October 6, 2003

IT Investor's Journal

- by Melanie Hollands -

I'll start with my bigger picture views about IT demand and then move into my thoughts about some recent events affecting tech stocks the last couple of weeks: Sun Micro's preannounce last Tuesday, the ongoing Oracle/PeopleSoft saga, AMD's Athlon launch, rumors around NetApp and McDATA, and InfoSpace - a cash-rich wireless company in need of a visible growth strategy.

IT Demand

On balance, I'm not observing anything that suggests a widespread IT upgrade cycle, a sustainable improvement in private sector IT spending, or activity at distributors is imminent. However, there has been an "echo PC boom" that is important to many companies. Many people don't see the "echo boom" happening because unit pricing is weak, but PC processor unit volumes are now higher than pre-Y2K (look at Intel guidance this quarter). This mini-boom could last for a few quarters.

I know of a company in Texas that is a Cisco-only reseller with ancillary, or add-on, product bought via distribution to fill out orders. This reseller seems to be doing well, having won some large bids for networking gear, but that's the only distribution/reselling business of which I'm aware that is really robust. Otherwise, it just seems rather quiet. I don't sense that business is that robust nationwide, or even in wide geographies from other resellers and distributors. A colleague of mine has been working with some business development managers on deals that have surfaced that are Asian-owned resellers, mostly internet/catalog based. Their businesses are strong and growing in the $200 million - $400 million revenue range. Some have doubled in revenue on low margins and appear to be taking share away from the larger players. In addition, I had one owner of a sizable distributor tell me it's not unusual for deals to get flipped on 2-4 points. So, it seems that there could be a shakeout in the distribution business coming down the pike.

To make this "uptick-in-IT-demand" debate a bit more interesting, I have colleagues involved in a restructuring with a quick-turn PCB player. The company has indicated a big uptick over the last two months, across most sectors. Lead times are growing, pricing is actually rising, and components manufacturers are having trouble supplying them. The company has also brought back some recently laid off people as they are bumping up against their "people capacity." This tells me somebody is doing something, buying something, or else inventory is building in a big way. I happen to believe it's the latter. A lot of what this PCB company says seems to be influenced by what they want out of the restructuring process, but TTMI and Merix (both of which I have been long for most of this year) seem to be saying the same thing, so this might actually be on the level.

I'm not trying to be evasive, but it's possible that both might be right, and the IT demand turnaround could be segmented by products and/or applications. If that turns out to be the case, then we'll get many mixed signals regarding early indications of a sustainable, real uptick in IT spending demand. It's also possible that people are building inventory expecting tax cuts and a last flurry of mortgage refinancings to boost short-term demand. It's so hard to quantify inventory build as the channel is so large with so many opportunities for inventory to build up! I have a hard time believing that demand is strong now outside of an inventory build and deferred orders due to Iraq and SARS now coming through. Based on my channel checks with distributors, resellers, IT Project finance companies, credit contacts, etc., perhaps demand stays in the maintenance, high unit volume/low price (or slow growth) mode.

SUNW - In a Rut or Digging Itself a Grave?
I don't like Sun Microsystems' business model. Last Tuesday, Sun pre-released negatively based on loss of market share (no surprises there) and announced a substantial $1 billion charge they will take. That said, I think it's worth looking at SUNW for a longer-term "buy and forget about it" position. (Like buying NT and LU at
I'm repeating below what I wrote about SUNW on July 21 this year. Clearly Sun needs to move if they want to be something other than a niche player. Perhaps they get bought, but I doubt that a merger would happen with the current Sun management team. One hypothetical option is that IBM and Sun could tie the knot. Then Linux begins to replace Solaris while Sun replaces IBM's server unit. Sun has some good niche products, although they are losing the battle in servers. In addition, Sun's "not built here" mentality has resulted in bloated development costs. With a hypothetical Sun/IBM merger most of the server and storage product lines overlap, but they could go after the market with multiple brands.

