Autonomous Driving AI: Companies, Startups & Research

Waymo LLC, Nvidia, Uber and Tesla Motors are among the active companies in autonomous driving AI. Waymo develops an integrated self-driving system and operates driverless ride-hailing services; Nvidia provides models and tools for developers; Uber works with autonomous-driving partners to offer rides. Automakers including Toyota Motor Corporation, Hyundai Motor Company and General Motors also develop automated-driving capabilities.

Wayve Ltd. stands out among the listed startups, with 165.6% growth in its recent document activity and a public robotaxi launch with Uber in London in September 2026. It develops end-to-end driving software designed to generalize across roads without relying on HD maps. May Mobility focuses on autonomous microtransit, while Kodiak Robotics develops self-driving technology for long-haul trucks.

Tsinghua University, Tongji University and Southeast University are prominent research organizations in this field. Tsinghua work includes uncertainty-aware accident anticipation and risk reasoning; Tongji studies pedestrian behavior and interaction-aware decisions; and Southeast University researchers work on lane detection and physics-informed reinforcement learning. Beihang University also appears in recent work on trajectory prediction and planning.

FACTS records $1.49 billion in grants since 2020, across 1,189 grants. It also lists $602.8 billion across 56 funding rounds, but those entries include broader corporate and project financing, so that figure is not a clean measure of investment specifically in autonomous-driving AI. Separately, Wayve announced a $1.2 billion funding round in February 2026.

Hyundai Motor Company leads the listed patent holders with 220 patents, followed by Toyota Motor Corporation with 193 and Waymo LLC with 146. Shenzhen Yinwang Intelligent Technologies Co., Ltd. has 123, and Kia Corporation has 118. These are the holder counts provided in FACTS, rather than a measure of patents filed in a single recent year.

No. Patent activity declined 48.6% in the last 12 months compared with the previous 12 months, according to FACTS. In contrast, publications in the last year grew 2.4%. The difference suggests that research output and patenting are moving at different rates; it does not by itself establish why activity changed or predict future commercial progress.

Recent developments include Uber and Wayve launching a public robotaxi service in London in September 2026, with a safety operator in the vehicle. Research is also examining the move toward large end-to-end driving models and whether explainable AI can help people understand self-driving decisions. Edge-case detection, uncertainty-aware decision-making and reliable prediction remain active technical challenges.

The main applications listed in FACTS are passenger vehicles and driver assistance, robotaxis and ride-hailing, trucking and freight, delivery and logistics, and industrial and off-road vehicles. Robotaxis and ride-hailing are the application area showing growth, at 6.1%; the other listed application categories show declines in recent activity. Each use case brings different operating conditions and validation needs.

The main approaches are perception and sensor fusion, localization and mapping, behavior and trajectory prediction, planning and decision-making, vehicle control, and end-to-end driving models. Modular approaches separate these functions, making individual components easier to inspect but requiring reliable handoffs. End-to-end models learn across a connected pipeline, but their decisions can be harder to interpret and validate.