NVIDIA (NVDA +1.38%) has made amazing advancements across the tech world over the past several years, and the stock price has reflected that. In fact, I visited the company and helped compile a special report on my findings there about three years ago.
Since then, the share price has soared over 1,000%.
Today, I'm looking at the company from 36,000 feet ... heading home after yet another visit to Silicon Valley and NVIDIA's headquarters. And I ended up just as impressed this time as I was three years ago.
The company's GPU hardware and software technology have opened up multiple revenue streams in many industries from gaming to data centers and cryptocurrency mining. But it's artificial intelligence and self-driving cars that excite me most, and they're the focus of the interview below.
The key to making self-driving cars a reality
One of the big advantages NVIDIA's leading AI efforts provide -- and something not entirely obvious yet to many investors -- involves simulation.
Self-driving cars need to be tested in more wildly varied environments and traffic conditions than they're currently able to get in the real world. However, the company's NVIDIA AutoSim technology addresses this problem by allowing cars to be "driven" in a simulaton, racking up the equivalent of billions of miles on the road.
One of the experts we spoke to at the company this year was Danny Shapiro, senior director of the company's automotive efforts. In the video below -- shot at the company's new Silicon Valley headquarters in front of the "Robocar" race car -- Danny talks about self-driving progress broadly and how the incredible benefits of simulation are pushing us toward the day people can regularly ride in cars without drivers.
A full transcript follows the video.
Rex Moore: Let's start with maybe one or two things you're really looking forward to in the self-driving car space in 2018.
Danny Shapiro: We're seeing a huge advance in terms of the types of vehicles that are out there, from cars to trucks to delivery bots ... to even race cars like the Robocar you see here. And NVIDIA is developing technology that will enable these cars to drive autonomously. So that's our NVIDIA Drive that goes in the vehicle but also the data center ... you're seeing our GPUs, our DGX-based system for doing AI computing, for training the neural nets that help understand what's happening around the vehicles. But also using simulation to create environments that would be too dangerous to actually have in the real world. So we'll use simulation in the data center to train the neural nets as well as then validate our hardware. So it's a combination of hardware running in the data center and hardware running in the car together that are going to enable autonomous vehicles.
Moore: In one or two sentences, explain the huge benefits the simulation is providing.
Shapiro: Simulation basically recreates what would happen in the real world when we're driving. But if you drive a car on the freeway, say you're going 60 miles an hour, in one hour's time, we've captured 60 miles of road. In the data center, we can simulate 60,000 miles in just one hour. And so we can have all different types of scenarios, we can test all different types of algorithms much faster, and we can recreate scenarios that would be too difficult to capture in the real world like a blinding sun at sunset or sunrise. And we can simulate that on every road.
Moore: And then looking ahead again over the next year, some partnerships you're excited about with NVIDIA?
Shapiro: We have over 320 different partners of our NVIDIA Drive platform, so it's the car companies, truck companies, robotaxi companies, virtually all the mapping companies, the sensor companies, a lot of start-ups and research institutions that are developing the algorithms. So we're seeing this explosion in the number of developers that are building on our platform. That's going to translate into different types of vehicles being introduced in the near future on our roads. A lot of testing going on, a lot of deployments and pilot projects happening. And in several years then, more vehicles that are truly autonomous on our roads.