From July 21, 2003: "It's [currently not a good use of] capital owning SUNW, in my opinion. Unless you want to roll the dice on one of two things: 1) a take-out play, or 2) a turnaround play. From time to time, it could also be a nice "occasional trading" stock. But for the next few months, I'd be short the name since I see no reason for upside, and plenty of room for more downside from current levels.

On last night's conference call, the company provided no inkling whatsoever on revenues EPS or GM and vague references to higher or lower R&D and SG&A, but not specific percentages. No numbers... no nothing. Every answer to such questions was "We are not going to comment", "We are not giving guidance", "We are not going to comment". A giant yawner. At least they used to try to communicate the good and the not so good of their business, give some color. But all they did yesterday was stonewall. And talk about their cash position; which is very strong but you'd think they were a bank or something. If it weren't for their balance sheet there's no company.

The fact that Sun has been at a crossroads and attempting to decide on what to do is a big internal issue and Sun's lack of focus is not encouraging. One main thing that I have been hearing is that products are really slipping on the ship date. Sales just eaked out a profit in their previous quarter, but I think it's severely impacted by cutting costs. Although two positives are that G&A has been seriously been reduced and that Sun has plenty of cash.

There is a consumer portal now with Christina Aguilera (!!!) as the sponsor. When Zander left the company there was a huge shift from hardware people and their agendas and the new leadership is mostly software. They're not worried about Silicon Graphics; they are worried about HP, IBM and Dell because they are moving towards the IBM approach - sell service and make low margins on the hardware. But since they are also deeply rooted now in software, they have a lot of competition. Their products are slipping away from being leaders to laggards. The frequent internal reorganizations show a lack of vision and only a short-term focus. One wonders if the company should split its organization structure and have a core R&D lab, then product groups, both software and hardware, a product support group, then a sales group. Things keep condensing over there at Sun.

The company has a cultural problem. It's very US-centric, and I think they don't really understand Asia although Europe is OK. Localization is a problem since a lot of marketing activity comes from the US groups. There is a lot of back and forth since they have product marketing and field marketing

Java seems to keep figuring out how to get more and more money from licensing. Sun's OS group is not a profit center. Blade servers (acquired from Cobalt) is slow but getting up to speed from what I hear. Other software groups seem to be doing OK, including iPlanet (a competitor to BEAs Web Logic) and Forte (a competitor to Borland). Another part of the problem at Sun is they have always been reliant on their rather-proprietary hardware and moving away from that is a good step. They have rather complex partnerships on the hardware side - IBM, Fujitsu - and also with EDS and other consulting firms. These partnerships are OK (albeit often messy) since Sun cannot be everywhere and are not specialist in every market segment and geography.

As for what Sun can realistically achieve going forward, I'd say survival on a "subsistence" scale is the best-case scenario. Focusing away from buying boxes and buying solutions and software and solving the management headaches should be their goal. There is such an "over-glut" of hardware out there. One good thing is that their release schedule for software is improved. A big danger is that Sun keeps toying with a consumer play since I don't think they have the right people who understand that space. For that they should heavily rely on consumer companies like Sony, Coca-Cola, Nokia, Mercedes, GM, GE and several others to promote their agenda. But I doubt that Sun will ever successfully enter the consumer space.


Oracle's quarter was weak but not a total bust. It continues to do poorly in the apps business - possibly because its natural constituency is less likely to buy right now than SAP's constituency, or more likely because we're still in the recovery phase of the apps business and there's going to be a lot more variability in results between vendors in any quarter, and variability of quarters for any given vendor (e.g. Documentum's slip in the second quarter). I like ORCL here, but would prefer to grab it at $11. They tend to miss the first quarter but the company releases the new apps product in November and has a good upgrade cycle going into next year. I'm thinking long ORCL for around 6-8 weeks, but wouldn't be in a rush to start a position. I think we get broader market weakness in October so there could be a lower entry point.

