At the Automobility conference in Los Angeles ahead of the city’s historic auto show this week, tech and automotive industry leaders are convening to share ideas on the future of transportation and mobility. Intel CEO Brian Krzanich, in a Keynote address to the conference on Tuesday morning entitled “Data is the New Oil,” presented his company’s view that for the car of the future, access to massive datasets will be just as imperative to propulsion as the fuel sources—namely oil—that cars currently depend on.
“The car of tomorrow will basically be a data center on wheels, capturing and processing terabytes of data.”
“The car of tomorrow will basically be a data center on wheels, capturing and processing terabytes of data,” Krzanich said, stressing the staggering volumes of information that will be both generated and processed by autonomous vehicles. “One million autonomous cars will generate the equivalent of 3 billion people,” he continued. “It’s hard to communicate just how immense this amount of data is.”
Krzanich, who leads one of the most influential tech companies in the country, stated that by 2020 the average internet user will use approximately 1.5 gigabytes of traffic each day. By contrast, each autonomous vehicle on the road will use a mammoth 4,000 gigabytes of data per day. Autonomous vehicles data needs will come in part from cameras, which use 20-40 megabytes per second alone, compounded by the significant data needs of radar, sonar, GPS, and lidar.
“Without data, an autonomous car won’t move. You will need to have data as much as you have any kind of propulsion,” Krzanich said. That includes technical data about the environment surrounding the car, societal and crowdsourced data which feed apps and other internal resource, and personal data, which includes user information about how many miles people travel, where they go, and why. “Each one of these data sets provides different opportunities and different assets as we move through this new world. It’s not enough to capture this data—you have to turn it into an actual set of insights.”
Data dependence to bring unique set of challenges
We don’t yet know if the car of the future will run on electricity, petroleum, or entirely on another fuel source. But the transportation sector’s current dependence on oil exacts an enormous toll on the economy and creates a range of challenges. Ninety two percent of transportation in the United States is fueled by petroleum. As a result, when oil supply disruptions or shortages cause a price spike, businesses and consumers are left without options or alternatives—they are simply forced to pay more for fuel.
Data won’t be subject to the kind of geopolitical instability, price shocks, or environmental impacts that oil entails. But the technical challenges of scaling the world’s data infrastructure to accommodate autonomous vehicles could still result in disruptions or outages that limit mobility in the world of driverless cars.
Data won’t be subject to the kind of geopolitical instability, price shocks, or environmental impacts that oil entails. But the technical challenges of scaling the world’s data infrastructure to accommodate the rapidly growing needs of autonomous vehicles could still result in disruptions or outages that limit mobility in the world of driverless cars. The challenges we face with our current laptops and cellphones—overburdened WiFi and cellular data networks that can’t send or receive data—could now translate to interruptions in traffic flow, a stranded vehicle, or other malfunctions. Moreover, the same way that cellphones and laptops are effectively useless when not connected to internet or cell networks, the car of the future will be similarly compromised when it can’t access the necessary data streams.
Additionally, much like with oil, tremendous power will be held by those who control all this data. Krzanich argues that, like oil, an entire industry will emerge around providing support for the capturing, holding, processing, and delivering of this data.
“We are committed to helping this ecosystem deliver on an autonomous future, but this isn’t a done deal,” he said. “We have four areas we see as challenges. The first is the size of the data sets: A single car can generate terabytes of data in 8 hours of driving. Next is speed: we need to be able to turn software that now takes weeks to do the same thing in a matter of hours. Third is security: The car is a special place we want to make sure is secure. We believe industry collaboration is key—no one company can provide all the computing solutions. Last is accessibility and scalability: This is about machine learning and artificial intelligence, but you have to do this through bigger and bigger data sets.”
In Intel’s view, industry will have to share and collaborate as much as possible if it wishes to address these challenges, understand how every car is using the best possible data, and update the decision matrix to ensure people are safe. Companies will still compete on user interfaces and eliminate redundancies, but the issue of user safety is something that everyone, collaboratively, will need to get behind. He concluded: “This open platform will enable much greater innovation across the entire industry.”