BrainChip Holdings Ltd
Australian public technology company developing neuromorphic processors and software for low-power edge artificial intelligence.
Last updated November 4, 2026
AI crawler reads
No AI crawler has read this profile since Oct 28, 2026.
Overview
BrainChip Holdings Ltd is an Australian technology company focused on neuromorphic computing, edge artificial intelligence, and machine-learning hardware. Founded in 2004 by Peter Van Der Made, the company develops processors and software intended to perform AI inference closer to where data is generated, including in cameras, sensors, industrial systems, vehicles, and other embedded devices. Its approach is based on spiking neural networks, a class of neural models inspired by the event-driven signaling of biological nervous systems. BrainChip’s principal technology platform is Akida. The first-generation Akida 1000 is described as an event-based neural processor containing approximately 1.2 million artificial neurons and 10 billion artificial synapses. Rather than continuously processing every input value, event-based systems are designed to respond to meaningful changes or signals. BrainChip positions this architecture as a way to reduce data movement and energy consumption while enabling real-time inference and, in some configurations, incremental learning on edge devices. The processor uses on-chip memory and a mesh of processing nodes, allowing neural-network layers to operate in parallel. The company also develops MetaTF, a software and development environment for creating, converting, training, and deploying spiking neural networks. MetaTF is designed to work with image, video, and sensor data and can support deployment on Akida hardware as well as conventional processors. This software layer is important to BrainChip’s strategy because it connects model development with embedded deployment, allowing developers to build applications for areas such as security, surveillance, autonomous systems, and industrial automation. BrainChip became a public-company vehicle through a two-stage corporate transaction in 2015. Australian mining company Aziana acquired BrainChip in March of that year, after which the dormant or inactive Aziana entity was used in a reverse-merger transaction that placed BrainChip on the Australian Securities Exchange in September 2015. The transaction enabled Van Der Made to pursue commercialization of the company’s AI-processor concept. In 2016, former Exar chief executive Louis Di Nardo became BrainChip’s chief executive, while Van Der Made moved into the chief technology officer role. The company began moving from technology development toward commercial availability during the early 2020s. It announced orders for Akida AI Processor Development Kits in October 2021 and for Akida AI Processor PCIe boards in January 2022. BrainChip subsequently expanded its ecosystem through collaborations with application and platform companies, universities, semiconductor-industry participants, and embedded-AI developers. Announced relationships or integrations included NVISO, the Rochester Institute of Technology, Intel Foundry Services, Edge Impulse, and Arm-related technology initiatives. BrainChip has also continued to develop later generations of its platform. In January 2023 it announced completion of the design of an AKD1500 reference chip. In March 2023 it introduced a second-generation Akida platform with support for 8-bit weights and activations, a Vision Transformer engine, and hardware support for temporal event-based neural networks. The company also announced integration of the Akida processor family with Arm Cortex-M85 technology. These developments reflect BrainChip’s effort to broaden the types of neural-network workloads that can run efficiently on embedded systems. BrainChip’s market positioning is specialized rather than consumer-facing. It competes for design wins and developer adoption in edge AI, embedded semiconductors, neuromorphic engineering, and low-power inference. Its commercial proposition is based on processing data locally, reducing dependence on cloud connectivity, lowering latency, and potentially limiting energy use and data transmission. The company’s long-term performance depends on the adoption of neuromorphic architectures, the availability of production silicon and development tools, ecosystem support, and the conversion of technical partnerships into sustained customer deployments.
