Participated · Equity · May 2026

SynapX builds a system architecture called SYNTH that links three domains—SYNAction (manipulation intelligence), SYNWorld (physical-world modeling), and SYNData (data systems)—to enable robots to execute complex operation tasks more efficiently and evolve autonomously. Its Omni-modal Physical Data System (OPDS) combines custom hardware design and AI algorithms to collect and generate high-quality multimodal physical signals such as vision, force, and touch at scale to continuously improve base models. The company plans to invest its recent $50 million raise into core technology and product R&D, construction of its multimodal data system, and hiring global talent. Founded in January 2026 by Tsinghua alumnus Du Dalong, SynapX positions itself as an emerging player in physical AI focused on large-scale data-driven model evolution. The article did not report revenue, user metrics, or a disclosed valuation.

Participated · Equity · Mar 2026

Octopus Dynamics is focused on building a unified system for embodied intelligence aimed at physical AGI and has proposed the SYNTH Deep Thought architecture as its core technology framework. SYNTH encompasses three pillars: SYNAction (an end-to-end manipulation architecture called REMA that layers high-, mid- and low-frequency control), SYNWorld (a Vision–Force–Tactile World Foundation Model integrating visual, force and tactile modalities) and SYNData (an Omni-modal Physical Data System, OPDS, for scalable multi-modal data capture). The company intends to use its recent financing to accelerate R&D, construct its data system and expand technical hiring. Founded in January 2026 by Du Dalong, the team includes researchers and engineers from Horizon Robotics, Baidu, Microsoft, Alibaba, ByteDance, NIO, XPeng and Li Auto. Leadership highlights include Du Dalong’s Tsinghua background, his role as an early employee at Horizon Robotics, and a prior co-founded company, Jianzhi Robotics, which was acquired at the end of 2025. The company emphasizes end-to-end execution cycles in real environments and claims system-level integration of model architecture, world modeling and data as central to achieving embodied intelligence.

Led · Series B · Mar 2026

KargoBot focuses on open-road, trunk-line autonomous freight, offering a full-stack solution that combines driverless ‘Robotrucks’ with a logistics operating system. Its KargoOmni platform uses mixed-platooning and single-vehicle intelligence to raise human efficiency by 83–300%, overall operating efficiency by more than 50%, and energy efficiency by about 30%. The company has already deployed more than 400 autonomous trucks across 10+ provinces, accumulating over 35 million kilometers and 1.2 billion ton-kilometers of freight. Seven mass-production truck models are in service, supporting over 20 enterprise customers and a 30-member partner alliance. Management plans to enable and deploy 1,000 autonomous trucks within 2024 and scale to tens of thousands of Robotrucks in the next few years. KargoBot has proven a profitable commercial closed loop on trunk routes and is positioned as a first-mover in China’s trillion-RMB long-haul logistics market. The new capital from its Series B round exceeds $100 million and will fund large-scale rollout and continued R&D.

Participated · Series B · May 2024

PhiGent Robotics builds next-generation autonomous-driving systems with 3D visual understanding as its core. The company emphasizes general-purpose, real-world AI and differentiates through binocular stereo vision. It offers a full-stack autonomous-driving product suite positioned for strong cost-performance. PhiGent targets global automakers and automotive parts suppliers as customers. The company recently completed a $30 million Pre-B financing.