Rackable Systems: Visibility Improving or Storm Clouds on the Horizon?

| About: Rackable Systems (RACK)

Anyone who has invested Rackable Systems (RACK), since the IPO in February 2005, has had a wild ride. The company came public at $12.50 and became a momentum darling. And why not? Initial customers for their AMD x86 servers included Google, Microsoft, Amazon and Yahoo.

Momentum buyers, supported by contributors at TheStreet.com, (including Cramer himself) drove the stock to an April 2006 high of over $57, despite two secondary offerings. First quarter earnings were “quite the buzz” at TheStreet.com and with all the analysts. “The Company under promises and over delivers” was the tone of every publication. The year-over-year comparisons triggered every earnings based advisory service on the Internet.

Analyst coverage at this point was pretty much limited to the bankers, who had now underwritten three offerings for RACK in less than 14 months. Insiders and management were allowed to be heavy sellers of “registered stock” in all three offerings, receiving more than half of total proceeds. Millions of additional shares that were registered and “locked-up” were subsequently sold in the open market. All these shares thus avoided the restrictions of SEC Rule 144, making insider sales at RACK especially egregious.

April and May were not good months for RACK shareholders. The stock started a steady decline, fueled by a rotten tech market, and closed on June 1, 2006 at $37.50. More insider selling pressured the stock down as low as $32 as over 2 million “locked-up” shares became free to trade.

What the market was beginning to realize was that while first quarter sales were up 176% over Q1/05, they were up only 1.5% quarter over quarter. GAAP earnings declined from $12.2MM to $9.3MM reflecting declining operating margins of 11% vs.14.7%, in a single quarter.

It was becoming apparent, to all but the emperor's court, that RACK was no longer "ahead of the pack" in offering AMD servers. R&D, which was basically nil in 2005, was now a growing operating expense. Their new product offerings in “scaled-out” servers and storage products were simply not selling. When “in line” Q2 earnings were released, combined with cautious guidance that indicated slowing sales, declining margins and some large delayed orders, the stock dropped to $22. Most of the decline came via a $13 after hours drop.

As the stock continued its decline into early August nearly everyone at TheStreet.com said “don’t buy, don’t buy” (Faulkner was the lone dissenter). For the first time, RACK had under promised, but had not out-performed. Sales of $86MM were up 95% year over year, but up only marginally from Q1. GAAP earnings continued to decline to $5.3MM.

However, RACK actually made some pretty good progress in the quarter. Scaled-out servers were beginning to sell and represented 17% of sales. Their DC power option accounted for over 50% of sales and they landed one large order for a new low-end storage product. The August low was $18.

The stock quickly began to recover, fueled by much better tech market, new analyst coverage from Hambrecht and word of new customers. New companies believed to be placing orders in the quarter included Raytheon (NYSE:RTN), AOL (NYSE:TWX), MySpace (NASDAQ:NWS) and YouTube. It was also believed that Microsoft (NASDAQ:MSFT) would begin buying servers for its new Quincy server farm. The stock closed at $28.80 on Monday, ahead of earnings.

Q3 numbers were not pretty. RACK reported sales of $80MM and GAAP loss of $389,000. Operating margins dropped to an abysmal 8%. RACK had now under promised and under performed.

But the future looks bright. The company gave some strong guidance for Q4 of $100MM-$110MM in sales and non-GAAP earnings of $.25-$.27. For 2007, management is projecting sales of $475-$525 and (non-GAAP) earnings of $1.28-$1.40. The market chose to ignore current problems and the stock traded up to over $33 before settling Friday at $32.66.

So the future looks bright, visibility has cleared and RACK's current problems were understood to be just growing pains. But earnings for Q4 and next year are based on increasing operating margins, something that is clearly going in the wrong direction. Margin increases will be largely affected by DDR2 memory and hard drive pricing, something hardly in the Company’s control. Most of RACK's production is from third party vendors, and thus not controlled internally.

The Company also claims to have had some “one time” items. But that does not explain margins dropping over the past four quarters from 14.7% to 8%.

Sales at RACK seem to be getting harder to predict as well. The Company indicated that 71% of Q3 sales came in the last month of the quarter. That is, in late August the quarter was shaping up to be a real stinker. Sales of their “scaled out” series which seemed to be gaining traction last quarter declined from over $14MM to only $4MM. Despite all the talk of server farms and competitive build-outs, sales to the “top three” (AMZN, YHOO and MSFT) declined from $57MM to $46MM.

Nearly 80% of RACK's sales are still in their low margin “foundation series” x86 servers. This market is becoming fiercely competitive with offerings from the likes of HP, IBM, SUN, DELL and a host of foreign manufacturers. RACK will now have to spend heavily on R&D to keep ahead of the pack and develop higher margin products.

Certainly the market for RACK's products appears to be robust. Q4 should be strong as corporations spend the last of their IT budgets. Analysts are raising their targets and the stock is now up over 80% from its August lows. RACK's products clearly have some competitive advantages and their large customers are committed to the RACK platform. They are growing their customer base amongst the largest Internet companies, while attempting to add large defense contractors and financial institutions.

But therein lies RACK's problem: a single customer can still make or break any quarter. The analysts’ targets are based on earnings and earnings depend on significant margin improvement. Year over year comparisons will begin to get very tough next quarter, as well. Sales in Q4 2005 were $83MM and non-GAAP earnings were $.39. Sales at the low end of guidance for Q4 2006 will be up only 20% and earnings will decline by 36%.

Hopefully, the future will still look bright. They can always use the “low bar” of Q3 as the new metric for business improvement. Meanwhile the insider sales, by management, continue as they sell their shares as quickly as they are vested. The only thing RACK has yet to do is over promise and under perform. This analyst will stand aside.

Disclosure: The author has been short and long RACK over the past few months but currently has no position.

RACK 1-yr chart: