At this year’s CES 2017, NVIDIA’s CEO Jen-Hsun Huang presented in the CES opening keynote his vision for the company’s future to the audience, explaining where he believes the company will succeed in the future. I also liked that Huang simplified AI and graphics so more could understand it.
Of the areas where NVIDIA participates in right now, two areas of growth for the company are gaming and automotive, both of which are growing and getting quite competitive. While NVIDIA is still a relatively new entrant into the automotive business, the company has proven that they can keep up with the legacy suppliers and surpass them by delivering innovative automotive solutions. NVIDIA’s leadership in technologies like DNN AI training have proven to be a valuable resource for the company and their partners who are looking to apply machine learning to their own products.
Audi “AI” car
When it comes to self-driving cars, NVIDIA has been partnering with some of the leading automotive manufacturers in the world, like first-mover Tesla. One of NVIDIA’s closest partners is Audi, whom they’ve been working with for quite some time and are now co-developing new technologies with. Specifically, this development of new automotive technology is something NVIDIA believes is important to enable self-driving cars. NVIDIA has the expertise in machine learning and artificial intelligence and Audi has the expertise in cars and automotive technology and what’s necessary for a self-driving car.
NVIDIA and Audi have announced at CES that they expect to deliver a self-driving AI car by 2020 which will be built on the partnership between the new companies. No more details other than “AI car” were presented and I look forward to details on which level of autonomy or co-pilot (see below) is supported in 2020. One of the ways this is possible in this timeframe is because NVIDIA already has a prototype vehicle codenamed BB8 which allows NVIDIA to test their new capabilities and prove them for commercial uses. In fact, at this year’s CES NVIDIA was offering demo rides in BB8.
New AI Co-Pilot and Xavier platform
Most people are enthralled with the notion of self-driving cars instead of what technology could do for human drivers. NVIDIA introduced the notion of “AI Co-Pilot”, which combines all the data about the driver with what’s going on outside the car and warning the driver and even correcting things if and when they go wrong. We’ve all been there, that time the kids are going crazy in the back seat and you take your eyes off that split second before you rear-end someone.
Co-Pilot performs facial recognition and looks for head and gaze tracking. It knows when and where the driver is looking and knows when to signal a warning. NVIDIA didn’t say it, but I can- this will help with distracted driving while looking at a phone or dash, crazy kids in the back, drivers falling asleep and even driving under the influence. This is brilliant and I love it.
I was wondering why we are just now hearing about AI Co-Pilot versus self-driving car. At a breakfast with NVIDIA CEO Jen-Hsun Huang I asked him why that was the case. He told me, “Co-Pilot” is actually harder than self-driving as it looks at the driver and the road”. Very interesting.
Last CES, NVIDIA unveiled what I called their water-cooled automotive supercomputer in a lunch box. The industry was pretty much blown away with this as no one had ever introduced anything this performant, power-hungry or water cooled for a car. This year, NVIDIA introduced “Xavier”, a much lower power derivative, air-cooled at 30 watts. It has 8 custom ARM-based CPU cores and a 512-core Volta-based GPU. Memory wasn’t specified. Xavier will likely be the actual shipping version of its “auto-supercomputer”, not the water-cooled version.
Mercedes, HERE and ZF automotive alliances
On top of the 2020 target date for Audi and NVIDIA to deliver a self-driving AI car, NVIDIA announced a partnership with HERE to partner on AI technology for HD mapping capabilities from cloud to car. HERE is a leading mapping technology provider that’s not Google, and that’s part of the benefit to carmakers. Some carmakers are spooked that Google could ask them for certain terms in the future, like demanding service additions that all carmakers aren’t comfortable with and HERE is an important element of the ecosystem.
NVIDIA also partnered with ZF, a leading automotive supplier to build self-driving cars trucks and commercial vehicles. ZF is well known for their involvement as a supplier for trucks and commercial vehicles, so it makes sense that NVIDIA would expand their involvement in automotive outside just self-driving cars. NVIDIA also announced a partnership with Mercedes to work with them on bringing a product to market within 12 months. These automotive announcements and partnerships underscore NVIDIA’s leadership in artificial intelligence and their commitment to accelerate the growth of automotive innovation. They also position NVIDIA as a central player in the automotive space and not just another PC hardware supplier that is getting into automotive.
