April 26, 2026 – Beijing, China (Auto China 2026) — During the 19th Beijing International Automotive Exhibition (Auto China 2026), ThunderSoft today announced the launch of AquaClaw, an in-vehicle AI agent operating system built on the NVIDIA NemoClaw reference stack, currently in early preview.
AquaClaw serves as the core runtime foundation for in-vehicle multi-agent AI systems, enabling seamless integration between intelligent agents and automotive use cases through a standardized, plug-and-play architecture. Designed for production scale, AquaClaw supports both real-time interactions and long-running, stateful agents that persist across driving sessions—unlocking continuous, context-aware experiences for drivers and passengers. By bridging development and deployment, AquaClaw accelerates the transition from pilot programs to commercial-grade implementations, establishing a foundational platform for the industry’s shift toward AI-defined, agent-driven in-vehicle experiences.
Addressing the Industry’s Shift to AI-Defined Vehicles
AI agents are rapidly emerging as a central engine driving the transformation of intelligent vehicles. As the automotive industry evolves from function-defined to AI-defined paradigms, vehicles are becoming one of the most critical environments for large-scale AI deployment.
This shift is driving urgent demand for an in-vehicle AI operating system (AIOS) that is secure, efficient, and highly adaptable—while being production-ready and capable of supporting core agentic AI functionalities at scale. Such a platform must accommodate diverse OEM requirements across vehicle lines and use cases, while significantly reducing development complexity and accelerating time-to-market. AquaClaw is purpose-built to meet these needs. The expanded collaboration between ThunderSoft and NVIDIA underscores a shared strategic commitment to solving critical industry challenges.
A Unified AIOS Architecture: Edge AI, Cloud Collaboration, and Built-in Safety
AquaClaw is built on a unified architecture that combines edge-first AI, cloud collaboration, and safety-by-design principles to deliver a full-stack in-vehicle AI operating system (AIOS). This architecture spans the base OS, AI runtime, and scenario-level adaptation—enabling scalable deployment of agentic AI across diverse vehicle platforms.
Its core capabilities are driven by the integration of ThunderSoft and NVIDIA technologies across two key domains:
ThunderSoft brings deep expertise in automotive edge computing and vehicle-grade AIOS, providing a customized base operating system, domain-specific ontology, and an intelligent vehicle skills library. These capabilities enable robust adaptation to complex in-vehicle scenarios while delivering low-power, automotive-grade performance at the edge.
NVIDIA contributes its industry-leading AI and agentic software stack, including the NVIDIA Agent Toolkit and NemoClaw, an open-source reference stack that installs OpenClaw in an NVIDIA OpenShell sandboxed environment with open models including NVIDIA Nemotron 3. This is complemented by NVIDIA TensorRT Edge-LLM for high-performance inference and the NVIDIA DRIVE AGX in-vehicle computing platform. Together, these technologies enable more secure, compliant, and low-latency execution of AI agents, delivering fast response times and stable, operation.
Lowering Barriers, Improving Efficiency, Ensuring Safety
At its core, AquaClaw delivers three key values: reduced development barriers, improved efficiency, and safety.
For automakers, AquaClaw eliminates the need to build an in-vehicle AIOS from the ground up. OEMs can rapidly integrate AI agent capabilities, tailor them to specific vehicle models and use cases, and significantly reduce R&D investment and time-to-production.
For end users, AquaClaw enables a new generation of intelligent, seamless in-vehicle experiences, empowering in-cabin AI control, driver assistance, and contextual interaction. Vehicles evolve into interactive, adaptive, and continuously learning AI-powered terminals.
Driving the Next Phase of Automotive AI
“AI agents are becoming a fundamental force reshaping both the automotive ecosystem and user experience,” said Larry Geng, Co-founder and Executive President of ThunderSoft. “Together with NVIDIA, we developed AquaClaw as a plug-and-play, full-stack solution for automakers. Our goal is to reduce development costs, accelerate mass production, and enable our partners to lead in the global intelligent vehicle market—delivering both commercial and user value at scale.”
“This collaboration brings together the best of both companies’ technologies,” said Rishi Dhall, Vice President of Automotive at NVIDIA. “By combining ThunderSoft’s deep expertise in automotive system customization with NVIDIA’s advanced AI computing architecture, we are creating a secure, efficient, and scalable pathway for deploying AI agents in vehicles -- delivering truly autonomous and adaptive mobility experiences for consumers worldwide.”
This initiative represents not only a powerful integration of complementary technologies, but also a coordinated step forward in advancing the in-vehicle AI ecosystem. Looking ahead, ThunderSoft will continue to invest in AquaClaw-driven innovation and ecosystem development leveraging NVIDIA technologies to accelerate the large-scale deployment of in-vehicle AI agents.
About ThunderSoft
ThunderSoft is a world-leading provider in AIOS products and technologies. Built around its core AIOS platform, the company focuses on cutting-edge on-device AI technologies and develops a full-stack AI operating system engine. It provides product and technology enablement for smart vehicles, intelligent hardware, software, robotics, and industry-wide digital transformation. By empowering industries with next-generation AI capabilities, ThunderSoft accelerates innovation and drives the global shift toward an intelligent, connected future. Website: https://en.thundersoft.com/