The PeopleSoft deal is too close to call, but I'm inclined to believe (as I have since the hostile tender was originally announced) that Oracle's bid will be unsuccessful. My view largely hinges on my belief that the deal won't pass muster with the Government. If enough states' Attorneys General file suit (possibly egged on by their IT people that don't want to have to deal with Oracle) Larry might decide that he has an out and decide to drop the bid. Or not. But Oracle didn't play their cards right. Oracle's attempt to talk down PeopleSoft's product line by saying it was going to discontinue the products was a misstep. It made a bunch of people angry and made a hostile takeover that much more hostile. Maintenance of Oracle's customer base, that's what it was about. Oracle was going to take the PeopleSoft products it liked and leave the rest while keeping the maintenance revenue stream. Even if that was Oracle's intention the company never should have made it public. Big mistake.

That said, from a shareholder value perspective it could be argued that the deal should get done. I don't think PSFT is worth the current bid amount of $19.5. Maybe $14-$15; $16 tops, but not $19.5. Perhaps PSFT shareholders they will see that the greatest value to themselves lies with approving the Oracle acquisition. I think a major shareholder who doesn't go for the Oracle deal, from a purely financial perspective they should have their head examined. I'm just saying that is one scenario whereby this deal could get done since, financially speaking, it really is in the shareholders' interest. Take the cash now and buy ORCL or SAP (which continues to be the dominant player in apps). Holding PSFT is not the best choice available for a shareholder.

According to Oracle (and who can necessarily believe that) they have a number of large shareholders backing them. We shall see later in October, unless the shareholder vote date gets pushed out yet again. Time was the enemy when they were thinking of preventing the JDEC deal, and it's still the enemy if they're seriously considering spinning the thing off or selling it. (Who wants it after the kimono's been opened to PeopleSoft and PeopleSoft possibly retains most of the development talent?). The arbs have still been trading PSFT the last couple of months, and there continued to be buyers of size below the 16.5 level, and again around the $18 level. But most of the arbs I know don't think that Oracle successfully buying PeopleSoft is a real possibility either. Too many hurdles.


AMD stock has roughly doubled over the last three months, propelled by bullish comments about the company's 64-bit processors, Athlon and Opteron, in the media, some trade magazines, and by financial analysts. There was also considerable chatter about new OEM wins for AMD's new parts and even potential for an AMD/IBM alliance. All of which has been positive for AMD stock the last few weeks. I had liked AMD stock for a short off of the back of the Athlon launch (on August 23), and the stock has pulled back since then. But I would not be short the stock when AMD reports on October 16, and might even be long the name. Even though Intel has taken back some share and, overall, AMD is in transition mode in terms of product cycles, I think the "echo PC boom" and better health of PC unit sales should bode well for AMD this quarter. I don't think the risk/reward is attractive to short AMD over the next 3-2 weeks; not ahead of the conference calls for Intel (October 14) and AMD (October 16) anyway. I might think about shorting AMD if it spikes again after they report on the 16th.

Some analysts believe that AMD's new MPUs might present a longer-term threat to Intel's. However, Intel has made some vast enhancements and over the past year has, in fact, fixed most of the bugs on Itanium2. But the more important area to focus on is notebook's, not PCs, and I still feel that Intel's Centrino has beaten AMD's Athlon XP-M to the punch. While I consider AMD's new MPUs to be better than what the company has produced in the past, I think these new parts are still not good enough to challenge Intel's dominance in the microprocessor business. I've not been hearing consistently bullish views from inside the industry. In fact, I've not really heard much positive feedback from a number of sources on Opteron. In the past, AMD has never been able to make any great strides in penetrating the server processor market despite the fact that many OEMs would welcome a second source to Intel.

While Opteron may indeed be technically better than AMD's past efforts, and may allow AMD to nibble away at Intel's share of the processor market, I doubt the new parts present any serious longer-term challenge to Intel's dominance. To do that, they would need Dell's business. For my part, I believe it's unlikely that Dell will use AMD's parts. I give that a 10% chance of happening, no time period assigned. Dell's tactics are to partner with/source from market share leaders, not niche players. But I could be dead wrong on Dell; and if I was wrong then I'd feel more positive about AMD. AMD has secured some solid OEM agreements with Sun and others, but would need volume deals with HP and Dell, coupled with a much stronger marketing focus, to present a challenge to Intel.

