NVIDIA and Faraday Future Press Unveil – CES 2016

6 PM Pacific and 8 PM Pacific on the 4th of Jan were supposed to be really important moments for the automotive industry. The first one was NVIDIA one of the fastest growing technology vendors for the automotive industry specifically in the ADAS and automated driving market unveiling its new product and the latter was the startup Faraday Future or FF as it calls itself unveiling its connected, automated and clean mobility solution for the future. Both were interesting in its own ways but the winner was NVIDIA and FF despite having its moments left a lot to be desired.

Let’s get to the specifics of both the press briefings

Faraday Future – FFZERO1 is not the Answer

Forget everything you know about cars, smarter cars, better used of parked cars, imagine the back seat as the new front seat and new generation of cars and ultimately the connected mobility vision for the future- these were phrases thrown all over the press event held by FF at CES in Las Vegas. The unveil turned out to be a disappointment in a few ways but some interesting things did come out such as

• Showcased largely by Nick Sampson, SVP R&D and Global Product Development, FF’s route to market is centered on 4 main areas – building a top class team, building top class alliances, aggressively pursuing a transformative vision and being fast to market. The entire press briefing was centered on talking to each of these 4 areas.
• As for the team part, FF has already assembled a team of 750 folks, 550 of whom are in the US. The interesting part is the diverse backgrounds these folks come from including the likes of Apple, Google, Tesla, BMW, NASA, Boeing, etc. Full marks to FF for assembling a rather interesting team that can definitely think out of the normal automotive box. Even Tesla when it grew did not acquire talent at this diversity; it chose to acquire talent from the conventional automotive world. Even a former NHTSA official is in FF’s payroll to make sure they are not crossing any driver distraction or safety boundaries.
• The transformative vision is where words were merely words. Despite throwing the idea that the company is not into making and selling cars (in this case EV’s) and is more into providing connected mobility options, nothing was disclosed on how this is going to be achieved and how this will differentiate FF from the biggest disruptor the auto market has seen – Tesla. The teaser videos that were previewed before CES emphasized a lot on new ownership models but nothing was showcased on this part.
• The alliance part is where the company is making rapid progress. On one side the announcement was made that the ground breaking for the NEVADA plant will happen very soon and more than 4500 jobs will be created through the $1.5 billion plus investment. The commitment is to bring its model/models to the market in the next couple of years. On the other side the LeTV alliance was mentioned many a times and this is where the possible internet connected vehicle ecosystem gets a strong push because of the background of LeTV and its position as the Netflix of China. With aspirations to sell its vehicles in China and gain a good market share in China, this alliance will go a long way and even help in US with a connected vision that goes beyond current imaginations.
• The fast factor was again mentioned many a times. Right from adopting digital design and engineering process to component testing digitally, FF is placing the right bets. The most interesting part was the announcement of the variable platform architecture (VPI) as FF calls it and proclaims to be better than the industry leading modular architectures (reference to VW’s MQB maybe). VPI allows FF to play around with the battery and wheelbase architecture (have a new battery placement called strings of batteries), motor and PT configurations, internet and autonomous pre-equipped thereby helping them to possibly create many different body styles and drivetrain architectures. Unfortunately if FF had gone a little deeper and unveiled possible battery combinations and range expectations, that would have helped them generate more positive buzz and gain an upper hand on Tesla (as a few pre CES articles proclaimed FF was going to bring).
• That brings me to the concept that was unveiled called FFZERO1. Classified into the fully BEV performance segment, the bottom line message we got is that this is probably not going to be the car that will be commercialized first. Nevertheless interesting concepts like inside out design philosophy, drive by wire instrument panel, zero gravity driving positions inspired by NASA and a whole lot of technology infused driving like gestures, AR HUD based copilots were discussed (not shown). Definitely the design though inspiring left a lot to be desired in terms of what the actual vision of the company is in terms of the product and solution it’s brining or transforming in the marketplace.

Bottom-line was that despite being dubbed as the biggest disruptor and a startup with transformative vision, little was showcased on both.

NVIDIA Showcase – Deep Learning and Artificial Intelligence All the Way

JHH the CEO of NVIDIA took the stage and the first and the only thing he talked about along with his colleagues was all things automotive especially automated driving. After showcasing the Drive PX module last year at the CES, this year NVIDIA showcased the Drive PX2 that is literally a supercomputer on wheels to support advanced levels of automated driving. The big challenge with Level 3 and above automated driving is not in highway driving but dynamic urban driving conditions where challenging scenarios are presented to the driver. Designed to support both Level 3 and 4 (NHTSA definitions), the Drive PX2 combines NVIDIA’s vision for software, supercomputers and deep learning. The Drive PX2 as NVIDIA claims can do about 24 trillion deep learning operations per sec (don’t ask me what that means) but simply put is a very fast machine that is capable of handling inputs from 12 cameras, LiDAR, Radar and Ultrasonic sensors (don’t know how many). The basic idea is that this giant supercomputer can differentiate between humans, cars, animals, signs, etc. and use that information in a real-time sensor fusion environment to help the car stay clear of any obstacles and on the safe course. The basic conclusions are the following

• NVIDIA is firmly establishing itself as a leader in the GPU-based deep learning space for automated driving. Maybe Mobileye will showcase something much more interesting later this week but with Audi, BMW, Daimler, Ford, Toyota and many others working on it and Volvo as a committed customer for its Drive PX2 platform, NVIDIA is firmly on a growing chart here. For two consecutive years in a row, NVIDIA has demonstrated its deep learning capabilities in this space at CES and no one else has come close.
• The Drive PX first gen solution cost was around $10,000 and it included input for 12 cameras and was based on two Tegra X1 processors. Though no pricing was mentioned for the PX2 module (availability for later 2016) it can be safely assumed that the cost will be much higher given the superhuman processing power that has been added and the advanced sensor input that it can use. That will make its usability to the wider market a big question beyond the premium OEMs. And given the fact that we believe that advanced automation features will be a convenience option for the customer, deep learning might very well be the technology answer that specific OEMs were willing to pay a price for.
• This growing expertise puts NVIDIA at a level that many Tier 1 suppliers will envy and am sure vendors like Mobileye (its business will not be directly affected but role will be diminished) will be watching this space very closely.