Pony.ai unveils next-generation autonomous driving controller based on NVIDIA… Will it accelerate the expansion of robotaxi services

Pony.ai ($PONY) has released a new generation of autonomous driving domain controllers based on NVIDIA’s latest autonomous driving platform. This move targets the autonomous mobility market, including L4-level autonomous taxis.

The platform released this time is based on NVIDIA DRIVE Hyperion, with the core equipped with DRIVE AGX Thor and NVLink. The company states that this platform supports not only single-chip configurations but also multi-chip setups, achieving a maximum computing power of 4000 FP4 TFLOPS when combined. This is interpreted as being specifically designed to meet the high-performance computing needs of L4 autonomous vehicles to process driving decisions in complex urban environments in real time.

Pony.ai claims that this domain controller improves energy efficiency and supports redundancy of key components, multiple cooling methods, and deployment options. For services like autonomous taxis that require long operational hours, energy efficiency and system stability directly impact profitability. This emphasizes that the product is not solely focused on performance competition but also fully considers real-world operational environments.

China market shipments have surged… while also emphasizing improved profitability.

The company states that its existing “Fangzai” controllers have seen shipments grow by over 500% compared to the previous year by 2025. This indicates that demand for autonomous driving hardware is rapidly expanding and also shows that Pony.ai has moved beyond pure technological development into mass production and commercialization.

Notably, Pony.ai announced that it has achieved break-even in “per-vehicle economics” in two major markets in China. Per-vehicle economics is an indicator of profitability per vehicle or service unit and is regarded as an important standard for assessing commercialization potential in the autonomous driving industry. While many companies have received large-scale investments but continue to operate at a loss, the fact that this company can achieve break-even in some regions is enough to attract market attention.

Goal: Operate 3,000 autonomous taxis by the end of 2026, covering 20 cities.

Pony.ai has set a target to operate over 3,000 autonomous taxis by the end of 2026 and expand services to more than 20 cities. This indicates that their plan is not just about technological demonstration but truly integrating autonomous driving services into urban transportation networks.

Looking at the entire autonomous driving industry, recent competition has shifted from “who has mastered the technology first” to “who can operate more stably and at larger scale.” Against this backdrop, the significance of Pony.ai’s release lies in showcasing both a high-performance semiconductor platform and commercialization metrics. The key in the future will be how to effectively combine high-performance controllers with vehicle scale expansion and profitability improvement.

TP AI Notice: This article is summarized based on the TokenPost.ai language model. The main content in the text may be omitted or may not be entirely accurate.