dorsaVi Ltd (ASX: DVL), an Australian wearable technology company, has unveiled its Sensor v6.5, marking the commercial launch of its first Physical AI product that performs movement analysis and delivers biofeedback directly on the device in real time without requiring cloud connectivity or a base device. The company update represents a significant step in dorsaVi's Ultra-Edge commercialisation strategy, transforming its FDA-cleared, TGA-certified movement measurement platform into an intelligent system capable of sensing, analysing, deciding and acting on human movement data at the point of contact. The announcement underscores dorsaVi's position as an early mover in edge artificial intelligence applied to wearable sensors for clinical, sporting, defence and occupational health applications.

Key Points

  • dorsaVi Ltd (ASX: DVL) is an Australian wearable technology company operating in the physical AI and movement analysis sector
  • The company has completed Sensor v6.5, its first commercially-ready Physical AI wearable that processes movement data and delivers real-time biofeedback on the device itself
  • Sensor v6.5 operates without requiring cloud connectivity, iPad or PC base devices, and performs measurement, analysis, decision-making and action within milliseconds
  • The platform builds on dorsaVi's decade-long FDA-cleared and TGA-certified movement analysis technology used across clinical, elite sport, defence and occupational health sectors
  • Future sensor generations are planned to integrate RRAM and compute-in-memory technologies as part of the Ultra-Edge Modular Design and Build program, with potential advancement toward always-on operation
  • The company's wearables have been deployed by leading organisations across military aviation, sport, resources, automotive, heavy industry, construction, transport, retail and healthcare

Physical AI Processing at the Edge: Sensor v6.5 Moves Intelligence Away From Cloud Infrastructure

dorsaVi's Sensor v6.5 represents a fundamental shift in how artificial intelligence can be deployed on body-worn devices. Rather than transmitting movement data to cloud servers for analysis and then waiting for feedback to return—a process that introduces latency measured in minutes or hours—the v6.5 sensor performs the complete intelligence loop on the device itself. This approach eliminates the bandwidth, latency and power costs traditionally associated with cloud-dependent wearables, enabling real-time intervention at the moment a movement event occurs.

The company positioning Sensor v6.5 within the broader emerging category of Physical AI, which it describes as systems that sense the physical world, interpret events, make automated decisions and act in real time at the point where the event occurs. dorsaVi notes that most deployed artificial intelligence currently operates in data centres, generating text, images and answers remotely from where physical events happen. The next technological wave—AI that senses and responds to physical movement in real time—is expected to emerge first in autonomous vehicles and robots, which can support the necessary power and computing budgets. Body-worn devices have presented a harder engineering challenge, as they cannot carry equivalent power or computational capacity. Sensor v6.5 signals that dorsaVi has solved this constraint for its movement-analysis application.

FDA Clearance and TGA Certification Underpin Commercialisation of Regulated Wearable Technology

dorsaVi has built Sensor v6.5 on a regulatory foundation established over more than a decade of clinical and operational deployment. The underlying platform holds FDA clearance and TGA certification, indicating it has met stringent regulatory standards for medical device and therapeutic goods classification in key international markets. This regulatory status is not incidental to the Sensor v6.5 launch; it is foundational. Rather than requiring a new regulatory pathway for the intelligent version, dorsaVi has extended its existing FDA-cleared, TGA-certified platform to include on-device artificial intelligence capabilities, effectively bringing Physical AI to market within an already-approved regulatory envelope.

The company's decade-long deployment history across clinical, elite sporting, defence and occupational health applications has generated both regulatory confidence and real-world validation data. dorsaVi notes that its wearables have been used by leading organisations spanning military aviation, sport, resources, automotive, heavy industry, construction, transport, retail and healthcare sectors. This installed base and operational history provide evidence of both platform reliability and market applicability. Sensor v6.5 therefore enters commercialisation not as an experimental technology, but as an evolution of a proven, regulated system already trusted across demanding operational environments.

