How do you measure success in the mobile application processor business? NVIDIA recently revealed that their Tegra 2 processor accounts for 70% of the non-iPad tablet market, but it cost them $2 billion dollars to get there and Android tablets haven’t been exactly flying off the shelves yet (until recently).
Company CEO Jen-Hsun Huang has previously said that if you want to build and sustain a world-class SoC (system-on-a-chip) business, then you need that business generating up to $1 billion in revenue to remain competitive. Analysts are expecting Tegra revenue to come in between $400-600 million this year, so sales would need to double in 2012 to reach that goal.
Most of us with Honeycomb tablets will agree that the Tegra 2 is a fine processor, but most of the 400 plus design wins that NVIDIA racked up were thanks to its advantage of being the lead platform for Android 3.0. Texas Instruments will enjoy that luxury as we migrate to Android 4.0 devices and the mobile SoC business is heating up with Intel expected to join the party, so competition will be fiercer than ever in 2012.
Can NVIDIA actually double their Tegra business and record $1 billion in revenue next year? This time last year Tegra had about 0% market share, so it would be a big achievement for them to pull it off. As we gear up for the Tegra 3 launch, it certainly looks like their strategy could pay off big time.
Quad-core ultra performance comes to Android
History has shown us that most new mobile processors only enjoy around 3-6 months as the new kids on the block before a competitor comes out with a slightly faster product and woos away the early adopters. However, when you bring to market the first mobile quad-core processor months ahead of the compeition, then you will attract a lot of high-end buyers during the important Holiday shopping season.
Just as Amazon has taken over the bargain-price tablet market with their Kindle Fire at $199, I expect ASUS will corner the market on high-end tablets with their quad-core Transformer Prime. We expect that device to go on sale next month, and I don’t see any other tablets on the horizon that will be able to compete this year.
comScore report bodes well for Tegra Zone
When it comes to differentiating their mobile platform and promoting their unique user experiences, NVIDIA has invested heavily in Tegra Zone to bring more premium games to Android devices. So far that has resulted in 22 games, a full-blown community site, and over 1 million installs of theTegra Zone app.
A recent report by comScore shows that two thirds of tablet owners play games on their device, so expect NVIDIA to keep pushing this strategy even more as their Tegra business grows.
ARM’s new Cortex-A7 big.Little processing concept validates NVIDIA’s “companion core” strategy
We have known about NVIDIA’s quad-core Tegra 3 processor (aka Project Kal-El) since early this year, but we were recently surprised to learn that it includes a stealth 5th “companion core” that promises to extend battery life.
This strategy of including an extra low-power core that runs at slower speeds has been debated by competitors. Qualcomm uses their own asynchronous symmetric multiprocessing (aSMP) technology and said it eliminates the need for ‘companion’ or ‘little’ cores since each core in an aSMP system can be operated in low power mode due to the independent voltage and frequency control per core.
On the other hand we just saw ARM announce a similar concept called big.Little processing which pairs a lower-power Cortex-A7 processor alongside a faster Cortex-A15 processor. ARM said this type of implementation could extend battery life by up to 70%, so it will be interesting to see how NVIDIA and Qualcomm’s solutions measure up.
So far it sounds like ARM is impressed by what NVIDIA has done with Tegra 3. Dylan McGrath of EETimes just posted an interview with Peter Greenhalgh, an ARM engineer who served as the lead designer on the A7, and he said, “”We find it very interesting to see what Nvidia has done. It’s fantastic, and it validates big-little.”
The ability to surprise is key
NVIDIA certainly surprised the industry with their “companion core” and they will need to keep pushing innovation to catch up to the large incumbents like Qualcomm and Samsung. At last week’s AsiaD conference, NVIDIA also revealed that Tegra 3 will include a special image-processing technique that “radically simplifies the way color is displayed , saving ‘tons of power’ without reducing the visual experience.”
We still don’t know how much Tegra 3 can extend battery life compared to the current generation of mobile processors, but NVIDIA says their 5th “companion core” uses 20 times less power than the four main cores. If this super low power mode works as described, we should see devices with industry leading standby time and battery life.
Larger competitors like Intel are already touting their advantage in process technology and the race to smaller transistors, so it will be key that NVIDIA can “surprise and delight” its customers with “magical” products.
Cars, TVs, and baseband processors
Tegra has dominated Android tablets, but there are some new areas where it should gain market share in 2012. Next year we expect to see Tegra gain share in smart TVs, cars, and baseband processors.
Not many details have leaked out about the Google TV 2.0 refresh, but Tegra 3 looks poised to play a role in some those products. It’s capable of decoding 1440p video, which would look awesome on the next-gen of HDTVs. Intel is winding down their Digital Home Group that was responsible for the processor in the first-gen of Google TV devices, so that opens the door for the ARM architecture to take over future products.
We have already seen Tegra inside several cars in years past and that will continue to grow next year. All Audis will be powered by Tegra in 2012 and others have reported we will “see a lot of cars with NVIDIA” early next year. Google has at least 50 people working on self-driving cars, so we should see more advanced mobile application processors coming to the road over the coming years.
Finally, another opportunity for huge growth is NVIDIA’s new baseband processor business, which they gained by acquiring Icera for $367 million this year. Outside of the application processor, the baseband processor is the next most important chip in most connected devices and NVIDIA thinks they have something special in Icera.
Expected to appear in both smartphones and tablets early next year, the Icera soft modem offers full Multimode LTE support. When paired with a Tegra 3 application processor, NVIDIA would now control the two most important chips inside a mobile device. They also plan to build a unified Tegra chip with an integrated modem called Grey, which will allow them to compete with Qualcomm’s fully integrated Snapdragon platform in the low-end part of the smartphone market.
It all comes down to product timing
When asked about Tegra’s outlook for the rest of this year during last quarter’s earnings call, NVIDIA CEO Jen-Hsun Huang responded that a lot of it came down time timing and well-positioned product launches.
And so the only thing that I can watch is how many design wins do we have? And are they high-quality design wins from high-quality customers and OEMs? And when the phones and the devices come out, are they well-positioned? Are they really well-designed? And are they well-positioned as a product?Jen-Hsun HuangNVIDIA CEO
A quick glance at Texas Instruments and their OMAP4 timing confirms most of this theory. In the span of a couple of weeks, we saw all the big holiday product launches (including the Droid Bionic, Droid RAZR, Galaxy Nexus, and Kindle Fire) go with a Texas Instruments processor, which will easily generate hundreds of millions in revenue.
It will take multiple well-positioned product launches for Tegra 3 to match those numbers, so we will have to wait and see which designs they win. ASUS will be first to market with a Tegra 3 tablet, but I expect many more existing Tegra 2 customers will return to launch new products built around Tegra 3. Tablets will get all the attention at first, but there will likely be more Tegra 3 chips inside smartphones (that currently account for 2/3 of all Tegra sales).
In closing it looks like Tegra still has the potential to become a multi-billion dollar business, but there are so many factors involved that it’s hard to predict when it will happen. NVIDIA is already working on the next three generations of Tegra at the same time, so they remain confident in the growth of mobile. Any delay of the next Tegra release cycle would be very costly, but NVIDIA still has around $2.5 billion in cash and Tegra is a long-term investment they are going to keep spending on.
Don’t expect NVIDIA to maintain their dominance in non-iPad tablets, but the rising tide of Android (and possibly Windows 8) will take Tegra to new heights in 2012.