History
BrainChip was founded in Australia in 2004 by Peter Van Der Made. Its central technical idea was to develop computing hardware inspired by biological neural processing rather than relying solely on conventional, continuously clocked digital architectures. The company’s work became associated with neuromorphic computing and spiking neural networks, in which information is represented through discrete events or spikes. This approach is intended to support efficient processing of sensory data and machine-learning workloads at the edge. A major corporate transition occurred in 2015. Australian mining company Aziana acquired BrainChip in March. Later that year, a reverse-merger transaction involving the dormant or inactive Aziana entity brought BrainChip onto the Australian Securities Exchange in September. The listing gave BrainChip a public-company structure through which it could pursue commercialization of its processor technology. Van Der Made continued to provide technical leadership as the company developed its hardware and software proposition. In 2016, BrainChip appointed Louis Di Nardo, formerly associated with Exar, as chief executive officer. Van Der Made became chief technology officer. The division placed commercial and operational leadership with Di Nardo while retaining the founder’s involvement in the engineering direction of the business. BrainChip’s technology stack developed around two linked elements: the Akida processor family and the MetaTF software environment. MetaTF was designed to help developers create or convert neural networks into spiking neural networks and deploy them on Akida hardware or, in some cases, conventional processors. The environment supports data types including images, video, and sensor streams. The intended application areas include security and surveillance, autonomous vehicles, industrial automation, and other embedded systems that require responsive local inference. The first-generation Akida 1000 was described as an event-based neural processor with approximately 1.2 million artificial neurons and 10 billion artificial synapses. Its architecture includes 80 nodes connected through a mesh network. Each node contains neural-processing units and associated memory, with the design allowing neural-network layers to operate in parallel. Processing and intermediate results are intended to remain close to the computation hardware, supporting low-latency inference and reducing the need to transmit raw data to a remote cloud service. The platform was also designed with incremental learning capabilities for selected edge applications. BrainChip began making the platform more accessible to developers in the early 2020s. In October 2021 it announced that it was taking orders for Akida AI Processor Development Kits. In January 2022 it announced order availability for Akida AI Processor PCIe boards. These products provided practical evaluation and development routes for customers and partners interested in testing Akida-based inference. The company expanded its external network in 2022 and 2023. In April 2022 it announced a collaboration with NVISO involving applications and technology. In November 2022 it added the Rochester Institute of Technology to its University AI Accelerator Program. In December 2022 it became part of the Intel Foundry Services technology-partner ecosystem. In January 2023 Edge Impulse announced support for the Akida processor platform, linking BrainChip’s hardware with a broader embedded machine-learning development workflow. Product development continued during 2023. On 29 January BrainChip announced completion of the design of its AKD1500 reference chip. On 6 March it announced a second-generation Akida platform. The updated platform added 8-bit weights and activations, a Vision Transformer engine, and hardware support for temporal event-based neural networks. On 12 March BrainChip announced integration of the Akida processor family with Arm Cortex-M85 technology. BrainChip’s history is therefore characterized by a combination of specialized semiconductor development, software-toolchain creation, public-market financing, and ecosystem building. Rather than presenting itself as a general-purpose consumer electronics brand, it operates in the business-to-business and developer-oriented markets for embedded AI and neuromorphic processing. Its commercial prospects are tied to the adoption of low-power edge inference, the ability of developers to use its tools and processors, and the conversion of partnerships and reference designs into production deployments.
- 2023
Edge Impulse supports Akida
[1]Edge Impulse announced support for BrainChip’s Akida processor platform.
- 2023
AKD1500 reference-chip design completed
[1]BrainChip announced completion of the design of its AKD1500 reference chip.
- 2023
Second-generation Akida platform announced
[1]The second-generation platform introduced support for 8-bit weights and activations, a Vision Transformer engine, and temporal event-based neural-network hardware.
- 2023
Arm Cortex-M85 integration announced
[1]BrainChip announced integration of the Akida processor family with Arm Cortex-M85 technology.
- 2022
Akida PCIe boards become orderable
[1]The company announced order availability for Akida AI Processor PCIe boards.
- 2022
NVISO collaboration announced
[1]BrainChip announced a collaboration with NVISO involving AI applications and technology.
- 2022
Rochester Institute of Technology joins university program
[1]BrainChip added the Rochester Institute of Technology to its University AI Accelerator Program.
- 2022
BrainChip joins Intel Foundry Services partner ecosystem
[1]BrainChip announced participation as a technology partner in Intel Foundry Services.
- 2021
Akida development kits become orderable
[1]BrainChip announced that customers could place orders for Akida AI Processor Development Kits.
