The self-driving car industry has had a turbulent start, marked by costly missteps and repeated delays. Despite this history, technology suppliers, chipmakers such as Nvidia, and a number of automakers are once again betting on artificial intelligence and a growing network of partnerships to reignite progress.
Even so, many automakers remain cautious. Beyond worries about rising development costs and the ability to scale production, companies are questioning whether consumer demand will be strong enough to justify the enormous investment required.
Autonomous vehicles have the potential to transform transportation, but making them safe for everyday use on public roads has proven far more complex and expensive than initially anticipated.
While some companies, including Alphabet’s Waymo and Tesla, have chosen to pursue self-driving technology independently, traditional automakers such as General Motors and Ford Motor have stepped away from developing fully autonomous systems in-house.
At the CES technology show in Las Vegas this week, Amazon Web Services and German supplier Aumovio revealed a partnership aimed at supporting the commercial deployment of autonomous vehicles. Meanwhile, autonomous trucking firm Kodiak AI and Bosch announced plans to scale production of self-driving truck hardware and sensors. Nvidia also introduced its next-generation autonomous driving platform, which will power a robotaxi collaboration involving Lucid Group, Nuro, and Uber.
Using Nvidia-powered technology, Mercedes-Benz announced it will roll out a new advanced driver-assistance system in the United States later this year. The system will allow vehicles to operate autonomously on city streets under active driver supervision.
Artificial intelligence—the core driver of autonomous vehicle development—is also becoming a crucial tool for reducing costs. AI and generative AI are increasingly being used to accelerate testing, simulation, and validation processes.
According to Ozgur Tohumcu, general manager for automotive and manufacturing at Amazon Web Services, AI is acting as a major catalyst for the industry by enabling extensive development and validation work with far fewer resources than before.
Western automakers are also facing growing pressure from China, which is moving aggressively to lead in autonomous driving adoption. Recently, Chinese regulators approved two vehicles with Level 3 autonomous capability, allowing hands-off driving under specific conditions. The industry defines autonomy on a five-level scale, ranging from basic driver assistance at Level 1 to full autonomy at Level 5, where no human oversight is required.
Despite this momentum, Jochen Hanebeck, CEO of Infineon, warned against unrealistic expectations that fully self-driving cars will become widespread in the near future.
Instead of committing new capital to Level 5 autonomy, major automakers are prioritizing revenue-generating Level 2 driver-assistance systems. These technologies are already on the market but still require drivers to remain attentive at all times.
Hanebeck said he does not see a rapid surge toward full autonomy, describing expectations of widespread Level 5 adoption as overly optimistic.
Although several limited robotaxi launches have recently been announced across China, the United States, Europe, and the Middle East, scaling these services remains a challenge. Jeremy McClain, head of systems and software at Aumovio’s autonomous mobility division, noted that expanding coverage requires vast amounts of data, large vehicle fleets, and complex logistics—all of which drive up costs.
The self-driving sector has long been fueled by ambitious promises. In 2019, Tesla CEO Elon Musk predicted that the company would have one million self-driving cars on the road within a year. In reality, Tesla launched only a limited robotaxi pilot last year, several years after that forecast.
One of the industry’s biggest challenges is the sheer number of unpredictable scenarios, known as “edge cases,” that autonomous systems must handle. For example, a human driver may slow down when seeing a ball roll into the street, anticipating a child could follow. A self-driving vehicle, however, may not react until the child becomes visible.
After the initial wave of enthusiasm faded, automakers such as Ford and GM shut down loss-making autonomous divisions. GM’s Cruise unit, in particular, faced setbacks after a serious incident in which a pedestrian was struck and dragged.
Still, Ali Kani, general manager of Nvidia’s automotive team, said advances in AI are helping to address some of the most persistent weaknesses in self-driving technology. He noted that recent breakthroughs have brought the industry closer to practical deployment.
Analysts at Morgan Stanley said Nvidia’s new Alpamayo autonomous driving platform could give traditional automakers a competitive edge and help them challenge Tesla’s lead, although Tesla remains years ahead in real-world deployment. Nvidia’s decision to make its platform open-source has also positioned it as a common foundation for automakers seeking alternatives to Tesla’s proprietary system.
Former Zoox product lead Russell Ong compared the situation to the rivalry between Apple and Android, with Tesla pursuing a closed ecosystem while Nvidia offers an open platform for the broader industry.