It’s interesting to see how the German carmakers are aligning their technology bets and how carmakers are signing up with NVIDIA.
SHIELD TV extends to Google Assistant and home automation with SPOT
NVIDIA also announced a brand-new NVIDIA SHIELD TV. The new SHIELD TV streamer is based on the latest version of Android TV and adds new capabilities that hadn’t previously existed on the SHIELD TV or any TV set-top box for that matter. The new SHIELD TV brings 4K HDR and even more interestingly, Google’s own artificial intelligence assistant known as Google Assistant. The SHIELD TV enables gamers to play all kinds of Android and streaming games on their TV and consume all the highest quality video and audio content available.
The new SHIELD TV also now functions as a smart home hub with the support for Google Assistant as well as support for Samsung’s SmartThings with SmartThings Hub technology integration. I’m starting to really like SmartThings as they have support for over 70 partners and over 200 certified devices. Google Assistant on SHIELD supports Nest but it would be great if SmartThings did too.
NVIDIA also announced the NVIDIA “SPOT” which is an AI microphone accessory that puts SHIELD at the center of the smart home, making things like Google Assistant across the whole home possible. SPOT plugs into any power plug and enables voice input and output in any room, like a mini-Echo Dot.
I like what NVIDIA is doing here because they are the first big player to try to integrate the full living room entertainment experience with home automation and AI. My entire house is “smart” and am looking forward to enabling it with the new SHELD and SPOT.
GeForce NOW game streaming enhancements
In addition to the plethora of AI announcements, NVIDIA also announced they would be enhancing the GeForce NOW service. Previously, the GeForce NOW service was only available via NVIDIA SHIELD devices. This technology was originally known as “GRID” and was subsequently branded as GeForce NOW, but is still the same cloud gaming technology that distributes gaming usage across large banks of GPUs in a datacenter at low latency.
The GeForce Now subscription service is coming soon to the PC and Mac. There are different levels of gaming performance in the NVIDIA GeForce NOW service which are dependent upon what level of visual effects the user desires. This service is specifically aimed at consumers who want to play games but have outdated or integrated graphics in their PC. I can also see it appealing to Mac users, who don’t have access to as many games as Windows PCs. It does cost $25 for 20 hours of game play, which gives NVIDIA another source of revenue and source of demand for GPUs. I like the ability to try a game out before I commit $60 to a lousy game.
I believe cloud gaming is still a technology in development and still has hurdles in a lot of places where latency or bandwidth can be an issue. After all, this is a game streaming service, not local gaming which usually uses much less data. This can be especially problematic when you consider that some ISPs are starting to cap their users’ monthly bandwidth. I believe that NVIDIA will see some growth with their GeForce NOW program by extending it beyond SHIELD devices, but they will need to market it to a tight audience for it to succeed.
A very strong CES for NVIDIA
NVIDIA had a very strong CES. I see it as a very big deal that NVIDIA did the opening keynote. If I look at the companies who did this before them over the years, they were most relevant and giants of their time, This is NVIDIA’s time as they are the current leaders in DNN training for AI, self-driving car technology, and the highest performance PC gaming and PC VR.
NVIDIA has long been known as a company with a strong gaming pedigree, but their recent leadership in artificial intelligence is starting to yield them meaningful relationships with big partners. Some of these big automotive partners like Audi, Mercedes, HERE and ZF should to be able to help NVIDIA get a bigger presence in the automotive market and bring more innovation to the automotive market.
I believe NVIDIA truly understands how big compute-heavy AI can be and they know that they are one of the few companies in the world that can solve this problem. Their solutions to the problems that self-driving cars present could ultimately cement their position as a crucial supplier in self-driving automotive technology. I also think that NVIDIA is taking a unique approach to the living room entertainment experience and pairing it with home automation and AI and it could pay off for them down the road.
NVIDA had a very good CES.