Although I regard the excitement about AMD's new parts seems as overdone, AMD can carve out a solid niche in microprocessors without attacking Intel head-on. There's an increasing number of people out there who don't see the need for the latest and greatest microprocessor anymore. Intel's strategy is still firmly planted in the inevitable and rapid move to new architectures and faster speeds. What that means is that some OEMs and customers are more and more inclined to listen to somebody (like AMD) who can give them a more efficient chip, at a better price, with less top-end competitiveness. Especially if AMD's new chip is capable of running new 64-bit apps in the future while remaining backwards compatible with what customers have right now - which is usually 32-bit. The one thing Intel does not offer is an "upgrade path" to 64-bit. Customers can stay 32-bit, they can convert, or they can run two different environments simultaneously. AMD's new parts seem to be more appropriate to a gradual upgrade strategy. That said, while this can give AMD a solid niche market I don't think it's sufficient to threaten Intel.

There has been chatter about a potential AMD/IBM alliance over the last couple of months. While a manufacturing deal is possible, a joint development or marketing alliance seems less likely. From IBM's perspective, I see little reason for an alliance with AMD. Why would IBM want to gain further exposure to the PC market? Slow growth, limited potential for market share gains, etc. (I think IBM made the right decision back in the 90's to move away from the desktop PC market and therefore would be surprised to see them go back at it again.) The only reason I could see is to fill their 300mm fab capacity. But I'm not aware that IBM is focused on morphing Fishkill to the foundry model, although they do have some customers (XLNX, NVDA), and recently forged a deal with IFX. Perhaps AMD could put up some financing and partner with IBM. AMD is in serious need of building microprocessors on 300 mm and IBM might be willing to help them achieve economies of scale for the right price and with minimal risk. If there were some kind of alliance with IBM, this would likely be perceived as positive for AMD's stock but probably neutral for IBM's.

AMD stock could potentially be a great contrarian play if you're willing to hold it for 6-12 months, and if the company gets more traction with OEMs. But I'm not a huge long-term fan of AMD stock; not as things stand for the company right now. At current levels, AMD doesn't look like the value it was three months ago.

NTAP - Rumors are Tumors

There was an NTAP rumor ("rumors are tumors") doing the rounds on trading desks earlier this week. It was that NetApp are going to have to guide down and that MSFT's new storage server products are knocking the pants off NetApp this month and killing the quarter, much less longer-term threat. Well, I don't buy it, particularly a full month prior to quarter end, so it seems a bit early to be talking about guiding down. In my opinion, the quarter would have to be a total bust, and completely "non-savable", for management to do this. And if this rumor was true, then given NetApp management's bravado they would more than likely wait until the last minute to guide down. If people just stopped to think about this stuff a bit, they'd see how silly this rumor is, at least at this point in time. Do they really believe Microsoft is taking that big a bite out of NetApp's sales? I don't think so.

For my part, I'm hearing that business is going quite well at NetApp, on strength in demand for lower-end storage products. That said, NetApp's sales are usually tough to figure due to the large direct sales percentage. Around mid-summer, NetApp gave all "non-National" accounts notice that they had to buy from one of two distributors - Arrow or Avnet. NetApp only sell direct to their largest accounts. Now, that is not unusual, but I think is a sign of either 1) slowing sales and the need to get distributors to push the product, or 2) growing pains, whereby NetApp management realizes they can't handle all those numerous accounts. Either way, there's a transition period with each reseller setting up a line with the distributor, if they didn't already buy from them. The transition to greater use of distributors should be factored into current analysis of NTAP since it could be a one-time event that helps spur sales in future quarters.

NTAP stock has been down a lot over the last couple of weeks, in my view due to anticipation of to EMC's new products. I have a feeling that this rumor about an imminent guide-down is nonsense at this point.

MCDTA - Inclined to Stay on the Sidelines

Whether you choose to go long MCDTA or not depends what your expectations are for the company. Will McDATA get purchased? Will they innovate or continue to compete in a shrinking margin product line? There has been some insider selling recently, which may (or may not) mean something. I realize that 460,000 shares is only 5% of the 10 million McDonnell and his wife own, but it's still not comforting. Plus, another officer decided to sell 50,000. Makes me wonder if the company pre-releases negative. They're famous for that.

