Nvidia's (NVDA) future is wholly tied to its mobile system-on-chip (SoC) line of processors, Tegra. With the desktop and notebook PC under pressure from the mobile and tablet market, Nvidia's traditional cash cow in discrete graphics processors has to be discounted in the same way as Nokia's (NOK) Symbian smart phone OS; having a limited lifespan. It will generate the needed cash flow to fuel the R&D and conversion of the company into a full-blown SoC producer because in 5 years no one will be buying or retailers wanting to stock big, expensive discrete graphics cards.
Think about how much real estate at Best Buy is devoted to graphics cards now as opposed to just three years ago. Now project that into the future. Best Buy won't even be the Best Buy we all remember in the next 5 years, as they are downsizing along with the devices they sell.
Top line revenue has been falling over the past four quarters as has bottom line profit. A lot of that drop in earnings has been from a large shift in R&D Spending. Nvidia is a company that traditionally spends a lot of money on R&D, but the recent trend is indicative of understanding what it will take to compete with larger competitors.
While recent design wins for the Tegra 3 line of SoCs is both necessary and encouraging company guidance for Q2 of revenues sub $1 billion still represent a 4% decrease year over year.
Moving Onto Phones
Nvidia's Tegra 3 faces very stiff competition from Qualcomm's (QCOM) Snapdragon S4 line. While the Tegra 3's are quad core A9 ARM variants with the ability to dynamically under-clock themselves depending on the task at hand, the dual core S4s are built on Qualcomm's Krait cores, which are architecturally-similar to A15 ARM cores; having the same ability as well as an integrated LTE modem while being built on 28nm die size. The dual chip solution that Nvidia's current Tegra 3 lineup is built on a 40 nm process which increases both die size and cost while providing performance that is better only under certain situations. That said, however, Qualcomm is having problems meeting demand due to 28 nm fabrication issues at TMSC. This has, for this product cycle, given Nvidia an opening to grab a number of phone design wins.
But none of them will be for Windows Phone, if that makes much difference. The latest reports are that Microsoft (MSFT) is shutting Nvidia out of Windows Phone 8 while shutting HTC out of the tablet market. Microsoft is acting a bit like Apple in this respect, dictating to hardware vendors' strengths to ensure that all versions of Windows during this transition to Metro and touch-centric computing deliver a good user experience. This makes sense. Windows Phone's big advantage is its ability to run well on lower-end hardware versus Android. By all accounts the Nokia Lumia 8/900 provides an equal level of user experience to that of the best Android phones, using a single core Snapdragon SoC. This edge will need to be maintained if Windows Phone is to aggressively build market share though price advantage. Qualcomm's S4s provide that advantage right now.
Most benchmark tests are a poor substitute for real world performance and even the best of them have the S4 and the Tegra 3 running similarly depending on what one uses phones for. Nvidia's strong relationship with game designers gives them a big advantage in the mobile gaming space. This will, though, more likely have a bigger effect in the tablet and hybrid arena than in phones. But in general purpose usage, the superior graphics performance of the Tegra 3 is less interesting and an unnecessary waste of power, real estate and, ultimately cost.
Given that Google's (GOOG) Android is notoriously poor at multi-threading, the quad-core Tegra 3 in its target applications only outshines the Snapdragons in graphics performance designed with the chip in mind. Hence this may be why Asus split its top end tablet into two SKUs with the wi-fi version running on a Tegra 3 and the other employing a dual-core Snapdragon.
In the short term the Tegra 3 is carving out place for itself in the smart phone market. But, the recent purchase of Icera will not produce an LTE version of the Tegra until 2013 at the earliest. A leaked Nvidia roadmap has it being added to a Tegra 3 in the beginning of 2013, about the same time that Qualcomm will begin shipping quad core S4s in numbers; according to Qualcomm's roadmap recently announced at Computex.
The Tablet Wars
It is in the tablet space that Tegra has a number of design wins that can make it a very promising technology for Nvidia. The rumored Google branded 7" tablet looks to be running a slowed down Tegra 3 on its Kai reference platform and will be priced to compete directly with the Kindle Fire and Samsung Note.
However, the 800 pound gorilla in the room is the upcoming Windows 8 and its ARM-variant, Windows RT. At this point there is only one Tegra-containing device announced in that space but it will likely hit the market before everyone else. Asus showed one off at Computex last week. Acer is holding off until 2013 saying that WinRT will not be ready until then. Microsoft has wedded itself to Nokia for Windows Phone who, in turn, is wedded to Qualcomm. Nokia's WinRT tablets will be Snapdragon-based. It's looking like a Microsoft-Nokia-Qualcomm axis is forming in the WindowsRT space.
For Nvidia, however, if that road map is accurate it means it is at least 9 months behind Qualcomm, if not more so. The quad core S4 was being shown off in reference hardware running smart TV's at Computex last week. If Qualcomm executes its tiered quad core S4 roll out over the next 6 months, Nvidia will be in serious trouble. Quad core S4 laptops will be challenging up the food chain, looking to take market share in the mobile space from Intel, leap-frogging Nvidia's current offerings entirely.
The Game's the Thing
Nvidia's niche will have to be its ability to provide superior gaming performance in the mobile market because it is not coming from the console gaming market. The rumored specs on the Xbox 720 are a design win for AMD(AMD) with its APUs. The new Wii-U is using a Radeon based GPU. The latest release from Sony confirms another design win for AMD using both CPU and GPU in the Playstation 4.
Nvidia's relationship with game developers is a major asset and one that Microsoft and Google will support in every way. While the headlines keep talking about Nokia as a potential buyout candidate, it would not be surprising for Nvidia to be the one bought up first. At a market cap of $7.5 billion with contracting revenues and a $3.5 billion cash and short-term equivalent position, the economics for an Nvidia M&A deal make far more sense than the more expensive and baggage-heavy Nokia. At this point, however, there is nothing that suggests Nvidia can make the transition without getting squeezed out by forces far larger than themselves. $12 per share with falling earnings and revenue implies at best a value trap for the next year and at worst a 50% haircut. Without a dividend to cushion the wait what is the compelling reason to buy at these prices?
Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.