Five years ago, in a blogging competition about “what will be the most important technological development over the next 10 years that will have the greatest impact in reducing climate change risks,” I concluded that the answer was robotaxis. If true robotaxis, broadly available and deployed in cities around the world, come to fruition, the potential reduction in emissions is immense. This is assuming they are electrically powered, but that seems most sensible for several reasons — especially by the middle of the decade.
Naturally, many of us think that Tesla is quite far ahead in the development of broadly applicable, cost-competitive robotaxi hardware and firmware. However, it certainly isn’t the only name in town, and there are also many who think that Tesla’s approach cannot lead to true robotaxis. One other tech company you have to keep on the table of possibilities is NVIDIA. Aside from being a tech giant in other realms, one of the advantages NVIDIA has is that it supplies hardware — and increasingly software services — for a bunch of automakers. Also, as the industry has evolved, NVIDIA has looked more seriously at providing integrated, robust technology partnerships with these automakers — not just as a supplier, but as a team working with automakers’ driver-assist or self-driving teams.
With all of that in mind, NVIDIA has rolled out a series of 6 news stories in the past two months related to autonomous driving. In this article, I’m going through 3 of those that relate to the tech giant’s NVIDIA DRIVE solutions. Let’s catch up and check those out.
AutoX Robotaxis in Service Now
Probably the biggest story of the batch is that AutoX, a self-driving vehicle startup out of China, has launched its 5th generation robotaxi platform and the platform uses NVIDIA DRIVE. The system’s “automotive-grade GPUs to reach up to 2,200 trillion operations per second (TOPS) of AI compute performance.”
We did cover the rollout of AutoX robotaxis in January, when they launched to the public in Shenzhen, the 5th largest city in China (population over 12 million). It’s a solid testament to NVIDIA that a company with robotaxis on the road just upgraded to the new NVIDIA DRIVE platform. “Safety is key,” said Jianxiong Xiao, founder and CEO of AutoX. “We need higher processing performance for safe and scalable robotaxi operations. With NVIDIA DRIVE, we now have power for more redundancy in a form factor that is automotive grade and more compact.”
Even more impressive that this service is in place in the high-traffic, highly complex streets of Shenzhen. NVIDIA notes, “Safely navigating such chaotic streets requires sensors that can detect obstacles and other road users with the highest levels of accuracy. The Gen5 system relies on 28 automotive-grade camera sensors generating more than 200 million pixels per frame 360-degrees around the car. (For comparison, a single high-definition video frame contains about 2 million pixels.)” Mind blowing. “In addition to cameras, the robotaxi system includes six high-resolution lidar sensors that produce 15 million points per second and surround 4D radar.”
Now, Tesla fans will quickly point out that Tesla recently ditched radar because it basically just got in the way, and that Tesla is working to solve broad, general AI challenges. Nonetheless, let’s not miss the fact that NVIDIA DRIVE is being used in robotaxis that are in service right this moment in one of the largest and most traffic-heavy cities on Earth.
“At the center of the Gen5 system are two NVIDIA Ampere architecture GPUs that deliver 900 TOPS each for a truly level 4 autonomous, production platform. With this unprecedented level of AI compute at the core, Gen5 has enough performance to power ultra complex self-driving DNNs while maintaining the compute headroom for more advanced upgrades.
“This capability makes it possible for the vehicles to react to high-traffic situations — like dozens of motorcycles and scooters cutting in or riding the opposite way at the same time — in real time, and continually improving, learning how to manage new scenarios as they arise.”
See — other systems can learn, too.
AutoX is just getting started, with plans to roll out robotaxis in cities around the world and with large automotive partners like Honda and Stellantis. And NVIDIA is just getting started, as well.
NVIDIA Acquires DeepMap
To further improve its mapping solutions for aforementioned autonomous driving systems, NVIDIA announced in June that it was acquiring DeepMap. Clearly, there’s an implication of deep learning in that name — it’s all AI all the time these days. The summary highlights from that announcement: “DeepMap expected to extend NVIDIA mapping products, scale worldwide map operations and expand NVIDIA’s full-self driving expertise.”
“NVIDIA is an amazing, world-changing company that shares our vision to accelerate safe autonomy,” said James Wu, co-founder and CEO of DeepMap. “Joining forces with NVIDIA will allow our technology to scale more quickly and benefit more people sooner. We look forward to continuing our journey as part of the NVIDIA team.” DeepMap cofounders James Wu and Mark Wheeler previously worked at Google, Apple, and Baidu, so going back into a tech giant must feel a little bit like going home after getting DeepMap off the ground and acquired by NVIDIA.
What’s so special about DeepMap? Well, we don’t have an insight into the coding (and seeing it wouldn’t help me much anyway), but the key appears to be the crowdsourced data collection from a broad fleet of vehicles, which “lets DeepMap build a high-definition map that’s continuously updated as the car drives.” Naturally, the coding must be good, too. Getting integrated into NVIDIA DRIVE, it will certainly collect a lot more data and benefit from fast-growing deployment.
The acquisition hasn’t closed yet — going through all of the paperwork and lawyers necessary, it’s expected to close this quarter.
DiDi Goes Public, Also Benefiting From NVIDIA DRIVE
Gigantic Chinese ride-hailing company DiDi just went public about a month ago, raising a ginormous $4.4 billion. Not too shabby, but note that DiDi has nearly 500 million active users across 71 countries and 10,000 cities. NVIDIA took the moment to note that DiDi “is developing its upcoming robotaxi platform on NVIDIA DRIVE AGX Pegasus.”
The question is, who isn’t using NVIDIA DRIVE?