Self-driving startup Turing secured ¥12.62bn ($78m) on 6 July in the extension round of its Series A funding. In total, the Series A raised ¥27.89bn ($172m).
The company was founded in 2021 and is headquartered in Tokyo.
Public-road driving with a VLA announced in March 2026
The mainstream approach to self-driving development has long been the so-called rule-based method: using LiDAR, a laser-based sensor, to measure the distance to surrounding objects, and then driving according to logical rules set in advance for how those objects will move and how the vehicle itself should respond. This approach was pursued on the assumption that the high cost of LiDAR would come down as self-driving spread.
In practice, however, LiDAR prices did not fall as hoped, and between 2023 and 2024 the leading self-driving developers switched from rule-based systems to End-to-End (E2E) approaches, or moved to combine the two (incidentally, LiDAR prices have been trending downward since the start of 2026).
E2E self-driving is, put simply, an AI-based approach: the AI judges the situation from data captured by cameras fitted to the vehicle and operates it accordingly.
In 2024, Turing’s founder and CEO Issei Yamamoto had likewise argued that fully autonomous driving was impossible with rule-based systems, and the company had been developing self-driving technology built on E2E. Then, in March 2026, it announced that it had driven autonomously on public roads using technology based on a VLA (Vision-Language-Action) model. A VLA is a form of physical AI for self-driving that operates a vehicle by combining camera data with a language-based understanding of the situation; it also falls within the broader concept of E2E.
Below is the video of the VLA-based self-driving released in March. It opens with the vehicle driving inside what appears to be a driving school before moving on to footage on public roads.
The aforementioned CEO Yamamoto is also one of the developers of the computer shogi program “Ponanza.” Ponanza competed in the Den’ō-sen, a shogi tournament that was a popular fixture of Japanese internet culture in the 2010s, and in 2013 it defeated the professional player Shin-ichi Sato, then a 4-dan (now 6-dan). It kept winning thereafter, and in 2017 it pulled off the remarkable feat of beating Amahiko Sato, now a 9-dan, who at the time held the rank of Meijin — one of the highest of the professional shogi titles.
AMD, Super Micro and others take part
The equity portion of the Series A extension was taken up by AMD’s corporate venture capital arm, AMD Ventures, Super Micro Computer and Tokyo Electron Device, among others. The debt portion was provided by MUFG Bank.
The company will use the funds to “expand its computing infrastructure, strengthen the business structure for real-world deployment, and hire the talent to accelerate those efforts.”
Turing’s CEO Yamamoto commented: “Completing this Series A round allows us to accelerate our challenge even further. Together with our partners in computing infrastructure, semiconductors and data centres, we will advance our technology development another step and do everything we can to bring fully autonomous driving, made in Japan, to society.”
