Jan 12, 2018
They played the #Lowrider song when they brought out the autonomous car 😆 #ces2018 #Intel #Autonomouscar #mobileeye #car #sensors #future
seen from China
seen from United States
seen from Türkiye
seen from United States
seen from Egypt
seen from Australia
seen from Russia
seen from Taiwan
seen from China
seen from United Kingdom
seen from United Kingdom
seen from Malaysia
seen from Singapore
seen from Sweden
seen from United States
seen from United Kingdom
seen from Netherlands
seen from United States
seen from Australia
seen from Japan
They played the #Lowrider song when they brought out the autonomous car 😆 #ces2018 #Intel #Autonomouscar #mobileeye #car #sensors #future
Nvidia pushes computing to the edge
The future of computing is on the edge. Powerful, self-contained devices at the periphery of networks will make cloud computing obsolete, say venture capitalists.
Nvidia Corp. (NVDA) is not waiting around to see how it all plays out. Last week it announced a partnership with Bosch, the German automotive-parts behemoth. The two will build energy-sipping, next- generation supercomputers for the self-driving cars of the future.
It’s a market Nvidia knows well. Chief Executive Jen-Hsun Huang invested billions to develop deep-learning algorithms and sophisticated silicon before it was commercially viable. That vision was vindicated in Drive PX and a seat aboard every new Tesla (TSLA).
The Bosch deal ups the ante. It pushes Nvidia technology beyond upscale electric cars, into the mainstream. Bosch is expected to use Nvidia AI technology in the mass production of autonomous cars.
This pushes the envelope. The new supercomputers will feature Nvidia’s Xavier AI-on-a-chip architecture. The company claims it’s capable of 20 trillion operations per second. That kind of oomph on board negates the need to beam most of the complex computation for self-driving to the cloud.
Venture capitalists are taking note. Peter Levine, general partner at Andreessen Horowitz, makes the case that self-driving cars need to become “data centers on wheels”. They are screaming for computing power on the edge of the network because latency is a safety issue.
It makes sense. Being dependent on a network connection for navigation while hurtling down the road in 2 tons of glass and steel is not a comforting thought.
Levine argues that cloud computing is essentially the mainframe on steroids. What it makes up in power, it surrenders in flexibility. The normal ebb and flow is to bring compute closer to the client.
This is something Nvidia has obviously given some thought. Current Drive PX supercomputers are as powerful as 150 MacBook Pros. Their ability to render 3D maps is cutting-edge. Still, they are part of a system that relies heavily on server-side Drive PX components to make everything click with precision. In the always connected Tesla, this works.
While the Nvidia and Bosch partnership is based on Drive PX technology, the secret sauce is Xavier. By bringing the AI onboard, the most important bits of self-driving are self-contained, inside the vehicle and on the edge of the network.
Intel (INTC) made news last week with the acquisition of our old favorite, Mobileye (MBLY), an Israeli maker of assisted-driving cameras. Intel sees synergy between future self-driving cars and its vast data center microprocessor business. For the record, that is not the edge.
Nvidia has been a terrific stock because management recognized future trends early. That has not changed. I like management. I like the technology. Shares are still a buy on dips.
Click here to sign up for my free VIP newsletter on the intersection of business, technology and culture