It's been a while since I've written about NVIDIA (NVDA) here. Interestingly enough, this has been one of the best performing long-term holdings in my portfolio, with shares up about 30% from my purchases well over a year ago. After such a nice run, it always pays to take a step back to evaluate what the "next step" is.
Well, spoiler-alert: I'm still holding on to my shares. I think that this is at least a $20 stock, and I believe that over the next couple of years, the dividend yield on my cost basis will continue to climb as the company continues to raise its quarterly dividend. Of course, these predictions are fairly meaningless without fundamental context to go along with them, so let me give you an update of my view of the NVIDIA thesis going forward.
Smartphone Upside Is Real
On NVIDIA's most recent earnings call, CEO Jensen Huang confirmed that the company's Tegra 4i (integrated apps processor + LTE baseband) had been certified by AT&T (T). Now, while this wouldn't be a big deal for a big smartphone chip name like Qualcomm (QCOM), this is great news for NVIDIA since it will finally have a chip with which to attack the mainstream LTE smartphone market.
That's what I've loved about NVIDIA's smartphone strategy - the risk/reward is just so incredibly asymmetric. Today, NVIDIA doesn't have a meaningful presence in smartphones (yes, I'm aware of the Xiaomi Mi3, but one design does not a meaningful presence make), but with the right product, it's difficult to see the company failing to make real inroads. But therein lies the rub: NVIDIA needs the right product.
I think the Tegra 4i is that product, even if it isn't perfect. Could it have benefited from ARM's (ARMH) upcoming Cortex A12 cores rather than the older Cortex A9 (albeit clocked higher and with some good architectural improvements over the very first A9 cores)? Sure. But if you're looking at what Qualcomm, Broadcom (BRCM), and MediaTek are going to have in this space, NVIDIA is plenty competitive. While it won't likely be an overnight sensation, I do think NVIDIA gains real marketshare here and can build on it from here.
I like the Tegra 4i and I like that NVIDIA has committed to being a meaningful player here in building/acquiring both apps processor and modem capabilities in house. It's not going to be an easy fight and there will be a lot of consolidation in this market (with plenty of players essentially forced to "go home"), but NVIDIA has the fortitude to keep the fight going.
Tablet Momentum Could Return
NVIDIA's management made the strategic decision to delay its Tegra 4 (for tablets) in order to pull in the Tegra 4i and "Logan" (Tegra 5) processors. Thanks to Tegra 4 being late (it is a competitive processor), it lost a number of key, high volume sockets such as the Google (GOOG) Nexus 7. Tegra ended up down Y/Y in FY2014 as a result.
However, as we head into FY2015, things look much better. The successor to Tegra 4 ("Logan") is NOT delayed and management seemed incredibly upbeat about this particular chip's potential at the recent UBS Technology Conference. The initial demonstrations of this chip have been impressive, and there is a good chance that NVIDIA can wrangle back one or more of the sockets that it lost during FY2015.
Of course, we can't look at NVIDIA in a vacuum. Qualcomm is a fiercer competitor in tablets than it ever has been, and Intel (INTC) has competitive products and the wherewithal to brute-force its way into the market. NVIDIA is a much smaller company than either Qualcomm or Intel, and its leverage with many of the tablet OEMs probably isn't as high as those two giants'.
However, NVIDIA has been incredibly clever about how it has been going about things, and is now helping to sell - by way of its traditional army of GPU add-in-board partners - Tegra-powered, NVIDIA-designed tablets. NVIDIA doesn't collect the hefty revenue per device, but this is a way for the company to sell chips. Hey, if anything, it shows that management is creative!
GPUs Keep On Winning
While many are bearish on the prospects of discrete GPUs, I think that this is - and has been proven to be - total hogwash. The very high end professional markets (Quadro for workstations, Tesla for HPC) continue to be dominated by NVIDIA, although I am not unaware that Intel's upcoming Xeon Phi may take share from NVIDIA's Tesla parts (although in a growing TAM, this serves to slow growth, not kill the business).
More importantly, it's difficult to ignore that the argument that integrated graphics will displace traditional GPUs is a bit unfounded. The argument would work in a world where games stopped getting more graphically intense, but until 3D games are photo-realistic, I see very little risk to NVIDIA's position as a discrete GPU vendor, particularly since the high end PC gaming segment is GROWING. I also see very little risk for material share loss to AMD (AMD), particularly as AMD's competitive products have failed to make meaningful inroads, signaling that NVIDIA's brand and vendor relationships are strong enough to sustain market-share.
GRID: The Wildcard
NVIDIA's GRID is probably the biggest wildcard. NVIDIA essentially sells boards/entire computing appliances that are basically used to allow graphically/GPU compute intensive applications to be run in the cloud. The TAM for this technology isn't clear, but meaningful adoption could drive material upside to NVIDIA's revenue base. While owning the NVIDIA business at current levels is enough, you get a call option on GRID thrown in for free.
Massive Buyback/Return Of Capital
Last fiscal year, NVIDIA returned about $1B to shareholders via buybacks and dividends, and the company announced its intentions to do so again during FY2015. This is a marked change from the company's previous shareholder unfriendly practices (no dividend and a buyback that went unused), but it's a change that has probably helped drive the slow and steady upside to the shares.
I also am happy with the company's dividend strategy. While $0.085/share on a quarterly basis isn't much, there is certainly room for NVIDIA to grow its dividend, particularly as the NVIDIA Tegra business starts to see some operating leverage (it's posting some pretty sizable losses, largely covered up by the Intel royalty payments). A more pessimistic view would be that if Tegra isn't profitable within a few years (i.e. when Intel royalties end), it will simply shut down Tegra and run the core GPU business for cash.
Next year, I'd like to see Tegra 4i win some meaningful smartphone designs, Tegra 5 claw back some lost share in the tablet space (and expansion into other adjacent compute spaces), Quadro/Tesla continue to grow, and the PC GPU business to see modest growth. If we get all of these (and it's not an unrealistic set of expectations), I believe NVIDIA could be worth $20+ within the next year given that next year's $4.28B/$0.71 consensus may prove to be too low.