CES 2017: AI Comes to World’s Largest Tech Show

By Jen-Hsun Huang, Founder & CEO, NVIDIA

Note: I was invited to give the opening keynote at this year's CES, which was celebrating its 50th anniversary. Below is a summary of my talk and our announcements. In the end, I described how the investments we've made over the past 24 years have put NVIDIA at the center of the most exciting time the technology industry has ever known.

50 Years of the Future

CES has been a bellwether of technology trends for five decades. From the first pocket radios and VCRs and video game consoles in its early days, to the PC revolution in the '90s, to the mobile revolution that put the internet in our pockets and forever changed how we connect and share, we've watched the future take the stage at CES.

This year, the world's largest technology trade show showcased the latest advances of the greatest computing challenge of all time - artificial intelligence. From intelligent assistants to smart homes to self-driving cars, it's clear that this new computing model will infuse consumer technology as much as it will reinvent enterprise computing.

NVIDIA - Amplifying Imagination & Intelligence

What was recently science fiction is now reality. This is unquestionably the most exciting time in the computer industry.

Our work at NVIDIA centers on a computing model focused on visual and AI computing. Built on top of the GPU that we pioneered, this computing model can solve problems that normal computing simply can't. It gives us a unique position in our growth markets - gaming, virtual reality, data center and self-driving cars - and allows us to tackle the most challenging computing problems in the world.

Several years ago, our computing model was discovered by researchers who were applying a new technique to rejuvenate the decades-old field of artificial intelligence. Deep learning allows software to learn from data, the machine equivalent of experience, but it requires an enormous amount of processing power to be practical. Enter the GPU. With its unmatched parallel processing capabilities, GPU deep learning ignited the 'big bang' of AI.

Breakthroughs have come fast and furious. A year ago, DeepMind's AlphaGo beat the world champion at Go, a game that has more moves than there are atoms in the universe. A few months ago, an AI network learned to play Doom, a game of maze navigation, resource finding and survival. Today, AI networks can write captions of photographs and imitate the painting styles of van Gogh and Monet. AI networks have learned the fine motor skills required to open a bottle of water, how to get up off the floor and walk, and even how to drive a car on dirt roads in the rain. The achievements of GPU deep learning have been nothing short of miraculous.

This breakthrough work is underwritten by an R&D budget fueled by the largest entertainment industry in the world - video games.

PC Gaming Is Thriving and GeForce Is PC Gaming

The single largest game platform, PC gaming doubled to $31 billion over the last five years. Our GeForce business grew in step, with its revenue doubling. Today, 200 million gamers around the world play on GeForce.

[Attachment]

And there's no sign of a slowdown. Gaming technology has increased in performance by a factor of 10 over the last five years. Now, 4K, HDR and virtual reality are coming. Gaming is the world's largest sporting event, with 100 million players and hundreds of millions of spectators. And gaming is now social - 2 trillion minutes of video games were watched on social sharing sites last year.

For the fast-growing market of gaming notebooks, we launched GeForce GTX 1050, which was adopted in a record 30 designs. We also added to our lineup of 60 G-SYNC monitors and laptops with the first HDR G-SYNC monitor, which PC World described as 'the holy grail' of PC displays. And to help fuel the growth of gaming as a social medium, we announced that the GeForce Experience application will seamlessly connect directly to Facebook's 1.8 billion monthly active users. Now, gamers using GeForce Experience to capture images and video can connect to Facebook Live and share in just two clicks. We showcased the technology with never-before-seen gameplay from Mass Effect: Andromeda, one of the most anticipated games of the year.

[Attachment]

While a GeForce gaming PC is the best choice for core gamers, we estimate that there are 1 billion non-core, new or infrequent gamers whose PCs are not ready for today's games. To serve this large market, we announced that the GeForce NOW game streaming service is now available for Macs and PCs. Built on GeForce GTX 1080 PCs in the cloud, GeForce NOW turns any computer into a powerful gaming rig.

It's integrated with the major game stores - Steam, Origin, Uplay and more - and supports multiplayer.

[Attachment]

GeForce is PC gaming and it's thriving.

