Friday, August 28, 2009

Marvell Technology (MRVL) Under the Hood

About half of Marvell Technology Group (MRVL) revenue comes from its disk drive segment. It does not make hard drives; it makes the chips that enable the drives read and write data and communicate with the rest of the computer. This was Marvell's first business and it now dominates that market.

Like most technology companies, Marvell had a rocky last half of 2008 and first half of 2009. Even after a 23% sequential revenue increase in its latest quarter (which ended August 2, 2009; second quarter fiscal 2009) to $640.6 million, its revenue was down 24% from the year-earlier quarter. On the other hand, Q2 2008 had been a blockbuster quarter, with Marvell proving that its investments in new technologies could pay off.

While Marvell may yet expand its market share in the hard disk market, and that market may expand as nations like India and China continue to expand their technology sectors, to understand (or at least guess at) Marvell's future revenue, profits, and stock value you must look at the other 50% of the company. This revenue already was approximately $320 million in the latest quarter, yet many of its products are just beginning to produce revenue or are still in development.

The networking end market accounted for about 20% of revenue in the most recent quarter. This segment includes the chips that go into network switches and routers. This is a very big market, and highly competitive, but Marvell has plenty of room to grow its market share.

Mobile and wireless networking is also an important end market. Here the combined analog and digitial expertise of Marvell, which are now integrated onto single chips, comes into play. One new type of product using Marvell chips takes a 3G signal usually used for cell phones and broadcasts Wi-Fi for local wireless internet connectivity.

Marvell also makes chips to control printers. According to HP printer sales have not been so great this last year, so expect this segment to rebound when people's legacy printers start dying.

Most fascinating are Marvell's application and communication processors (based on Kirkwood and Discovery Innovation). These are designed for the type of high performance users now expect from mobile computing devices and use very low amounts of power. More important, Marvell is able to put these processors on a single chip with one or more of its other technologies. It already has chips that combine Bluetooth and Wi-Fi capabilities. Rather than just sell discrete chips, it works with customers to design whole systems-on-a-chip. Other companies do this, but Marvell is the leader in the field.

Some of this last year's profits have been made by cost-cutting, particularly in Research and Development. But the key to profitability has been the model of integrating functions on a chip for customers particular products. To paraphrase management: We will continue to have above-industry margins as long as we are able to build products that our competitors cannot match.

Marvell has video processing capabilities. It says it has many design wins, but does not announce anything until its customers make their own announcements. You can imagine what might be coming out in 2010: devices with a digital processor, video processor, and either wireless or wired communications capabilities all on a single chip. Which will lower costs for consumers, driving unit sales, while allowing both the OEM and Marvell to maintain high margins.

The usual risks and uncertainties apply in the highly competitive technology field.

See also my August 27, 2009 Marvell analyst conference summary.

And keep diversified!

Marvell site

Openicon Marvell main page

Wednesday, August 19, 2009

Dot Hill Beneath the Surface

Dot Hill makes data storage components, and they have had a rough time the last five years or so. If you just look at the bottom line, it has been a sad story. Dot Hill had (and has) a strong cash balance with no debt, and for years profitability has been just a little over two quarters away.

Yet a lot has changed. If you look under the hood, you can understand why I am not alone in thinking that Dot Hill is likely to start showing profits in Q4 2009 and then really take off in 2010. [Assuming the usual caveat: if the economy does not implode]

At the end of Q2 Dot Hill had $57. 1 million in cash (it has a $0.7 million note payable for a recent acquisition as its only debt). Revenues for the quarter were $54.3 million, barely up from Q1 and down 24% from $71.0 million in Q2 2008.

But expenses have been slashed as well. Part of this expense reduction is a response to the recession, but much of it has to do with Dot Hill's product development cycle. The heavy R&D expenses were in 2007 and 2008. Now much of the R&D expense is devoted to lowering the production costs of the current generation of data storage components.

A few years ago Dot Hill had one major customer for its products: Sun Microsytems. But Sun decided to buy a much larger data storage company, so the writing was on the wall: new products would be developed internally. Dot Hill was relegated to supplying parts for expanding or replacing its old components used by existing Sun customers. And, of course, now Sun has been eaten by Oracle (sound like a Greek myth, doesn't it?).

So Dot Hill used its cash to develop new products for new customers (equipment vendors, as opposed to end customers). Reports are the new products are very good and end customers like them. Dot Hill sells the components to an increasing number of OEMs. Sun represented only 5% of revenues in Q2. A year ago Sun repesented 28% of revenue, and two years ago that was 65%.

HP is now the single largest Dot Hill customers. The new products only became available from HP in March of 2008. In Q2 2009 they represented 51% of revenue. The next largest customer is NetApp, which represented 24% of revenue. All the remaining customers accounted for 20% of revenue.

With sales to Sun unable to drop much further, the ramping HP revenues should cause Hill revenues to ramp in coming quarters.

Dot Hill has hit an economic sweet spot with their new products, which greatly improve storage capacity, efficiency, and power savings at a very attractive price point for end customers.

To a large extent costs of goods sold will ramp along with revenues, as will some operating costs. But right at this moment Dot Hill is running a lean system, so much of the coming revenue increases should increase the bottom line. In Q2 GAAP net income was negative $4.1 million. Non-GAAP net income was negative $3.0 million. But cash flow from operations was positive $3.4 million.

Since cost of goods sold represented 85% of revenues (GAAP), it will take a considerable ramp to get to GAAP profitability. But management believes that as the product lifecyle lengthens cost of goods sold will decrease as a percentage of sales (which is typical for this type of equipment) and there will be savings as production runs scale. Also, Dot Hill (through HP) is starting to sell software to help with storage management, and that has much better gross margins.

As always, things could go wrong with the economy, with clients, or with end-customer demand. But Dot Hill management (and employees) have been through the fire and come out looking pretty good. I believe they are well tempered and able to respond rapidly to changes in technologies and economic conditions.

As always, keep diversified!

More Data:

Dot Hill web site
My summary of the Q2 Dot Hill analyst conference

Saturday, August 15, 2009

Dot Hill, Applied Materials, Dendreon, Microchip and NVIDIA

Soon I should be writing Dissecting the Bull blogs on a regular basis again. The first half of August was an exciting time for the economy and investors, but my head was buried in a project for Microsoft, which I turned in Friday.

I did post several analyst conference summaries. Probably the most interesting is the summary for Dot Hill, which posted huge percentage gains in the week after the conference.
In this post I am just providing links to my most recent conference summaries:

Applied Materials (AMAT) August 11, 2009 analyst conference summary
Dendreon (DNDN) August 11, 2009 analyst conference summary
Dot Hill (HILL) August 10, 2009 analyst conference summary
Microchip (MCHP) August 6, 2009 analyst conference summary
NVIDIA (NVDA) August 6, 2009 analyst conference summary

Wednesday, August 5, 2009

Analyst Conference Summaries for Onyx, Hansen, Akamai, etc.

I have been very busy with projects lately, hence few posts. Much as I would like to take the time to comment on them, all I can do right now is let you know that I have posted summaries of analyst conferences recently for the following companies:

Akamai (AKAM) July 29, 2009
AMD (AMD) July 21, 2009
Biogen Idec (BIIB) July 16, 2009
Celgene (CELG) July 23, 2009
Gilead Sciences (GILD) July 21, 2009
Hansen Medical (HNSN) August 4, 2009
Intuitive Surgical (ISRG) July 22, 2009
Onyx Pharmaceuticals (ONXX) August 4, 2009
TTM Technologies (TTMI) July 29, 2009

Later this week: Cisco, Maxim Integrated Products, NVIDIA, SGI