Real-Time Biofeedback Architecture: Distinguishing Movement Quality and Injury Risk on the Device

The core capability of Sensor v6.5 is its ability to assess movement quality directly on the wearable and deliver configurable biofeedback within milliseconds. The sensor performs three analytical functions: it distinguishes normal movement patterns from elite performance characteristics, identifies patterns associated with elevated injury risk, and delivers haptic, audible or visual feedback to the wearer in real time. This feedback loop operates at the moment the movement event occurs, theoretically enabling intervention or correction before harmful loading patterns establish themselves.

dorsaVi emphasizes that this real-time feedback principle is not theoretical but clinically validated. The company references a randomized controlled trial (P Kent, 2015) in chronic low back pain demonstrating that feedback delivered at the moment of a movement fault improves patient outcomes. Sensor v6.5 embeds this clinically proven principle directly into the wearable's firmware and decision-making architecture. The sensor's configurable feedback modes—haptic (vibration), audible (tone or speech) and visual (light or display)—allow different applications and users to receive information in the modality most suitable to their operational context, whether that is an athlete on the field, a worker on a production floor or a patient in clinical rehabilitation.

RRAM and Compute-in-Memory Roadmap: Path Toward Always-On, Ultra-Low-Power Operation

While Sensor v6.5 operates on conventional silicon, dorsaVi has architected the device with a technology roadmap clearly defined. The company's Ultra-Edge Modular Design and Build program, announced on 6 May 2026, incorporates emerging memory and computing technologies—specifically resistive random-access memory (RRAM) and compute-in-memory architectures—designed to push future sensor generations toward always-on operation within the power envelope of a coin-cell battery. These are not speculative technologies; they represent the direction that edge artificial intelligence development is moving across robotics, autonomous systems and industrial computing applications.

RRAM and compute-in-memory technologies operate fundamentally differently from conventional semiconductor architectures. Rather than shuttling data between separate memory and processing units, these technologies enable computation to occur directly within memory itself, dramatically reducing energy consumption per decision cycle. Neuromorphic processing approaches, also part of the roadmap, further optimize power efficiency by mimicking biological neural processing. The announced program positions dorsaVi to integrate these technologies into future sensor generations as they mature and become commercially available, progressively extending the duration for which Sensor v6.5's successors can operate on minimal power without cloud connectivity or battery replacement.

Movement Measurement and Response Loop: From Millisecond Detection to Intervention Timing

The functional architecture of Sensor v6.5 operates as an integrated four-step cycle: measure, analyse, decide and act. Raw movement data is captured on the body in real time using the sensor's inertial measurement capabilities. Rather than transmitting this raw data to external processing infrastructure, the sensor performs real-time processing and interpretation on the device itself, achieving analysed output with no cloud round-trip and no latency penalty. The decide step follows immediately, with onboard assessment algorithms evaluating whether the detected movement pattern falls within normal operating parameters, represents elite performance characteristics or indicates elevated injury risk.

The final act step delivers the intervention. dorsaVi notes the timing is critical: biofeedback is configured to reach the wearer within milliseconds of a movement event, at the moment when intervention can theoretically change the outcome. The company provides practical examples without explicit quantification: whether an athlete displays excessive knee valgus (inward collapse) on landing, a runner crosses a safe impact-loading threshold, or a soldier begins to fatigue, the Sensor v6.5 system is designed to detect and signal these events in real time. This represents a departure from conventional wearables, which typically log data for later review by clinicians or coaches rather than enabling immediate intervention during the movement event itself.

Addressing the Body-Worn Power and Compute Constraint: Why On-Device Intelligence Has Been Difficult

dorsaVi articulates a key technological constraint that has historically prevented advanced artificial intelligence from operating reliably on body-worn devices. While the feasibility of Physical AI in vehicles, robots and industrial systems has long been established—these platforms can support large batteries, substantial processors and continuous power supplies—the same approach is not viable for wearables. A body-worn device operates under severe power, size and compute budget constraints. A coin-cell battery provides finite energy. Form factor must remain practical for extended wear. Processing power is limited by the thermal and power budgets of miniaturized silicon.

Until Sensor v6.5, the open question in wearable artificial intelligence has not been whether Physical AI works in principle—the concept is sound and proven in larger systems—but whether it can operate reliably on the body, run continuously without external wifi or cloud support, function in remote locations without network connectivity and stay within the power envelope of a wearable battery. This represents fundamentally a memory and compute efficiency problem. Sensor v6.5 demonstrates that dorsaVi has achieved sufficient algorithmic and architectural efficiency to run the complete intelligence loop on conventional silicon within these constraints. The company's roadmap toward RRAM and compute-in-memory integration signals an intent to push this efficiency boundary further in future generations, extending operating duration and reducing dependency on battery replacement.

