DorsaVi Shares Jump 18% on Acquisition of Neuromorphic Technology

DorsaVi Enters US$47bn Neuromorphic Market with Technion IP Acquisition

DorsaVi (ASX: DVL) shares surged 18% this morning following exciting news that the company has acquired a neuromorphic Process-in-Memory (PIM) intellectual property portfolio from Technion, led by Professor Shahar Kvatinsky, a globally recognised pioneer in neuromorphic hardware. This acquisition positions DorsaVi at the forefront of one of the most promising frontiers in computing.

The neuromorphic computing market is projected to grow from roughly US$5.3 billion in 2023 to more than US$20 billion by 2030, according to IBM, and could reach as high as US$47 billion by 2034. For investors, this represents a significant long-term opportunity as DorsaVi transitions into a high-growth, innovation-led space with a potential runway for value creation.

Check out the research coverage of DorsaVi by Pitt Street Research!

What are the Best AI ASX Stocks to invest in right now?

Check our buy/sell tips

Processing-in-Memory Redefines How Computers Work

For investors trying to understand what this next-generation technology actually is, it helps to start with how today’s computing works. In conventional systems, processors such as CPUs and GPUs are separate from memory chips, like NAND flash and DRAM. These components constantly exchange data, which creates latency, the time it takes to transfer and process information, and consumes large amounts of energy.

Processing-in-memory (PIM) changes that architecture entirely. It enables the memory chip itself to perform computations, effectively combining the roles of storage and processing into one unified stack. Because the computation happens directly where the data resides, there’s far less need to move information back and forth between processor and memory unit. This design significantly improves speed and energy efficiency, making it a breakthrough for data-heavy applications, like AI and advanced analytics.

Mimicking the Biological Brain to Power the Next Wave of AI

Another feature that makes this technology especially intriguing is its neuromorphic design. The human brain remains one of the most efficient computing systems ever created. Unlike traditional computers that continuously process information, the brain operates on an event-based model, it only responds when stimulated. Neurons fire in response to specific signals, allowing the brain to conserve energy and process information far more efficiently.

Neuromorphic computing aims to replicate this principle. It allows different parts of a chip to process information in parallel, much like various regions of the brain working together. This creates a synergistic system capable of handling complex tasks, such as pattern recognition, sensory processing or adaptive learning, with minimal power consumption. When combined with Processing-in-Memory architecture, it forms a powerful foundation for next-generation AI hardware that is faster, smarter and dramatically more energy efficient than today’s solutions.

Humanoid Robotics Could Be the Next Trillion-Dollar Market

When we look at companies like Tesla and Apple working toward bringing AI-powered humanoid robots to market, one of the biggest challenges lies in developing the hardware that allows these machines to think and react like humans. To operate efficiently, humanoid robots need a form of computing that mimics how the human brain learns, processes and responds to stimuli. This is exactly where neuromorphic and Processing-in-Memory technologies come into play, they replicate the brain’s event-based and parallel processing structure within a chip, enabling faster learning and adaptive behavior with lower power use.

According to investment bank Morgan Stanley, the global humanoid robotics market, currently worth around US$3 billion, could grow to as much as US$5 trillion by 2050. This is a massive, emerging industry, with Asia-Pacific projected to be the fastest-growing region, supported by strong government initiatives in Japan and China. The long-term potential for neuromorphic chips is significant because they represent the closest hardware model to how the human brain functions, and will likely become the most efficient and scalable way to power humanoid robots in the years ahead.

The Investors Takeaway: DorsaVi can potentially address the very high-potential Robotics market

While the potential within robotics is highly compelling, DorsaVi remains primarily focused on its core sensory technology business for now. This is likely where its newly acquired neuromorphic PIM IP will have the most immediate impact, enhancing how the company captures and interprets human movement data. DorsaVi’s system measures motion, muscle activity and acceleration in three dimensions, then applies cloud-based analytics to assess injury risk, rehabilitation progress or athletic performance. Integrating neuromorphic PIM could significantly improve the system’s ability to process raw biosignals in real time, enabling faster, more adaptive insights.

In our view, though, the longer term opportunity is in potentially supplying the neuromorpic/PIM IP to OEM’s, such as chip designers and manufacturers as well as electronics companies in the Robotics space.

That said, investors should recognise that DorsaVi is still an early-stage business. The company has yet to demonstrate consistent commercial traction or widespread adoption of its technology, both within its sensory applications and its longer-term ambitions in robotics. Another point worth noting is that the current Board and management team may need more experience in advanced computing, AI-driven hardware and ReRAM development as the company transitions into these more complex, technology-intensive markets.