Monday, April 30, 2007

Rackable Systems (RACK) Priced to Scoop Up

Rackable Systems (RACK) had a bad Q4 2006 in which revenues hit a record but net income plunged (see my analyst conference summary). Last week it reported Q1 revenues that plunged and a bloody net loss (see summary). I bought stock at several points last year; it has been a major drag on my portfolio. So why am I scooping up as much as my portfolio rules allow? Am I still trying to catch a falling knife?

Could be. Falling knives sometimes fall to zero; and that has happened to stocks in my portfolio in the past. That is one good reason for diverified portfolio rules. If you take three risky positions (normalized to 1) and one falls to zero but one goes to 2 and one goes to 4, you have doubled your money with very little net risk. The question is, did you do your research, do you really know what the risks are with each stock and therefore with the portfolio.

In an auction system like the stock market a substantial portion of stocks will have prices that are unsustainably high or low. When there is enthusiasm stocks can get priced as if their revenues have already grown by factors of 10 (or back in 1999, by infinity for some stocks). The same thing happens when people all dump stock on bad news: often the auction value goes far beneath the underlying value of the stock.

This is easy to see with Rackable Systems. On Friday, April RACK ended at $11.27 per share giving it a market capitalization of $322 million.

The company had $170 million in cash on March 31, 2007, which is 52.8% of its market capitalization. It has positive cash flow in Q1 of $9.6 million. If it starts making money again its effective tax rate will be very low due to NOLs (Net Operating Losses), which are losses it accumulated during its startup phase that can be credited against future profits.

Q1 was its worst revenue in a year, yet it had $72 million in revenues, or an annual run rate of $288 million. Q1 is typically seasonally weak.

Belt tightening alone can restore the company to profitability. As I explained in Rackable: Playing with the Big Guys, server technology competitors saw how well RACK was doing. Unable to compete on technology, at least one began bidding at below cost in order to stop Rackable in its tracks. In Q4 Rackable's management went for market share, reducing bids in order to keep clients. Since then they have realized that failure to make a profit is just as great of a threat as losing market share. So now they are continuing to compete with their superior technology, cutting their own costs, but refusing to match what management called "irrational competitor pricing."

I expect that, with its cash cushion and existing level of sales, even if things are tough going forward Rackable's stock price is at or near its low.

The upside potential is substantial. One good quarter (revenues above $80 million with GAAP net income of over $2 million) should send the stock price into the high teens.

Rackable is expanding from just selling servers to selling storage equipment as well. They are expecting about $20 million in storage revenues this year that they did not see last year. Storage is a very competive field, just like servers, but reports are Rackable has great technology.

Then there is always the possibility of a buyout. Rackables technologies are protected by patents. For Sun or HP an acquisition of RACK would make a fair amount of sense given the ability to acquire customers, cash, and NOLs all in one swoop.

Thursday, April 26, 2007

When Will Housing Recover?

Here in California, you can feel the panic. It is beginning to look like housing prices are not going to drop substantially, except in a few specific localities.

And why should they? Despite the building frenzy between 2001 and 2006, there is an overall shortage of housing in California. Right now it is hard to find rentals, much less reasonable rentals. Plenty of houses are for sale, but the prices can't be said to be bargains.

I know many couples around the state who have high incomes ($50,000 per year or higher), live in apartments, and have been hoping to buy a house for years. 2007 finally seemed like the year of the promised decreases in house prices. Some individual houses for resale have lowered their asking prices, but statistics for April show that the median statewide housing price was flat compared to year-earlier.

If housing prices were to drop, buyers would step in. As people stop hoping to save by waiting for a drop, the market overhang will be absorbed.

Since prices did not drop, even most of the fools who bought in 2006 have not lost any money.

There are some exceptions, mostly large housing developments built in 2006 in the outer reaches of the Central Valley.

All this fits into a pretty pleasant macroeconomic picture. Jobs are plentiful. For those who are steady workers and have saved a down payment, 30 year interest rates are historically low, near 6%. Having heard horror stories of people who bought housing on at adjustable rates, sane people are thinking about how nice steady monthly interest payments would work out over 30 years.

It does not appear that things are much worse in the United States as a whole. Immigrants, legal and illegal, create a constant push into the housing market. Some areas like Florida were severely overbuilt, and some cities have lost jobs and will continue to have weak housing markets as a result. But already the New York City and Boston housing markets are heading back up.