Bringing AI to the Home with Most Advanced Streamer

CES has showcased the state of the art in TVs since its early days. About a year and a half ago, we set out to reinvent TV with SHIELD, the world's first Android TV. At CES, we introduced an updated version of the most advanced streamer. With the addition of Amazon Video in 4K HDR, SHIELD offers the largest, most open catalog of media in stunning 4K - also supporting Netflix, YouTube and Google Play Movies. SHIELD, which boasts more than 1,000 games, is also the first Android device to integrate the Steam app.

[Attachment]

We didn't stop there. We advanced SHIELD to bring AI into the home. SHIELD will now support Google Assistant hands-free as well as SmartThings - turning it into a smart home hub that can connect to hundreds of devices. And to extend SHIELD's reach beyond the living room, we introduced SPOT, a SHIELD accessory and AI mic that extends intelligent control throughout the house.

[Attachment]

Here's a taste of how SHIELD and SPOT can make AI ambient in your home.

Reinventing Transportation with AI

CES has become the show where the technology and automotive industries come together. This year was no exception.

[Attachment]

AI will revolutionize the $10 trillion transportation industry. There are 1 billion cars on the road. In a single day, 20 million rideshares are hailed on just Didi and Uber. There are 300 million trucks on the road that drive 1.2 trillion miles per year. And there are half a million buses in just the largest cities of the world. The transportation industry is the basic infrastructure of our society, yet it produces a lot of waste and harm.

Autonomous vehicles hold enormous promise to modernize our transportation system, and AI has the power to make them a reality. With deep learning, we can now perceive the environment surrounding the car. We can also reason about where your car is, where everything around the car is, and predict where everything will be in the near future, so we can plan a safe path forward. This is all supported by HD maps in the cloud, which we can update continuously as the world changes around us.

[Attachment]

All of this requires an enormous amount of AI computing power. We dedicated ourselves to building a fundamentally new computer to handle it. We call it an AI car supercomputer. Xavier is its latest edition. It runs a new operating system, DriveWorks, that takes in data from multiple sensors, fuses it, recognizes and perceives the world, localizes the car, reasons about its surroundings, connects to an HD map, and drives. It's incredibly powerful - with the performance of a high-end gaming PC shrunk into a tiny chip. Xavier can reach 30 trillion operations per second of deep learning in just 30 watts.

Here's a video that will give you a sense of what an AI car supercomputer can do:

We also announced new AI Co-Pilot capabilities. Now, even if the car's not driving for you, it can be looking out for you - alerting you to dangerous situations on the road as well as recognizing your identity, knowing where you're looking and reading your lips to understand your commands in noisy situations.

[Attachment]

At CES, we announced a series of new partnerships to build out our platform and accelerate the race to self-driving cars. With long-time partner Audi, we announced that we'll achieve Level 4 autonomous driving in production cars no later than 2020. We also announced that cars from Mercedes-Benz will incorporate products built with NVIDIA within a year. Bosch and ZF, the No. 1 and No. 5 tier 1 auto suppliers, respectively, have adopted the DRIVE PX car computer.

And we added ZENRIN and HERE to our recent announcements of Baidu and TomTom for HD mapping. The global HD mapping services adopted NVIDIA MapWorks for map processing and live map updates in the car. The DRIVE PX AI car supercomputer does continuous mapping. NVIDIA-powered cars can now achieve Level 4 self-driving throughout the world.

[Attachment]

NVIDIA - Shaping the Future

There's never been a more exciting time to be in the gaming, smart home and automotive industries. We're thrilled to be playing a vital role across them all. We want to bring video games to a billion people who currently don't have the computers necessary to enjoy them at their full potential. We want to turn your home into an AI. We believe that your home will engage you and you will engage your home in natural ways, and it will just make life better and more fun. And we would like to turn your car into an AI. By applying AI computing, we can revolutionize the automobile and bring joy and delight and safety to billions of people.

With 200,000 attendees and 3,800 companies, this year's CES was one for the record books. Its 50th anniversary was not only a milestone for the largest technology conference in the world, it marked a historic moment for the technology industry itself. With AI, we're now able to realize the dreams that we've had for so many years. I was honored to be a part of the moment.