Quantcast
 
 


CES 2017: NVIDIA Keynote

by MAC     |     January 5, 2017


Jen-Hsun Huang – CEO and founder of NVIDIA – delivered one of the major keynotes of CES 2017.


This is the first year that NVIDIA has had the opening keynote slot at CES. The keynote is supposed to provide insights into the technologies changing the consumer electronics industry and people’s day-to-day lives. While NVIDIA has been a staple of the PC hardware industry since shortly after its inception, 2016 is the year that it unexpectedly emerged as one of the hottest names in technology... and surprisingly it has almost nothing to do with graphics cards.

While gaming GPUs are indeed NVIDIA’s bread and butter, and by far their largest source of revenue, that is not why many people started paying attention to this 17-year old company this year, nor why their stock priced has tripled in a single year. That gaming money is just what NVIDIA has been using to subsidize all their other initiatives, and their eclectic portfolio of products have really started to gain traction.

While it is really nerdy and not all that interesting from a consumer product point-of-view, NVIDIA’s role as a leader in artificial intelligence and deep learning was highlighted throughout the hour-plus keynote. NVIDIA’s deep learning solutions for cloud computing – powered by Tesla GPUs – are transforming the everyday lives of ordinary people, because they are utilized by the likes of Google for their searches, Amazon’s Alexa digital assistant, Netflix‘s movie recommendations, and Tesla’s self-driving cars. Those Tesla GPUs are essentially the brains behind most artificial intelligence simply because GPUs have massive computational power. They excel at parallel tasks and can process trillions of deep learning operations per second. That is one for GPU, imagine the potential when you pair a couple thousand together in one or many datacenters.


NVIDIA has a few GPU-powered datacenters too, and they just announced a new cloud service that is going to put them to good use. Mr. Huang revealed the GeForce NOW game-streaming service, that is obviously based on their NVIDIA GRID graphics virtualization technology. We wrote an entire article dedicated to the specifics, but essentially this is cloud gaming on a virtual PC or more specifically Gaming as a Service (GaaS).

They announced some new features for GeForce Experience, which is a simple way of helping gamers capture and share their gameplay content with others. They added Facebook support, which give gamers the ability to broadcast live to Facebook and to upload those saved recordings and screenshots directly to their Timeline. Overall, this is going to bring gamecasting to a huge new and more mainstream audience.


The announcement of the updated SHIELD TV ended being more exciting than expected. The design has been streamlined, and a remote control has finally been added to the bundle, but the internals appear unchanged. That is okay, since it remains the most powerful media streamer on the market. The announcement wasn't really about the hardware though, it was a declaration that NVIDIA now had the best home entertainment device on the market, period.

Thanks to Android 7.0 - this is an Android TV box after all - the SHIELD TV now supports Netflix 4K HDR, Amazon Video 4K HDR, Google Play Movies & TV 4K, VUDU 4K, and YouTube 4K. And in a few months it will also be able to play YouTube 360° videos. That is an unmatched level of access to 4K and HDR content.

The SHIELD TV is not only a media playback device though, it aims to be a serious gaming console. There are now over 1000 fully optimized games in the SHIELD Store, over 100 of which NVIDIA describes as "console-class. With GameStream you can cast games directly from your computer to the SHIELD TV - in 4K HDR - and if you're home computer isn't up to the task you can use the new GeForce NOW cloud gaming service. Another huge announcement was that gamers would soon have access to both their Ubisoft and Steam libraries, which is pretty awesome.

Those who already own a SHIELD TV will be glad to hear that all of these features will be provided via a software update. If you want to know more about this device, we have written a more thorough article right here.


What is that little ball that Jen-Hsun Huang is holding? Well it actually still relates to the SHIELD TV. In an effort to bring AI to the home, Google Assistant has added to the SHIELD TV. No other device aside from Google's high-end Pixel phones have had this unique voice-based intelligent personal assistant. How will you interact with the Google Assistant? Well the newly included remote has an always-listening microphone built-in and it supports a number of hands-free TV commands. To expand the AI functionality to the entire home, NVIDIA have also just announced a peripheral for the SHIELD TV, the NVIDIA SPOT microphone/speaker accessory. This golf ball-sized Wi-Fi-connected gadget plugs into an electrical outlet, and it can hear your commands/questions from up to 20 feet away. It then wirelessly routes the queries to the SHIELD TV and playbacks the reply the nearest SPOT device. Availability is a few months away, it will retail for $49 USD.


Next up was a focus on transportation technology. But first, a little background information. While NVIDIA’s Tegra chips have never really found a home in mobile devices, they were perfectly suited to power in-car displays, be it consoles or dashboards. As infotainment systems started featuring higher-quality graphics, and more & more features (think navigation, rear cameras, media streaming, etc), NVIDIA was easily able to supply more powerful and more power hungry Tegra chips that just wouldn’t be suitable for mobile devices. As a result, they have developed very lucrative contracts with the automotive industry, and they are about to get even more lucrative. Modern cars, especially fully autonomous self-driving cars need to crunch through a lot of data, like one gigabytes of data per second worth of data. And it just so happens that GPUs can handle that specific workload better than CPUs can.

Now only do they have the hardware – they unveiled the Xavier AI Supercomputer SOC back in September – but they also have the software in DriveWorks OS. With this combo, NVIDIA are aiming for Automotive Safety Integrity Level D, which is the most stringent level of safety measures for self-driving cars. Level D is a 100% purely autonomous car that requires zero human input while driving. And they even had an example of it when they showed a brief video of their very own BB8 autonomous car.

During the keynote, they revealed their DRIVE PX 2 AI computer - a 250W liquid cooled module capable of 24 trillion deep learning operations per second - and their overall AI car platform that combines numerous features that detect the world inside and outside the car. Self-driving cars are the way of the future, and NVIDIA looks to have everything required to make it happen. There is an unbelievably amount of money to be made if they succeed, and that is why the world is now paying attention to them more closely than ever before.


Last, but not least, it was announced that Audi and NVIDIA would partner to develop next-gen AI for cars, and that they would have the world's most advanced self-driving car in 2020. It was specifically mentioned that they were working to make sure that their kids would never have to learn to drive. That is a pretty amazing statement if you think about it.

So overall, the keynote set out the company's vision for the future of AI, gaming, TV, and self-driving vehicles. NVIDIA aims to be at the technological spear-point of all those areas, so far it looks like they are succeeding.
 
 

Latest Reviews in Events
January 8, 2017
Jen-Hsun Huang – CEO and founder of NVIDIA – delivered one of the major keynotes of CES 2017....