As to the subprime mortgage market, it is a problem for the fools who did the lending, but its impact on the California market will be marginal at best. There are people with real money who want those homes, especially if they can get them at even the slightest of discounts.

The only thing that could substantially depress prices in California as a whole is a recession. But with employment solid, consumer spending okay, and the rest of the world's economies booming, that is an unlikely possibility.

Note to the Dow is at record levels. That means some people who were worried about the value of their portfolios back in 2002 (even if they did not over invest in Internet stocks) are not so worried.

In a way it is too bad. Housing is astonishingly expensive in California. It is hard for people to get started just out of college, or if they are spendthrifts. But those have always been the choices: spend or save and invest. A job, few years of saving, a down, and a thirty year mortgage: that was a formula for success 40 years ago, and it is still a formula for success.

Tuesday, April 24, 2007

Altera, Texas Instruments See End to Chip Slowdown

Texas Instruments (TXN) and Altera (ALTR) issued quarterly results and held analyst conferences yesterday. Both have reason to believe that general demand for semiconductor chips hit bottom in Q1 2007 and is rising Q2. Are they right?

You can see my full summaries of the conferences from these links: TXN; ALTR.

Texas Instruments (TXN) is the bigger player, with Q1 revenues of $3.2 billion, down 8% sequentially and down 4% from Q1 2006. They believe Q2 revenues will be $3.32 to $3.60 billion, which would be up 3 to 12%. Their main reason for the belief is that daily sales were up 20% in March and April compared to February. It takes time for a sale to be shipped and booked as revenue, but some of that 20% increase should start appearing in Q2.

Altera (ALTR) is more of a niche player, specializing in programmable logic devices like FPGAs and CPLDs. Their Q1 revenue was $305 million, down 4% from Q4 2006 but up 4% from Q1 2006. The guided revenues to 1% to 4% sequential growth for Q2. They were more cautious about the future, saying they are seeing a slow rebound, but some end markets are still weak.

Both companies and many other chip makers argued that the downturn that started roughly in Q3 2006 was more a result of inventory management changes than long-term demand weakness. It seems that in 2006 many finished goods makers decided to move to just-in-time acquisition of parts. In the transition to that they used up their own inventories and paused their buying. They only started ordering specific parts when they were ready to use them. This caused a slump in demand at the manufacturing end. Yet the chip makers could not cut their own inventories because the end users still wanted rapid shipments after orders were made. Another factor was macroeconomic fear for 2007. So far those fears have been wrong. Consumer demand has been steady in the U.S. and markets like China, India and Germany have been booming.

Nothing is certain in the world of economics, but there is a consistent story here: in order to meet demand late in 2007, end-product makers are going to need to seriously bump up their chip orders. They can no longer just reach into their own inventories, which are too lean to allow that.

Some end markets may be stronger than others. Texas Instruments complained (following Motorolla, which it supplies chips to) that the cell phone market has been increasingly difficult to make a profit in. Their solution is to continue to integrate more functionality on each chip to cut costs and restore the profit equation.

Altera is in a better position in some ways than TI. Their products are more specialized. But in Q1 that had a 35% sequential decrease in their Hardcopy chips for computer storage systems. They believe that is a one-quarter event, but added that since these chips are custom manufactured, it takes some time to go from an order to shipping and getting paid.

I have to wonder how wise it is to fight it out in the cell phone market. It is true that it is a huge market and 3G is just coming into play. But it seems that a lot of chip makers have their eyes on the same pie.

Friday, April 20, 2007

Google and Microsoft

Google (GOOG) reported Q1 2006 results yesterday. The headline numbers hid some sequential weakness in earnings growth, and so far Google has not shown significant revenue beyond its search advertisement and AdSense franchises. [I don't own Google stock, but I use AdSense.] There was some good discussion with analysts. You can check out my summary of the analyst conference at

With the possible exception of Oracle, of all the "next Microsofts" investors have bet on in the last 20 years, Google has come closest to fulfilling that promise. Though it had years in the wilderness, it market cap swelled for more quickly than Microsoft's did in the mid-1980s.

