Marvell Technologies (MRVL) is a fabless semiconductor chip maker focused on SoCs (System of Chips) for data storage, networking, and wireless communications including smartphones. 2013 was a good year for Marvell investors. The stock ended 2012 at $7.08 and 2013 at $14.38, generating a 103% return for the year.
It could be argued that 2013 was an anomaly because the $7.08 starting price was artificially low, a product of fear about patent litigation and about the supposed decline of hard disk drives (HDDs). I have concluded that Marvell will likely see a good revenue and profit ramp in 2014, resulting in a price closer to $20 per share. As the basis for that argument I'll look back to Q3 2013 results (actually, because Marvell is on a fiscal year, results for the quarter that ended on November 4, 2013; Marvell's fiscal Q3 2014).
The whole Marvell mechanism summed up to GAAP earnings of $0.21 per share, up 75% sequentially from $0.12, and also up 75% from $0.12 year earlier. On a non-GAAP basis (mainly reached by excluding stock-based employee compensation) EPS was $0.32, up 39% sequentially from $0.23, and up 60% from $0.20 year-earlier.
First note that the reported quarter is typically the seasonal high for the year due to higher sales of chips for devices sold to consumers during the holidays. However, much of Marvell's product ends up going into the enterprise market, which tends to be less seasonal.
Marvell has the largest market share of any supplier of HDD controller chips. As a result the number of HDD units shipped by manufacturers like Western Digital (WDC) and Seagate (STX) is critical to Marvell revenues. In Q3, despite a tepid global HDD market, Marvell revenues were up 3%, indicating another market-share gain. Marvell also had record sales of controllers for SSDs (solid state drives). While Marvell does not yet command the market share in SSDs that it commands in HDDs, it is well-positioned to substitute SSD revenue for HDD revenue. Recall that in 2010 some futurists predicted HDDs would be obsolete by the end of 2012, resulting in dramatic reductions in stock valuations for MRVL, STX, WDC, and others in the HDD business. Those of us who bought the stocks rather than the hype have done well. Today HDD still has a large cost advantage over SSD, and we also know that SSDs have problems with corruption and degradation that must be addressed by controllers, just like HDDs.
The SSD and HDD controller business provided 46% of Marvell's Q3 revenue. I figure it will be flat for practical purposes in 2014, and seasonally down in this current quarter.
Networking, both wired and wireless, represented 17% of Marvell revenue in Q3. It is a complicated market, with strong competition, but while Marvell expects some growth in 2014, I'll call it flat until I see data proving otherwise.
The argument for Marvell's growth in 2014 centers on the mobile and wireless space. Two recent announcements confirm my long-held hypothesis that Marvell would be one of the winners in the global smartphone SoC market. Marvell took a beating when its main cellphone client, Blackberry, lost most of its market share after 2009. But MRVL kept up its heavy R&D efforts, despite the impact on the bottom line, and worked with Chinese regulators and service providers on the then-new Chinese 3G TD-SCDMA standard. Marvell gained a major share of the TD-SCDMA mobile phone market in China. The strong results in Q3, with the mobile and wireless segment up 60% sequentially, and representing 31% of Marvell's total revenues, largely reflected 3G sales in China. (The segment also includes other wireless connectivity solutions, notably for game consoles).
Marvell has been promising announcements in the standard now being introduced in China, TD-LTE. Back in November it announced a new TD-SCDMA slot for its PXA 1088 ARMADA mobile processor in the Samsung Galaxy Win smartphone. The first 4G, TD-LTE win announced was certification of the Yulong Coolpad 8736 smartphone containing the ARMADA PXA 1802 in early December.
The new year brought a flurry of announcements. A "mass market" Coolpad 8720L based on the PXA 1088 that will sell for 1,000 renminbi (about $170) puts a full featured LTE smartphone within the grasp of most Chinese consumers. It is important to note that service providers do not generally subsidize devices in China, so a $600 iPhone or other expensive phone is going to continue to have limited sales appeal.
On January 6 the ZTE (OTCPK:ZTCOF) Grand Memo LTE based on the AMRADA PXA 1802 was announced. In addition to the LTE standard, the 1802 modem supports older standards and another important new standard WCDMA.
Also announced was another Yulong Coolpad, the MagView 8970L, a high-end, 5.9" high-def smartphone, again using the 1802 chip. Yulong called Marvell's modem chip "a twin-turbo attached to the powerful engine to boost the 4G TD-LTE industry that is taking off in China."
These announced wins may be driving the rise in Marvell on January 6, but they also are milestones showing Marvell's overall strategy is working. Marvell lost out in the U.S., but China is a bigger market. Marvell's chips are now being certified by U.S. carriers. We may see Chinese-made, high-quality, lower cost smartphones in the U.S. this year, and also in many more nations outside China. The WCDMA standard will be important to sales in developing nations, enabling a rapid transition from older cellphones to smartphones.
I should not need to point out that the competition is sharp, particularly from Qualcomm (QCOM), and Intel (INTC) is certainly throwing billions of dollars towards buying a share of the smartphone processor market. There are several up-and-coming Chinese chip makers to look out for as well.
So Marvell can't stop running, it has to keep its R&D budget at high levels, and it is dependent on OEMs for final sell-through. Despite all that I expect Marvell to gain global market share this year. 3G chips will continue to sell in 2014 (China Mobile is just starting to turn on 4G networks), while the LTE chips will ramp.
The trick is putting a number on these trends, when no one really knows how sell-through will progress. Marvell's guidance for the current quarter (fiscal Q4 2014) which ends January 31 is for revenue to drop from Q3's $931 million to between $880 and $920 million. Even with the seasonal drop, revenue will be 13% to 19% above the $775.3 million year-earlier revenue.
EPS should grow faster than revenue, but taking a conservative 16% growth rate for EPS, by fiscal Q3 2015 (ending November 2, 2014) non-GAAP EPS would be $0.37 per share, GAAP EPS $0.24. That is likely the seasonal peak for the year, so for annual EPS I would use $1.32 non-GAAP, $0.80 GAAP as a range.
Using a rule-of-thumb P/E of 20 for a stock showing good profit growth, that would project out to a range of $16.00 to $26.40 per share. Of course by making different assumptions you can get different target prices. You might also want to look at the patent dispute. I think it amounts to nothing except a prejudiced local jury that did not understand patent law or technology, but you may think differently.
There is a race on to bring smartphones to developing nations, where Nokia once dominated the cellphone markets. If Marvell can win a substantial share of this market, there can be substantial upside even to the rosy scenario I have painted above. The downside risks should be obvious. China is a tricky environment to work in, and nothing guarantees Marvell will succeed in getting its chips into significant numbers of smartphones sold outside China. Competition is intense, and if it becomes pricing competition, margins could erode for all the smartphone chip operators.