NVDA Stock Predictions: Nvidia Should Also Take Lead In Mobile-Centric Artificial Intelligence

motek 1The article was written by Motek Moyen Research Seeking Alpha’s #1 Writer on Long Ideas and #2 in Technology  – Senior Analyst at I Know First.
NVDA Stock Predictions – Summary:

  • Nvidia has spent billions of dollars developing its GPU-accelerated artificial intelligence computing platform.
  • Apple used its WWDC 2016 event to unveil its on-device deep learning initiative for iOS devices.
  • Nvidia’s big lead in GPU-accelerated machine learning will help it compete against Apple’s on-device intelligence technology.
  • Nvidia is now a leading chip supplier for machine-learning computers. It should also try to become a lead chip supplier in machine-learning mobile devices.
  • In spite posting a +41% YTD performance, NVDA still has positive short and long-term algorithmic forecasts.

Nvidia (NVDA) spent $2 billion to research and develop the world’s most powerful GPU for artificial intelligence computing, the Tesla P100. I suggest that Nvidia should also spend on researching how to compete with Apple’s (AAPL) upcoming on-device deep learning technology.

Nvidia’s huge 9-figure investment toward making its league-leading GPUs the go-to processors for artificial intelligence deserves a bigger total addressable market. Yes, Amazon (AMZN) will likely buy a lot of Tesla P100 GPUs for its deep-learning computers. However, investors will appreciate it if the Nvidia GPU-accelerated deep learning platform gets expanded to include phones, tablets, and Internet of Things devices.

The commercial applications of artificial intelligence is not limited to data center servers. Consequently, Nvidia’s growth prospect is better when it can expand its GPU-accelerated artificial intelligence ((AI)) solutions to serve all sectors. As you can see from the chart below, supplying Artificial Intelligence to the Digital Assistants product category (which I believe includes smartphones/tablets) is predicted to grow to a $2.18 billion industry by 2019.


Nvidia Already Has A Deep Learning Mobile Processor

Apple is bullish that its A-series of processors can do on-device machine learning. The more powerful integrated GPU of the Tegra X1 should also make it a potent on-device deep learning processor for tablets and smartphones. The Tegra X1’s 256-core Maxwell GPU is certainly going to be a better on-device deep learning processor than the integrated 6-core or 12-core PowerVR Series 7xT GPUs of the Apple A9x.

An Nvidia pre-emptive strike against Apple’s forthcoming on-device artificial intelligence is sweet revenge over Tim Cook’s recent shift toward Radeon GPUs for Macs. When it comes to AI, the early bird advantage belongs to Nvidia. Aside from buying Siri and several AI-related companies, Apple is still a non-factor in the world of commercial AI services.

I really do not know how Apple will implement its on-device artificial intelligence framework for phones. Apple has yet to announce any Artificial Intelligence SDK (Software Development Kit) for its A-series processors. On the other hand, Nvidia already has the Tegra X1 and a solid reputation for GPU-accelerated AI computing.

Nvidia previously announced the Tegra X1 as the go-to processor for AI computers inside autonomous automobiles. If the Tegra X1 can be the brain of very complex self-driving cars, it should also be suitable for on-device artificial intelligence inside smartphones/tablets.


(Source: Nvidia)

Just like the Tesla P100, the Tegra X1 is CUDA programmable. CUDA is Nvidia’s proprietary parallel computing platform that allows C or C++ programmers to harness the power of GPUs to accelerate custom Artificial Intelligence software. The Tesla P100 can rule the deep learning computers of data centers. The Tegra X1 (or a modified version of it) could flourish inside deep learning gadgets like smart TVs, tablets, and smartphones.

The battery requirement that a 256-core Maxwell GPU will need is no longer an issue. There are now Android phones that are equipped with 5,000-7,000 mAH batteries. I’m sure Nvidia has enough money to find a way to make a deep learning Tegra X1 compatible with Android phones equipped with 5,000 mAH batteries.

Nvidia has already managed to put a Tegra X1 inside its upcoming 8-inch Shield 2 Android tablet. The Shield 2 tablet has a 5,100 mAH battery. Squeezing the Tegra X1 inside a 5 or 5-inch phones is not an impossible task.

Nvidia Cannot Ignore Mobile-centric Artificial Intelligence

Self-driving cars are moonshot projects that are still years away before they become commercially available. Smartphones like the iPhone are now available to accept GPU-accelerated artificial intelligence programming. Yes, the Tegra X1 never received any design win from Android phone manufacturers.

However, I believe Nvidia could make the Tegra X1 more palatable to phone vendors if it imitates Qualcomm’s (QCOM) recent move to market its Snapdragon 820 as an Artificial Intelligence-compatible processor. Aside from Apple, Qualcomm is also serious in bringing artificial intelligence to mobile devices. Qualcomm’s Project Zeroth is an optimized SDK or toolkit dedicated to bestowing machine learning features on Snapdragon-powered smartphones and tablets.