MCDTA is not my favorite storage name and I'd be inclined to stay on the sidelines, for the most part just playing the volatility. It's a tough call making a directional bet on MCDTA. Worst-case scenario, I can see MCDTA at $6; best case scenario $14-15. The stock nearly reached $16 in early July, but only after management made overly bullish comments, which they later retracted. If MCDTA gets to a more attractive valuation (say, high single digits) then it could be a good stock to own for the last quarter of the year. I think storage is one of the better-positioned areas in tech, but I'd prefer to be long EMC as that company seems to be getting its act together.

I don't particularly like the fundamentals of the switch business. Hardware companies seem destined to become more like cyclicals. In particular, I don't see much growth in the segments in which McDATA is currently active - organic growth is just not there. I expect McDATA will continue to look to grow through acquisitions. It recently acquired Nishan, and while the technology Nishan sell sounds good it seems that few companies are buying it. McDATA has made other acquisitions, but mostly companies that have no proven markets for their products. Furthermore, McDATA are doing another couple of investments in Sanera System and Aarohi Communications, and these two will hurt earnings.

I doubt that there's a need to connect backend storage devices to an IP network, which is what front-end applications already do. McDATA has a director class switch that is currently one of the few things on the market that has been selling well. However, while I think McDATA has some okay products, as IT shops try to create simpler, cheaper SANs I think it might lose out in the space to Cisco and other commodity-oriented switches. Offerings from "standard" suppliers should be very appealing in situations where premium performance is not necessarily required. Consequently, Cisco and other "standard" networkers should be the big winners from this trend.

High-end storage companies can get squeezed more and more by more robust technologies based on simple standards like those from Cisco and Microsoft. At an analyst dinner in July, I needed a calculator to count the number of times the McDATA CEO mentioned the Cisco threat. Chambers had personally negotiated several marquis wins for switches. Now, most were greenfield, but no matter. It's cutting into McDATAs (and Brocade's) revenues and would have been in the IDC numbers for these two companies, and now it won't be. McDATA's CEO also said that if Cisco had McDATA's Sphereon 4500 in their product lineup, then it would be game/set/match for Cisco. So why doesn't Cisco just buy them and be done with it? It's either that or they develop their own product and put McDATA out of business too, but that would take longer to get to market.

I don't like to own stocks based on take-out speculation, although there are a couple of hypothetical options. One suitor could be Cisco, whose expertise is in the software that enables their IP switches, not the hardware. Their type of software doesn't work well with storage applications, which have other requirements. Cisco bought a switch company, Andiamo, last year, so I doubt that Cisco will expand further into storage switches if they don't see results from the Andiamo purchase. Maybe Dell could buy McDATA since it already has a good relationship with EMC. I believe that Dell will eventually get into the storage switch/director market and McDATA could be a good fit.

Unless you see a new market opportunity for McDATA, or see it being a buyout target, it could fade away into history like Wang and SGI. Management has often been wrong. Recently, they got it wrong about the Cisco threat. Very wrong. I don't think management is deliberately misleading, just that they are often "not right". And that makes me concerned about what else they can be "not right" about. Business concentration risk is another concern; too much of McDATA's business comes from two companies - EMC and IBM. It needs to diversify its customer base or it could be hurt going forward. An ongoing price war between Cisco and Brocade isn't helping either.

Fundamentals aside, from a purely technical perspective MCDTA looks OK as a long with an upside target of around $14+. I think it's not a bad risk/reward for MCDTA. A break above $12.75 would be good, there's support at the 50 day MA at around $11, which appears solid. So, an entry point at around these levels looks good with a tight stop at its 50 day MA around $11. I might start here with a tracking position and add on an anticipated move to $13. The fact that the gap fill is being attacked with higher than average volume looks promising for the stock. If MCDTA had moved to fill that gap on lighter volume, then I'd have said it looks more like a short. But it's moving there on good volume, and that could support a technical up move. Then, I'd look for a move to $14.5-ish (where a "pair of peaks" would likely form).

