In 2012, NVIDIA (NVDA +0.97%) won the Google (GOOG +2.11%) (GOOGL +2.10%) Nexus 7 tablet. However, when 2013 rolled around, Qualcomm (QCOM +0.49%) had snatched that particular design. Google didn't bother releasing a new Nexus 7 tablet last year, but did launch an 8.9-inch tablet known as the Nexus 9. This was powered by NVIDIA's 64-bit Tegra K1 processor.
While winning a design in a given year does not guarantee a chip vendor will still hold that design in the following year, I believe NVIDIA is well positioned with its Tegra X1 to win the next Nexus 9, should Google roll one out.
Tegra X1 will likely offer the best performance for a tablet processor
In the Android merchant mobile system-on-chip market, only a handful of vendors still develop high-performance tablet-focused chips. These vendors basically come down to NVIDIA, Qualcomm, and Intel (INTC 1.43%).
For tablet processors launched in the 2015, Qualcomm's highest-end chip is likely to be the Snapdragon 810, Intel's will be Cherry Trail, and NVIDIA will have the Tegra X1. While Cherry Trail seems likely to offer excellent performance and power for CPU and graphics, I believe it will be underpowered in a number of key metrics for a flagship device (lack of H.265 decode/encode support, for example, a feature present on the NVIDIA and Qualcomm processors).
This leads me to believe Cherry Trail will be out of the running as a "halo" device. This leaves the Snapdragon 810 and the Tegra X1. While both will likely feature roughly the same CPU performance (both are implementations of ARM's (ARMH +0.00%) Cortex A57 in Taiwan Semiconductor's (TSM +4.75%) 20-nanometer process), I believe NVIDIA's X1 will deliver superior graphics performance.
What's the potential risk to this thesis?
One thing to note is that while most tablets are Wi-Fi only, Google did offer cellular versions of the Nexus 9. Those devices featured the NVIDIA Tegra K1 chip but paired it with a Qualcomm stand-alone baseband. If Google purchases a single integrated package, then what I believe to be NVIDIA's graphics performance advantage might be offset by Qualcomm's ability to offer a single integrated chip.
That said, while integration of baseband into the main system-on-chip is critical in the more cost-sensitive low-end and midrange portions of the mobile device market, higher-end tablets and phones still often ship with stand-alone cellular baseband. The Nexus 9 family is premium-priced, so the risk to NVIDIA of baseband integration being a make-or-break feature seems reasonably low.
Winning this design is relatively important to NVIDIA, not so much for Qualcomm
Another reason I strongly believe a next-generation Nexus is likely to go to NVIDIA is that it's more meaningful to the company than it is to Qualcomm. The Tegra K1 only found a home in a handful of tablet devices (Nexus 9, Xiaomi MiPad, and Shield tablet, to my knowledge), likely because the non-Apple/non-Samsung market for premium tablets isn't particularly huge.
Qualcomm has significant chip volume in phones and tablets of all tiers, so the Nexus 9 "slipping away" would by no means be the end of the world. It wouldn't be the end of the world for NVIDIA, either, particularly as the company seems focused on other market segments (automotive and embedded, for example), but it would be meaningful since losing those designs would create a difficult year-over-year revenue comparison for Tegra next year. To that end, NVIDIA might be a bit aggressive on pricing to keep that design.
We'll have to see how everything pans out, but I'm fairly confident the Tegra X1 will find a home in a next-generation Nexus and a next-generation Shield tablet. I also would not be surprised by a Tegra X1-powered Xiaomi MiPad 2.
These designs in aggregate will not drive NVIDIA's Tegra to new heights, but maintaining those designs while pushing further into automotive, embedded, and other segments will certainly help the company get that much closer to breakeven (and potentially profitability over the longer term) in its Tegra division.