Nvidia’s management will be remiss in its duty if it allows Qualcomm to lead in mobile-centric artificial intelligence. Qualcomm previously knocked out Nvidia’s $352 million investment in Icera modems. Nvidia cannot afford to just watch Qualcomm to again knock it out of mobile-centric artificial intelligence. Complacency could kill Nvidia’s current lead in GPU-accelerated AI.

Qualcomm’s integrated Adreno GPU on its Snapdragon processors is courtesy of the GPU assets it bought from AMD. Apple is also a part-owner of Imagination Technologies, the British firm that makes PowerVR GPUs. Apple even tried to buy Imagination Technologies three months ago. These loud warning signs should be heeded by Nvidia. Apple and Qualcomm are serious in their quest toward becoming relevant in commercial artificial intelligence. They both pose a clear and present danger to Nvidia’s GPU-accelerated artificial intelligence platform.

Nvidia has no choice but to immediately challenge these two ambitious companies on mobile-specific artificial intelligence.

Smart Home Appliances Are Also Ideal For Tegra X1

The smart home assistant Echo made by Amazon (AMZN) is an ideal product for deep-learning Tegra processors. The Alexa-powered Amazon Echo used an ARM-based processor made by Texas Instruments (TXN). Nvidia could call Bezos and offer him an irresistible bulk discount deal on Tegra X1.

A Tegra X1-powered Amazon Echo can definitely match the deep learning capabilities of Alphabet’s (GOOG) upcoming Google Home smart home assistant. I believe Amazon and Nvidia have a cordial working relationship. In spite having its own Android smart TV products, Amazon has not yet banned Nvidia’s Shield Android TV and Shield tablet/console products.

Empowering household appliances with GPU-accelerated artificial intelligence is a future endeavor that could help Nvidia increase the total addressable market of its processors. As per Mary Meeker’s chart below, smart home appliances like the Amazon Echo is just starting to gain mainstream acceptance. They could become as popular as smartphones.

Sans titre

Final Thoughts

I am staying long NVDA. The stock has notably outperformed QCOM and AAPL this year. This stock is unlikely to suffer any untoward dips this year. NVDA’s YTD performance is already +41.47% but I still see upside potential. It saddened me that some people (like Seeking Alpha editors) cannot comprehend the killer advantage of a 256-core GPU-equipped Tegra X1 when it comes to deep-learning smartphones.


(Chart Source: Finbox.io)

Nvidia has a big lead on GPU-accelerated deep learning computers. However, it should also try taking the lead in deep learning mobile devices. Letting Apple or Qualcomm take the initiative on mobile-centric artificial intelligence could eventually diminish Nvidia’s presence in data center-centric deep learning computers.

Apple and Qualcomm are both cash-rich firms that has 9-figure R&D budgets. They are also both trying to diversify their revenue streams. We cannot discount the possibility that Apple and Qualcomm, who both have GPU IP assets, could eventually create GPU-accelerated artificial intelligence platforms that could challenge Nvidia’s CUDA Deep Neural Network.

NVDA Stock Predictions based on Artificial Intelligence

My still bullish sentiment over Nvidia is backed by the still-positive algorithmic forecasts from I Know First’s deep-learning neural network computers. The 1-month, 3-month, and 1-year algorithmic trend signal scores of NVDA are all positive. There is obviously more probability that NVDA will go up higher in price this year.


I Know First Algorithm has previously predicted the stock movement for NVDA like in this forecast from the May 22 2016 to June 5 2016  showing the bullish signal of 34.43 and predictability of 0.16  achieving to return 5% in just 14 days.

About I Know First’s Algorithmic Forecasts

The underlying technology of the algorithm is based on artificial intelligence, machine learning, and incorporates elements of artificial neural networks and genetic algorithms through which we analyze, model, and predict the stock market. The algorithm is adaptable, scalable, and features a Decision Support System (DSS) to optimize the information produced by the years of data inputted. The algorithm produces a forecast with a signal and a predictability indicator.

The sign of the signal tells in which direction the asset price is expected to go (positive = to go up = Long, negative = to drop = Short position), the signal strength is related to the magnitude of the expected return and is used for ranking purposes of the investment opportunities.

Predictability is the actual fitness function being optimized every day, and can be simplified explained as the correlation based quality measure of the signal. This is a unique indicator of the I Know First algorithm, allowing the user to separate and focus on the most predictable assets according to the algorithm. Ranging between -1 and 1, one should focus on predictability levels significantly above 0 in order to fill confident about/trust the signal.