INSP - Long-Term Long

INSP should have more upside for patient investors. I don't see INSP trading above $25 in the next 2-3 months; and with any broader market softness the stock could pull back into the mid- to high-teens before moving into the mid-$20's. Basically, InfoSpace has cash but not much else going for it at this point. But the market loves a Cinderella turnaround story, and INSP probably has more upside over the next 12-18 months. I'd wait for a pull back before starting/adding to a position and be patient.

I liked INSP long in January and it was an absolute no-brainer then. InfoSpace was trading at around $7, well below its $10 per share in cash at that time, and in 4Q 2002 reported that it was treading water from a cash flow perspective. So, INSP was like investing almost in cash, with so much cash on the books and a break-even operating cash position. The stock had been punished as the company's ability to survive the economic downturn was in doubt. Hence, the no-brainer call. Since then, former CEO Naveen Jain (not well liked by the Street) has been removed and a new management team is in place. The company has preserved its strong cash position, although it still lacks a clear growth strategy and execution plan.

InfoSpace has some positives. First, it has $304 million of cash or cash-like investments ($302 million current + $2 million long term), and interest income on that cash. Second, zero debt, which, with about 32 million of "reasonably" diluted shares, implies cash of about $10 per share. Third, it recently became EBITDA positive for the first time. Ignoring pre-2001 results, as they aren't correct with restatements, poolings, etc. (InfoSpace is a total mess in that respect for comparability purposes.) there has been upward progress on EBITDA over the last six to eight quarters. Fourth, capex runs about $1-2 million per quarter and since this is a very low capex business, with hardware costs dropping each year, capex should never go up. Fifth, a new management team should help it repair its reputation with the Street. Finally, Internet and wireless is still at the bottom offering potential for upside.

The INSP stock price ramped over 70% during the summer for two reasons. First, a court order for Naveen Jain, InfoSpace's former CEO, to repay $200 million+ for allegedly fraudulent stock trades. Although it's unclear when, if ever, the company will get this money from Jain, who is fighting to not pay back money he owes. Second, the new management team has refocused InfoSpace around its original businesses - White Pages and Yellow Pages. That said, the company still has no clear growth strategy and has narrowed its product offerings. New product introductions have been weak, especially products with any market traction. There has been modest organic growth from InfoSpace's Authorize.net service and Search, and going forward there's potential to derive a healthy income from Authorize.net, which helps small and medium sized businesses to authorize credit card transactions. Beyond this, other offerings in the merchant and wireless space seem to have fizzled. Other sore points have been legal costs and recent fraud issues associated with Authorize.net, but those are subsiding.

While INSP may still have upside longer-term, I'd be prepared to be patient. I don't see INSP trading above $25 in the next 2-3 months; and with broader market weakness the stock should pull back perhaps to mid- to high-teens before resuming its up trend. InfoSpace's cash position is one strong point, so management has time to figure out the company's strategy. But that's not a compelling short-term investment rationale. Other positives include the $200+ million suit against former CEO Naveen Jain and the prior management team's issues are history, but the removal of these overhangs is already reflected in the current stock price. Longer-term, with an eventual industry turnaround or a modest uptick in end demand, INSP could move up into the high $20's/low $30's.

Melanie Hollands has over a decade of experience covering the technology and telecommunications sectors, from positions held in business strategy (McKinsey & Co., Bain & Co.), corporate finance (Salomon Smith Barney) and fundamental equity research (Merrill Lynch). While at Merrill, she received an Institutional Investor All-Star Honorable Mention for her coverage of PC hardware and wireless stocks. She follows PC/server/storage hardware, enterprise software and application software, wireless hardware and software, data networking and telecom equipment, optics, semiconductors, semi capital equipment, and digital media. Hollands is President of hedge fund Koala Capital, which focuses on trading/investing in technology stocks. She is also an advisory board member for a start-up technology venture, the Semiconductor Futures Exchange , and has been a guest lecturer at Columbia Business School, where she earned her MBA.

Editor's note: The opinions and conclusions in this article are solely those of its author and may or may not be shared by OSDN editors and management.

Click Here!