There seems to be a lot of confusion around NVIDIA's (NVDA 2.13%) Tegra program. As a quick reminder, Tegra is a line of ARM Holdings (ARMH +0.00%)-based system-on-a-chip products from NVIDIA that is intended to go into select smartphones, tablets, and automotive. Though the company's strategy surrounding the product line has shifted, it has not been "killed" as some would like to suggest.
Explaining the shift
When NVIDIA was first promoting its Tegra product lineup, the big selling point was that it would have an absolutely gigantic total addressable market to go after. Thanks to Google's Android operating system enabling the mass-market mobile computing revolution, NVIDIA saw an opportunity to sell a vast number of chips.
Additionally, Microsoft had promoted its ARM-based version of Windows 8 -- known as Windows RT -- as bringing more processor choice to Windows. NVIDIA, which had long coveted an X86 license in order to compete with Intel and Advanced Micro Devices in the PC processor space, saw Windows RT as an opportunity to compete in PCs without the X86 license.
It all sounded great, but of course reality didn't quite match up with the initial optimism.
Windows RT flopped
The bad news for Tegra on the Microsoft front was that Windows RT was a complete failure. The problem here was that Intel managed to release largely comparable, price-competitive low-power parts that also had full Windows 8 compatibility. What was the value proposition in buying a Windows system that didn't run any of the millions of pre-existing Windows applications?
Android smartphones -- a low-cost battlefield owned by Qualcomm and MediaTek
In the Android smartphone market, which is where NVIDIA's Tegra would be able to play, the market is bifurcated into essentially the low end and mid-range, and then the high-end "hero" devices. The high end of the Android market is largely dominated by Samsung Electronics (NASDAQOTH: SSNLF), which uses its own custom-designed processors when possible, though it often sources from Qualcomm (QCOM 2.63%) for its high-end devices, and other smaller players (as well as Qualcomm) for the low end.
The low end of the market isn't dominated by a single player, but this is also a market where performance matters far less than cost structure. NVIDIA, as a performance-focused vendor, probably wouldn't be able to successfully play in this market with the assets that it had built up as Tegra has evolved.
Tablets are different
One of Qualcomm's structural advantages is that it has the world's best cellular modems and outfits even its lowest-cost chips with top-shelf modems. Given that the cellular piece of a smartphone is so vital to the experience, competing with Qualcomm is just really hard. Just look at how much Intel is sinking into this venture to try to catch up.
The very low end of the tablet market is probably difficult for NVIDIA since, once again, the company is designing relatively large-die (i.e., expensive to make), high-performance products. However, for higher-performance tablets (particularly ones for which graphics performance is important), NVIDIA's world-class graphics IP is certainly helpful. The Xiaomi MiPad, for example, offers performance that runs circles around an iPad Air for $240 -- enabled by NVIDIA's Tegra K1.
NVIDIA's Tegra focus vis-a-vis consumer tablets is in trying to win sockets where this kind of performance is appreciated.
So, Tegra isn't dead -- but it has evolved
Even though NVIDIA is probably never going to have the volume of microprocessor shipments that Intel or Qualcomm enjoy, Tegra doesn't need those kinds of volumes to be considered successful in the consumer space. Now, consumer probably isn't enough to drive Tegra to profitability on its own, but consumer combined with the fairly sizable automotive opportunity that the company has been exploiting for quite some time could be enough.
Has the scope of Tegra changed? Sure. Is Tegra dead? I seriously doubt it.