Meanwhile, investors show little confidence in Microsoft (I own the stock and worked freelance for Microsoft from 2000 until 2006). Microsoft is due to report on April 26th, with the analyst conference beginning at 2:30 PM Pacific Time (see my MSFT page for previous conference summaries, and this one when it is posted). Today, using figures, Microsoft stock is selling at 24.5 times trailing earnings and Google stock is selling at 49 times trailing earnings. Clearly that is a vote of confidence in Google's growth potential.

Microsoft revenues in Q4 2006 were $12.5 billion. Google had $3.21 billion in revenue the same quarter, so Microsoft is still about 4 times larger. It is also much more diverse. This company that began by selling the BASIC programming language to early microcomputer enthusiasts made its first really big money when it was selected to supply the operating system for the first IBM brand personal computer. That operating system has evolved into Windows Vista. In the meantime Microsoft added its Office Suite, the Visual Studio programming interface, and the XBox 360, to name some notable successes. It has also had some failures.

One big question going forward is whether Microsoft can gain market share in the highly-lucrative search market. Which is to say, in the targeted ad market. While the crew at Google is quite impressive, there is some real danger to Google, and not just from Microsoft. Yahoo has a new search algorithm for ad placement, and smaller search engines like Ask should not be ignored. But Microsoft's profits would go up if it abandoned Internet search altogether. Google has not yet proven it can make any other revenue model stick. Given its bankroll and its brains, I do expect Google to bring out products (or turn some existing products) with revenue streams that will please investors. But Google's big acquisitions, YouTube and DoubleClick, are extending the current franchise rather than branching out into truly different IT areas.

Microsoft had, for all practical purposes, a monopoly on microcomputer operating systems that it still has today. Linux is still more of a threat than an actual competitor. Google has no such safety margin. Microsoft search is not a danger to it, but Yahoo, another "future Microsoft", in many ways is ahead of Google in diversification and has a truly substantial share of the search market.

The days of hyper growth at Microsoft are over, but the stock pays a nice dividend, there is a huge stock buyback program, and growth keeps happening. Google has more potential, but it also has more risk. Choose either, both, or neither; you will probably be fine.

Wednesday, April 18, 2007

IT Outlook: IBM, Intel, Yahoo Report

Yesterday, April 17, 2007, IBM, Intel, and Yahoo reported 1st quarter results and held their analyst conferences. You can look at my summaries of the analyst conferences if you would like. See IBM, INTC, YHOO.

Mostly this blog is about revealed trends and their implications for other information technology companies, but first let's look at what was revealed about the companies themselves.

IBM did really well in Asia, but revenues were essentially flat in the United States. In total revenues were up 7% from year-earlier. IBM expects EPS growth to continue at a 10 to 12% annual rate, but that depends on getting the US back on track. Their software and business services segments did best, up 9%, with their hardware segment lagging, with 2% annual growth. Within that segment high-end servers did better than low-end servers.

Intel had revenues down 1% from a year ago and down 9% sequentially (Q1 is usually a sequentially down quarter for them). But they seem to have dodged the AMD bullet. If AMD had continued to gain market share, Intel would have seen lower revenues. Flash products are still losing money for them. Q2 to be down about the typical seasonal 4%. Some remarks about selling to systems integrators seemed in the direction of saying Intel can still play hardball despite the AMD v. Intel lawsuit on anti-competitive practices. Unit production costs of microcontrollers fell faster than average sales prices (ASPs) fell, so margins were up. Server chip ASPs took a bigger hit than desktop/notebook ASPs, probably because AMD is more competive in server chips.

Yahoo revenues were up 7% over year-earlier, and down only 2% sequentially. Considering the difference in Q4 v. Q1 advertisement, that is not too bad. But it certainly does not show any market share gain against Google (See my GOOG page). If Google gained share this last year (and it almost certainly did) it was at the expense of Microsoft. Yahoo execs characterized the online ad market as going through big transitions.

Those big transitions are the news for 2007 and 2008 against which everything else must be measured. It isn't just MySpace and YouTube. Nor just the Internet. The whole global information processing structure is going through one of those periods of rapid transformation that may create totally new winners and thrust some of yesterday's winners into the loser column.

