\n<\/p>\n
Nuro<\/a>, the Silicon Valley-based startup backed by Nvidia<\/a>, Uber, and SoftBank, is testing its autonomous vehicle technology in Japan.<\/p>\n Toyota Prius vehicles equipped with Nuro\u2019s self-driving software \u2014 and human safety operators behind the wheel as backup \u2014 began testing on public roads in Tokyo last month. The testing marks the first overseas expansion for the startup, which upended its business model<\/a> two years ago.<\/p>\n Nuro said testing in Japan introduces a number of new challenges and different driving styles and rules. For instance, vehicles drive on the left side of the road, and Tokyo\u2019s streets have dense traffic. Road signs and lane markings are also different in Japan. The company, which opened offices in Tokyo last August, did not disclose how many test vehicles are in its fleet or when it might remove the human safety operator from the vehicles.<\/p>\n The company did suggest, in a blog post announcing the testing in Japan, that there will be future expansions. <\/p>\n \u201cOur autonomous operations in Tokyo are the beginning of the compounding benefits of global deployment,\u201d the company wrote.<\/p>\n Nuro, founded in 2016 by early Google self-driving project engineers Dave Ferguson and Jiajun Zhu, initially focused on developing and operating a fleet of low-speed, on-road delivery bots. Nuro\u2019s pitch and pedigree got the attention of SoftBank Vision Fund, which invested $940 million<\/a> into the startup in 2019.<\/p>\n Nuro had a buzzy start, but the cost of development and a wave of consolidation forced the company to cut staff and assess its business model. In 2024, it ditched the low-speed bots and decided to license its technology to automakers and mobility providers, like ride-hail and delivery companies.\u00a0<\/p>\n The company\u2019s autonomy stack is built on an end-to-end AI foundation model that allows the system to learn as it drives, according to Nuro. This AI strategy, which it calls \u201czero-shot autonomous driving,\u201d allowed Nuro\u2019s software to autonomously navigate public roads in Tokyo without any prior training on Japanese driving data, the company\u2019s blog post said. U.K.-based startup Wayve, which recently raised $1.2 billion<\/a>, has taken a similar end-to-end AI approach to its self-driving software.<\/p>\n Nuro says that this AI approach, which is designed to be broadly capable, doesn\u2019t mean it\u2019s disregarding safety. The company said that it conducts closed-course testing of each new release of its universal autonomy model and evaluates performance and tests edge cases using simulation. Once the autonomous vehicles are on the road, they are manually driven while Nuro\u2019s software operates in \u201cshadow mode.\u201d Nuro said the foundational AI model produces what the software would do, but the commands are not sent to vehicle controls.<\/p>\n Nuro checks the results to determine if the system is ready to operate autonomously on public roads. <\/p>\n Nuro has gained some traction and investors for its approach to self-driving software. Last year, Nuro raised $203 million<\/a> in two tranches in a Series E round that included existing backer Baillie Gifford and new investors Icehouse Ventures, Kindred Ventures, Nvidia, and Pledge Ventures. Uber, which has said it would make a \u201cmulti-hundred-million-dollar\u201d investment in Nuro as part of a\u00a0broader deal<\/a>\u00a0with the electric car maker Lucid, also participated.\u00a0<\/p>\n<\/div>\n