dorsaVi Limited (ASX:DVL), a Melbourne-based medical-grade sensing and Data Analytics company, has announced it has begun engaging with robotics companies to explore applications for its established human movement data library and bespoke motion-capture capabilities. The initial focus is on humanoid robot balance, coordination and recovery — areas where dorsaVi's decade-long accumulation of real-world movement data may complement robotics developers' control architecture requirements. The engagement connects the Company's sensing expertise with its exclusively licensed neuromorphic Reflex Engine IP and an ongoing RRAM hardware development programme.

The Movement Data Resource

Over more than a decade, dorsaVi has accumulated an extensive library of movement recordings across clinical, workplace and sporting settings, collected through its ecosystem of FDA-cleared wearable sensor devices. The data library currently serves as a baseline for occupational health and safety studies. Following an internal review, the Company is presenting this resource to robotics developers to identify which datasets may be relevant to the specific movement and control challenges associated with humanoid robot development.

Alongside its existing library, dorsaVi is also capable of undertaking bespoke movement capture using its medical-grade wearable sensor technology and Video AI systems. This capability enables additional recordings to be collected around defined tasks, body segments and operating conditions. The Company's capture experience extends to animal applications, including equine biomechanics — a breadth of capture capability that may allow it to generate datasets across a wider range of physical movement scenarios than those covered by its existing human-focused library.

The Humanoid Balance Problem

dorsaVi's announcement articulates the specific robotics challenge it is targeting: while humanoid robots have achieved reliability on flat, predictable ground, they continue to struggle with dynamic corrections to unexpected perturbations — uneven surfaces, unexpected loads, slips, pushes or missed steps. The human neuromuscular system addresses these challenges through a rapid, distributed correction response spanning ankle, knee, hip and trunk, completed in a few hundred milliseconds and involving multiple joints acting in a temporally coordinated sequence.

For a machine to replicate this capability, three components are identified: real-world movement data showing how a stable human body detects and corrects instability; local processing that can respond across multiple joints without routing through a central processor; and low power consumption to allow continuous operation within battery constraints. This framing positions each of dorsaVi's three key Assets — movement data, neuromorphic IP and RRAM hardware — as addressing one of the three required components.

Neuromorphic IP and RRAM Hardware

dorsaVi holds an exclusively licensed Reflex Engine IP designed to support rapid local responses to sensor inputs — potentially addressing the local processing requirement identified in its robotics thesis. The Company intends to assess how its movement data can support the development and evaluation of models using this technology. Additionally, the NTU-ITRI programme is progressing 180-nm RRAM validation toward a future 22-nm platform. The Company intends to investigate the materials' analogue characteristics for selected Compute-in-Memory functions, including neuromorphic processing and physical artificial neural networks. Results from RRAM validation are intended to inform subsequent hardware evaluation and potential joint development with robotics companies.

Strategic and Commercial Context

dorsaVi's robotics engagement represents a strategic extension beyond its existing medical and occupational health sensor markets. The announcements frames the initiative as an integrated Physical AI strategy that connects established sensing and data capabilities, licensed computing IP and developing semiconductor hardware. However, investors should note that the robotics engagement is at an exploratory stage — discussions are underway to identify application requirements and potential technical evaluations, not binding commercial partnerships. The timeline from initial engagement to any Revenue contribution from this initiative is uncertain and subject to the pace of both dorsaVi's own IP development and the broader humanoid robotics industry's adoption curve.

Key Insights

dorsaVi's pivot toward humanoid robotics is strategically coherent in that it leverages existing movement data assets and licensed IP toward a high-growth adjacent market, without requiring new capital-intensive data collection infrastructure. The three-component framing — data, local processing, low-power hardware — provides a clear articulation of how dorsaVi's assets map to robotics developers' needs. The key uncertainty is the commercialisation pathway: engagement with robotics companies is in its early stages, the RRAM programme is progressing through validation milestones rather than production readiness, and the neuromorphic Reflex Engine IP has not yet been deployed in a commercial robotics application. The opportunity is conceptually well-constructed; execution risk and timeline remain significant variables.