Chip and hardware makers will benefit as ever more storage and processing power is needed. While desktop processing and storage is more than adequate at this point, servers are barely keeping up, despite dual and quad processors and virtualization technology. Sun, Rackable, Dell and others are doing their best, but don't be surprised if some relative unknowns make breakthroughs in the next few years. AMD and Intel could fly if they compete on technology instead of price (which Intel signalled it is willing to do; AMD was doing it until mid 2006 and would probably be happy to return there). Innovators like Marvel and Maxim should do well -- if they make the right innovations.

While Google looks set to remain king of the online ad roost for the short term, that does not mean a swarm of innovators aren't going to be stepping up in the ranks. Not all the money is directly in ad serving, as Akamai has proven.

Asia is set to become a world leader again. With Japan, China and India all on the upswing, the United States will be technologically challenged. This will be partly offset by globalization. Innovators will do well whether in the US or India; laggers will be crushed.

I expect chip makers in general to have a relatively good Q2 and a great Q3. Server-focussed computer makers will do better than client-focussed makers. New media companies will continue to emerge rapidly. This means that venture capitalists will have a substanial advantage: the real money is made (or lost) before the IPO is set up. But stock market investors can also do well by spotting companies that are good at buying small innovators before the markets see their true value. Yahoo and IBM are both in that class.

You can check out the list of companies I do analyst conference summaries for at

Monday, April 16, 2007

Red Hat (RHT): Growing Value in Linux?

Red Hat (RHT) was the darling of investors in 2000: they paid over $150 per share for it, even though it had not shown a profit and had only a few million dollars in annual revenue. Now it is no longer even part of the Nasdaq 100 index. Is Red Hat an old dog no longer worthy of investor attention? I don't think so. Yet I would not characterize the stock, at this point, as very underpriced.

To get management's current assessment of themselves see my summary of their March 29, 2007 analyst conference.

I think many investors, and sometimes even professional analysts, fail to pay enough attention to history. Going back two or three years in a company's history, or even going back a full decade, can give some real insight into the present and future. Red Hat's past is a glorious example of how investor enthusiasm, or disregard for fundamentals, can affect a stock's price.

Today Red Hat has a market capitalization of $4.3 billion, based on a price of $22.28 per share. Full 2007 revenue was a bit over $400 million, so the stock is valued at over 10 times revenues. GAAP net income was only $20.5 million in Q4 2006. If you believe in Non-GAAP net income, you could use $32.7 million instead. Let's compromise and say 2007 net income will be in the range of $100 million. That gives an EPS of over 40 per share.

So unless you think revenues and profits are going to grow rapidly at Red Hat, you would be better off putting your money in a CD.

But revenues may grow rapidly. They were up 40% between Q4 2005 and Q4 2006. If net income grows in a similar fashion, and you are willing to pay prices now based on projections two or three years out, Red Hat's price today could be proven to be undervalued. I also like the fact that they have $1 billion in cash.

How can you inform a decision where the numbers are unclear? In this case you need to know about technology and about history.

The reason so many investors were buying Red Hat in 1998 (as venture capitalists) and in 1999 and 2000 was simple: it was billed as the new Microsoft (See my Microsoft page). Most investors, including me, regret not buying Microsoft stock when we were young. But there was a reason we did not buy Microsoft: there appeared to be better opportunities available. A whole bunch of companies in some aspect of the microcomputer business in 1983 ended up being nearly valueless by 2000. The big exceptions were Intel, Microsoft, Apple, and Compaq (which is not part of HP).

In retrospect I can be as brilliant as any analyst: I recommend that you buy Red Hat at $3 per share in July 2001. But there were good reasons RHAT did not look to be worth even that at the time. Neither past revenues nor EPS justified it. Only future growth justified it.

Whatever the technological merits of Linux, the problem with it from an investor standpoint is that it is free. If you want you can build your own Linux server package and deploy it to as many servers or clients as you have and not pay anyone a penny. Unlike Microsoft Vista or Server.

But Red Hat is making money. They are able to charge for their support services and do charge for their enterprise Linux distribution. Most companies find it cheaper to pay Red Hat or another Linux provider for their package and expertise than to try to do it "free" in-house because the cost of labor is greater in-house than when outsourced. It does not make sense for hundreds of thousands of IT people around the country to be fiddling with Linux details when one person at Red Hat can do it for them.

