 1GHz TM5800 is great evidence that TMTA can produce the fast chip in quantity, and for OEMs other than HPQ. <<
It will be reassuring to see these NEC 1GHz-machines actually on the shelves. Huge relief that there is evidently no supply problem (?any longer).
It’s odd that the article gives a price for the NEC “eco” in the Japanese market but not for North America. The older model using a 900Hz Crusoe is widely believed to be overpriced. Dew
No idea. Why is Krispy Kreme under drugs?
Excellent analysis, wbmw. The only thing I would add is that what Hector says in an interview with the press might be somewhat different from he says (or thinks) in private. FWIW. Dew