With all of the hype surrounding Tesla and autonomous driving, you’d think Tesla cars would be farther along than its competitors when it comes to a reliable solution, but that isn’t the case. The first consumer cars to be certified for Level 3 autonomous driving are Mercedes EQS and S models which, I believe, is directly related to Mercedes’ close partnership with NVIDIA.
Let’s talk about why NVIDIA is the critical partner for autonomous driving and why Mercedes was able to beat Tesla to market.
The Autonomous Driving Mess
Compared to companies that make advanced technology, automobile manufacturers tend to be insular and inbred in terms of their core engineering staff. I was once brought in as a tech expert for GM’s OnStar platform, and I was shocked at how bad the technology was given how advanced it should have been. This was because GM wasn’t leveraging what was in the market so much as trying to create something from scratch with little understanding of the tech. I left, realizing I couldn’t help them because they didn’t seem to want to be helped.
This is why Tesla was so successful. It was created using a company model that was more like a technology company and less like a car company. But as Tesla advanced, it too became insular. When it came to AI autonomous driving, it tended to favor internally developed technology rather than looking outside for something better. Tesla’s biggest problem was naming what it developed “Autopilot.” The term convinced drivers they had an actual autopilot when what they really had was simply a more advanced cruise control (Level 2+ system). While it was good, it was never meant to be hands free.
This resulted in a number of high-profile accidents and dead Tesla drivers, and caused people to begin to distrust the technology which will significantly lower demand for it.
NVIDIA
NVIDIA has been developing its autonomous driving technology for decades. Its Omniverse platform allows these autonomous driving systems to be virtually trained on most roads in every potential driving condition at computer speeds, effectively completing months of training in minutes to better prepare the system for almost any driving problem.
And given there is an audit trail through the training, if a problem does result, you can more easily prove it didn’t result from negligence. So, it not only helps you get to market more quickly with a more advanced solution, it also helps protect the company from mistakes that might have been made through the labor intensive processes of training through driving. That process relies heavily on human operators who too often aren’t paying attention which potentially results in training errors and an extended training time. That lengthens time-to-market.
Working with NVIDIA should provide the fastest time-to-market with the least back-end liability risk.
Mercedes’ Smart Move
Mercedes partnered with NVIDIA to advance its self-driving efforts. The partnership paid dividends by enabling the company to get certifications before anyone else. One side benefit for drivers of new Mercedes EQS and Model-S automobiles is that while the full Level 3 self-driving capability costs around $2,500, many of the safety benefits should be available to those who don’t opt in to the full autonomous driving service. This capability (that Toyota named Guardian Angel) should be in every autonomous car as a base feature, so drivers who can’t pay for the full service will still get one of the biggest benefits: more aggressive accident avoidance.
And Mercedes dealers get what amounts to a halo offering. People are likely to want to go to the dealer to check out this capability even if they can’t afford an EQS or Model-S. Because the self-driving technology is so much better in these cars it will interest those who have to drive in freeway traffic and want the car to handle the tedium for them.
Wrapping Up:
Mercedes has surprised Tesla by being the first to get certified for Level 3 autonomous driving in the United States. This was related to its partnership with NVIDIA which currently has the most mature and complete autonomous driving platform in the world. This certification should speed up certifications for other automotive OEMs using NVIDIA’s autonomous driving solution.
In the end, the success of self-driving cars will have a lot to do with the maturity and testing of the related technology. Right now, no platform is more mature or more tested than NVIDIA’s.