It is clear that Linux is a very real competitor to Microsoft. It has mostly driven all other versios of UNIX from the field except for Sun's Solaris.

So the real problem is predicting Red Hat's future growth is their competition within the Linux market. Just as Napster, RealNetworks and others fight over the scraps left from Apple's dominant music download position, a bunch of companies, small and large, fight over the Linux market.

Sun, of course, is a formidable competitor. They recently made their Solaris product free, believing they can make money on support services. Oracle is now in the Linux business; no one says they are not a formidable competitor. Both Sun and Oracle have a lot of clients to cross-sell to gain Linux market share.

So I would not place any big bets on Red Hat unless you know something the rest of us don't. I think, on the whole, it is slightly underpriced, but that assessment jumbles a lot of variables that are hard to quantify. However, if Red Hat continues to grow in 2007 as fast as it did in 2006, then the stock should move up pretty quickly.

See also my Red Hat page.

Thursday, April 12, 2007

What Price Dendreon? (DNDN)

Nobody knows today what, exactly, the correct price for Dendreon is because nobody knows what the FDA will decide on May 15. I'm not one of those people who believes that the market price is always right. I believe much of the time the market is out of equilibrium for one or more of several reasons, the prominent one being human miscalculation. With sell-side (brokerage) analysts and even the great Cramer miscalculating just about as often as the rest of us.

I called Dendreon right, which you can see in my blog of March 30, 2007 and from my coverage on my analyst conference summary pages (See DNDN). But before I update my thoughts, let me tell you a different biotechnology story.

There is a company called Genta (GNTA) trading today for $0.34 per share. That gives it a market cap of $62.5 million. They have $29.5 million in cash to support that, plus a drug candidate tradenamed Genasense. But they aren't a small startup that is going to use the cash hoping to get some good clinical results. Genasense had major Phase III clinical trials for 3 types of cancer. They had a larger pipeline. They had a market capitalization of about $2.8 billion five years ago. They were a "next big thing" based on some early results for Genasense. They told investors that Genasense had great results in the first completed Phase III trial, for malignant melanoma. They claimed they met primary endpoints and showed prolonged survival of patients.

Only when the FDA looked at the data, they tore it apart. The study as not double-blind and there were solid reasons to believe there was investigator bias. The FDA decided there was no positive significant result that outweighed a side of the studies that Genta press releases avoided: some "adverse reactions" (side-effects) that included, in a different indication, an almost doubled death rate from adverse reactions.

If you look at the success rates for drug candidates, the statistics are grim. Lots of promising test tube drugs fail when given to animals (pre-clinical studies). The FDA requires three phases of clinical (people) studies: I, II, and III. At each phase far more candidates bomb out than go on to the next phase. It is not real surprising to have an apparently favorable (apparently to investors, based on information released by drug companies) drug with good Phase III results turned down by the FDA. Then there is always the question of whether an approved drug will actually sell. Finally, when widely distributed, the drug could cause adverse reactions not seen in the studies, resulting in the FDA requesting that it be re-evaluated or pulled from the market.

Each drug costs a lot to develop and test. Larger companies know they are in a crap shoot; with lots of candidates, some work out and some don't. Biotech investors should know it is a crap shoot. No matter what the merit of a drug candidate, it is going into an extremely complex machine, the human body, and something could go wrong.

When I bought Dendreon I knew the odds were against approval of Provenge. In fact, the FDA can still gut Provenge, despite the positive recommendation of the advisory committee.

So why am I holding on to Dendreon stock when I could sell it today for three times what I bought it for? Well, for one thing I cut back my holdings; I did take some profit on the run up after the good news.

The main thing is the odds. Given my knowledge of the FDA (I first worked for a biostatistics firm some 30 years ago, but I only started looking closely at these types of decisions recently) my guess at the odds is: 50% chance of outright approval; 40% of getting a "letter" (requiring more study for full approval); and 10% of outright rejecting Provenge. This range of odds is based on the proven safety of Dendreon plus some fair indications that it is effective. That is good enough for dying cancer patients without other options, and it may be good enough for the FDA.

