Nvidia (NVDA) is one of those stories in the market that takes brass to make a call against given its current position. I did earlier in the year and the stock moved against that call, but the arguments laid out in that article are looking more valid by the day. But when we look beyond the upcoming quarter and into the next couple of years the argument against Nvidia and its technology becomes much clearer. In my mind there are very few avenues in which Nvidia's current trajectory can offset its past. For while the current financials and market position in traditional computing are strong, lurking beneath that is a hollower shell than those bullish on the company care to admit.
Let's start with the Tegra 4. By now even Nvidia's excellent marketing department cannot overcome the massive failure this chip represents. There are no design wins to speak of for it beyond low volume, high price 'performance' tablets. This started back before the chip was even delayed as the previous Tegras over-promised and under-performed on price, heat and capability. OEMs simply will not put up with this for long, ask AMD (AMD) about Llano and Trinity some time.
And, that's what this Tegra situation feels like. The only difference is Nvidia's dominant place in discrete and high-end laptop graphics is covering up a multitude of Tegra-based wins. Nvidia has been able to take market share in these spaces while AMD has gone through its very painful reorganization and implementation of its long-term APU strategy, the fruits of which are just now beginning to take shape. And, still, AMD is having to fight for even one or two decent laptops to showcase its current crop of APUs, Kabini and Richland.
So, now think about that in the context of Nvidia announcing it is finally willing to license its Kepler graphics architecture to chipmakers. Who is going to work with them at this point? It's well known that Nvidia was not allowed to even submit a bid for the PlayStation 4. Do you think Sony (SNE) is the only one not happy with their past association with NVDA? Tablet OEMs were burned multiple times by the Tegra 2 and the Tegra 3, the latter of which only sold well in 2012 due to Qualcomm's (QCOM) inability to ship Snapdragon S4's in sufficient quantity for the Surface RT, which no one bought because it was too limited and too slow.
In a vain attempt to recoup development and manufacturing costs on the Tegra 4 as well as its being shut out of the development of the PlayStation 4, Nvidia revealed its desperation with its handheld Project Shield game controller. On its surface it is a cool device but in no way can it compete in the land of handheld gaming devices at its price point, which had to be dropped even before its release from $399 to $299, just to get someone to order one. According to Semiaccurate, pre-orders for Shield have been so dismal that not even a pre-launch drop in price could stimulate demand from channel partners.
The lunacy of its prices was revealed shortly after Sony announced the pricing on the next PlayStation console. So, let's do the calculus, shall we? For $399 I can buy the latest full-fledged game console supporting the best the gaming industry has to offer and a cutting edge 8-core SoC or I can buy a clunky, small screen handheld device that will ship with its best features in beta and had to be pulled at the last moment due to "mechanical issues?"
By contrast, pre-orders for the PlayStation 4 are at record levels, which is great news for both Sony and AMD, validating both companies' strategies. Console gaming is not dead, despite the recent drops in sales, instead these pre-order numbers for both the PS/4 and Xbox One put paid to the idea that the older generation consoles were simply well past their "born-on" dates.
Both Nintendo and Sony have been burned by building poor handheld gaming devices in recent memory with the Nintendo DS and PlayStation Vita, neither of which sold well and were never going to now that one's smartphone can do most of what those devices could for most people. The disruption of older device markets by smartphones and tablets was obvious for the company to see well before they ever attempted to design Shield.
In essence, Nvidia compounded the failure of the Tegra 4 by building it into another device in which it could lose even more money per unit built than simply selling the chip alone.
So, the Tegra 4 is pretty much dead on arrival both as a tablet and smartphone chip and as a gaming device chip. And any design wins Nvidia has for it, like with Hewlett-Packard (HPQ), are very likely at or near cost of production, which is substantial given its idiotic 4+1 core design.
Neither Nvidia nor AMD have a viable smartphone chip now and AMD is simply not even going to attempt building one, which I believe is a smart play on their part. As good as its Jaguar cores are they simply are not capable of being competitive in the sub 2 watt space needed for a smartphone. Nvidia has the upcoming Tegra 4i, which will offer last year's performance as a quad-core Cortex A9 CPU with a software programmable LTE modem. A nice idea that would have been brilliant at the end of 2012. Unless it can make them for next to nothing and sell into low-end LTE devices by the millions over those of Qualcomm or MediaTek's quad-core Cortex A7s I remain circumspect about its chances either.
All in all, Nvidia's Tegra strategy is in the process of hitting the iceberg while AMD's APU strategy is about to go to the next level with the release of Kaveri in late 2013. This should put even more pressure on any 2 chip CPU/GPU computing solutions as Kaveri's HSA architecture should offer dramatically improved overall system performance per watt in laptop/desktop level systems. Moreover, Intel (INTC) is moving quickly to no longer need discrete graphics solutions in its laptops with its Iris Pro GPUs and Broadwell likely having limited PCIe bandwidth anyway.
Sales of high-end laptops which use discrete graphics are under serious pressure and with new competition at the mid-end by AMD's solid value proposition in the $350 to $550 price range (check out the prices of the latest HP laptops to get an idea) it will be difficult for the Intel/Nvidia solution to compete overall.
Looking at Nvidia's chart over the past few months it is obvious that its price is being supported by both its stock repurchase program along with strong performance by technology ETFs like the SPDR Technology Select ETF (XLK) which has seen $1.22 billion move into it since May 1st. XLK is heavily weighted towards Apple (AAPL) and Microsoft (MSFT), two companies whose stocks are moving in different directions. XLK is off its high but has seen strong investor inflow.
I would watch for further margin erosion from Nvidia as a sign of longer-term weakness, which should weigh on the stock price. Its share buyback plan will also end once free cash flow begins to erode. I would sell any rally in the stock above $15.50 per share as that would constitute a multiple of 16.7, which is too rich to justify such poor performance from one of its core future products and a tenuous relationship with OEMs not named Intel.