Market Applications Across Clinical, Occupational and Defence Sectors

dorsaVi's platform deployment history reveals a diverse addressable market spanning clinical rehabilitation, elite sport, occupational health and defence applications. The company's wearables have already achieved adoption across military aviation, where pilot fatigue and postural stress are safety-critical factors, and in sport, where movement analysis informs both elite performance coaching and injury prevention. The technology has also gained deployment in resources extraction, automotive manufacturing, heavy industry, construction, transport logistics and retail operations—sectors where worker safety, injury prevention and fatigue management directly impact operational efficiency and health outcomes.

Each of these application domains stands to benefit from on-device intelligence that detects movement patterns associated with injury risk and delivers real-time feedback. Clinical rehabilitation environments can provide immediate feedback to patients executing therapeutic movements, potentially improving compliance and outcome velocity. Occupational safety applications can alert workers to fatigue-related postural degradation before injury occurs. Defence operations benefit from continuous monitoring of personnel without dependency on communication infrastructure. The breadth of this deployed base suggests that Sensor v6.5 enters an established market with proven use cases and existing customer relationships, rather than creating demand from a novel category.

Clinically Validated Feedback Principle: Outcomes Research Supporting Real-Time Intervention

dorsaVi grounds Sensor v6.5's design principle in published clinical evidence. The company references a randomized controlled trial by P Kent (2015) examining chronic low back pain outcomes, which demonstrated that biofeedback delivered at the moment of a movement fault—in real time, as the movement occurs—produces superior clinical outcomes compared to conventional delayed feedback or observation-only control conditions. This is not an assumption about what physical feedback might achieve; it is an empirically demonstrated clinical principle. Sensor v6.5's architecture directly embeds this principle into the wearable's firmware.

The practical implication is significant. Rather than asking whether real-time feedback could theoretically benefit movement outcomes—a question already answered affirmatively in clinical literature—Sensor v6.5 translates validated clinical methodology into a commercial product that operates independently in occupational and athletic settings. Organisations deploying the sensor can expect outcomes aligned with clinical evidence showing that immediate biofeedback at the moment of a movement fault improves intervention efficacy. This evidence base distinguishes Sensor v6.5 from consumer wearables that provide step counts or activity summaries; it is explicitly designed around a clinical principle validated in peer-reviewed research.

Competitive Positioning Within the Edge AI and Wearable Device Market

dorsaVi's announcement of Sensor v6.5 positions the company within the emerging category of edge artificial intelligence applied to wearable sensors. The company emphasizes that it is not building speculative technology or pursuing a future roadmap; it is shipping a commercialisation-ready product today that solves the constraint—on-device intelligence without cloud dependency—that has historically defined the boundary of what body-worn AI can achieve. Most wearable devices currently available to market either operate without artificial intelligence, offering only data logging and transmission, or require cloud connectivity for analytical processing, introducing latency and dependency on network availability.

dorsaVi's competitive advantage rests on three factors. First, it has solved the on-device compute constraint, enabling real-time analysis on the wearable itself. Second, it operates from a regulatory foundation—FDA clearance and TGA certification—that many newer entrants to the wearable AI space lack, providing immediate market access in regulated sectors. Third, it has established a decade-long deployment history across demanding applications, generating proof of concept and customer relationships that newer competitors cannot quickly replicate. The RRAM and compute-in-memory roadmap signals intent to extend this advantage further as emerging memory technologies become commercially available.

Ultra-Edge Strategy and Long-Term Product Architecture Evolution

dorsaVi frames Sensor v6.5 as the initial commercial expression of its Ultra-Edge commercialisation strategy. The term "Ultra-Edge" reflects the company's focus on intelligence operating at the furthest edge of networked systems—not in data centres, not on nearby servers, but on the sensor device itself, at the point where sensing occurs. This positioning aligns dorsaVi with broader industry trends in edge computing, where processing and decision-making are being pushed to peripheral devices to reduce latency, bandwidth consumption and dependency on centralised infrastructure.

The company's announced Ultra-Edge Modular Design and Build program indicates a structured approach to evolving Sensor v6.5 and its successors. Rather than pursuing a single-generation product, dorsaVi has established a development roadmap targeting integration of emerging memory technologies (RRAM), advanced computing architectures (compute-in-memory) and neuromorphic processing approaches. This modular design approach theoretically allows the company to refresh and advance sensor capabilities as new semiconductor technologies mature, without requiring complete redesign of the platform architecture. The immediate product is Sensor v6.5 on conventional silicon; the strategic intent is progression toward always-on operation on minimal power in future generations.