The 40% spectrum leaves us in the broad price range we are in today, $12 to $25 per share, depending on the exact wording of the letter. The 10% chance puts us back at $2 per share. The 50% approval chance puts us over $25. How much over in the short run depends on investor enthusiasm, something I have been weak at predicting. In the long run the price depends on if Provenge is successfully marketed. More important, can Provenge's success in a particular type of prostate cancer be extended to other types of prostate cancer and to other cancers as well? If the answer is yes, then the sky is the limit for Dendreon. If they can make cancer immunotherapy work it may take years to get approval for other types of cancer, but the long-term value makes it more than worth the risk that the FDA will fail to approve Provenge in its coming meeting.

The market is often wrong, and that can be in either direction. Back in 2000 Dendreon traded in the mid-20's, before it even had anything much more than a theory to work with. In the summer of 2002 it traded below $2 per share. For the last few years it has mostly been between $4 and $6 a share. That isn't just volatility. That is risk. More peculiarly, it is a difficult-to-quantify risk. So decide how much risk you want in your portfolio before you decide what your buy and sell prices are for Dendreon.

Thursday, April 5, 2007

Rackable (RACK): Playing with the Big Guys

Today Rackable (RACK) had plunging stock price syndrome after the company announced that Q1 2007 earnings would be below expectations, that is to say, negative. How negative, we will have to wait until the earnings release.

At the analyst conference for Q4 2006 (see my summary) guidance for Q1 2007 was revenues between $70 and $75 million due to seasonality and lumpiness in largest accounts. They said to expect gross margins at low end of range. There are lots of details about why Q4 worked out the way it did and what Q1 might be like in the answers to analyst questions.

What did the press release say? "Total revenue for the first quarter of 2007 is expected to be in the range of its previous projections of $70 - 75 million." No change there. But "the company anticipates a net loss for the first quarter of 2007, on both a GAAP and non-GAAP basis," because "gross margin for the first quarter [is] to be approximately 30% lower than previously communicated." The details: price competition, a cancelled order, a severance package, and some tax obligations.

Management leaves some bright looking crumbs for investors: a greater than one book to bill ratio, the strongest backlog of orders ever, and cash balances improving to $170 million at the end of the quarter.

So how much is the stock worth today (which is to day, should you buy it because it is beneath that price, or should you sell it if you own it above that price)? It was at $55.75 per share within the last year; today the stock is below $15. What changed?

Rackable is in a position that is common in all industries, including technology. It had a good idea and executed well on it. It grew rapidly, from $53 million in annual sales in 2003 to $360 million in 2006. Investors (including me; I own the stock) got overly enthusiastic by projecting a straight line of increasing sales and profitability.

Sales revenues had not been bad, continuing to increase through Q4 2006. But Q1 2007 revenues will not only be sequentially (and seasonally) down; they will be below Q1 2006. In addition, starting with Q3 2006 earnings have taken a beating.

Rackable was noticed by the big guys because it was becoming a big guy. It was taking out full page ads in Infoworld; it was winning big contracts from YouTube, Microsoft, and other Internet search/portal players who liked Rackable's innovative approach to server farms. Sun, IBM, HP and Dell took a look and said: this has got to stop. Any one of these companies has the financial resources to take some losses in order to underbid Rackable. And reports are that this has been happening lately.

In order to keep its customers Rackable has cut its prices in specific situations. But it does not sell servers one at a time. It sells systems on a large scale. A few price cuts in a few specific situations and it is in the red.

Just as when the stock was bid up high based on future expectations, the pricing of the stock today depends on your model of future sales and earnings. No one actually knows what will happen; but some of us will guess right, and will at least guess in the ballpark, and some will be way off.

I think Rackable will pull through, but I see plenty of downside danger. To a fair extent they should be able to meet anything Sun, Dell, or HP can throw at them because Rackable has a superior (and patented) technology. They have the advantage of being a niche player: the other guys might be willing to underbid for a while, but they have other distractions and no one wants to be the head of the division that is bleeding money within a company that has many divisions.

The need for the type of server installations that RACK sells is growing rapidly. I think management was right to do what it took to keep their market share. They have enough cash that they can't be forced out of the game easily.

Rackable is not, today, a stock for short-term investors. It is a risky stock for long term investors. But today it has cash equal to 42% of its market share; the immediate downside is limited. I admire the guys at IBM, Sun, Dell, and HP, but I think Rackable may be the nimble running back that dodges the heavy defenders and makes a good run.
See my summary of the Q4 2006 Rackable analyst conference