- 2016
Louis Di Nardo becomes chief executive
[1]Former Exar chief executive Louis Di Nardo was appointed CEO, while founder Peter Van Der Made took the CTO role.
Products and positioning
Neuromorphic computing and low-power edge AI for embedded inference, sensor processing, and real-time machine learning.
Neuromorphic processor
Akida 1000
The Akida 1000 is BrainChip’s first-generation event-based neural processor. It is described as containing approximately 1.2 million artificial neurons and 10 billion artificial synapses. Its architecture uses 80 mesh-connected nodes, each containing neural-processing units and associated memory. The design runs neural-network layers in parallel and is intended to support low-power, low-latency inference and incremental learning on edge devices.
Neuromorphic processor platform
Akida
Akida is BrainChip’s processor family and embedded AI platform for spiking-neural-network workloads. It is designed to process event-based inputs locally, allowing applications to analyze sensor, image, or other data without always sending raw information to a cloud service. BrainChip has developed the platform for applications including surveillance, autonomous systems, industrial automation, and other power-constrained embedded environments.
AI development software
MetaTF
MetaTF is BrainChip’s software environment for developing and deploying spiking neural networks. It supports Python-based workflows, training and conversion of neural networks, and deployment on Akida processors or conventional processors. The environment is intended for image, video, and sensor data and serves as the software bridge between model development and embedded edge-AI implementation.
Developer hardware · 2021
Akida AI Processor Development Kits
These development kits provide hardware and tools for evaluating BrainChip’s Akida processor technology. BrainChip announced order availability in October 2021, giving developers a practical way to experiment with event-based processing, spiking neural networks, and edge-AI applications.
AI accelerator board · 2022
Akida AI Processor PCIe boards
Akida PCIe boards are accelerator products intended to connect BrainChip’s processor technology to compatible host systems. Announced as orderable in January 2022, they provide a development and evaluation path for customers testing Akida-based inference in workstation or embedded computing environments.
Reference chip · 2023
AKD1500
The AKD1500 is a BrainChip reference-chip design associated with the continuing development of the Akida platform. BrainChip announced completion of the design in January 2023. The reference chip is intended to support platform development and evaluation rather than function as a consumer-facing product.
Neuromorphic processor platform · 2023
Second-generation Akida platform
Announced in March 2023, the second-generation Akida platform expanded the processor family with support for 8-bit weights and activations, a Vision Transformer engine, and hardware support for temporal event-based neural networks. These additions were intended to broaden the range of modern AI models that could be implemented in power-efficient edge hardware.
Brand decisions
2023
Second-generation Akida platformGeneration changeBackground
BrainChip aimed to extend the types of AI workloads supported by its event-based and neuromorphic architecture.
Action taken
The company introduced support for 8-bit weights and activations, a Vision Transformer engine, and temporal event-based neural-network hardware.
Aftermath
The updated platform broadened BrainChip’s technical proposition for embedded computer vision and time-dependent sensor workloads.
2022
Akida PCIe boards offered for orderProduct launchBackground
BrainChip expanded its hardware evaluation options beyond development kits.
Action taken
The company announced that Akida AI Processor PCIe boards were available for order.
Aftermath
The boards provided a host-system integration path for developers evaluating Akida-based edge inference.
2021
Akida development kits offered for orderProduct launchBackground
BrainChip sought to move Akida from a technology concept toward practical developer evaluation and application development.
Action taken
The company announced order availability for Akida AI Processor Development Kits.
Aftermath
The announcement gave developers and prospective customers a route to test BrainChip’s processor and software ecosystem.
2015
Use of Aziana acquisition and reverse merger to create a public listingM&ABackground
BrainChip required a public-company structure to support the commercialization of its artificial-intelligence processor technology.
Action taken
Aziana acquired BrainChip in March 2015, and a later reverse-merger transaction involving the dormant Aziana entity placed BrainChip on the ASX in September.
Aftermath
The transaction established BrainChip as an Australian listed technology company and enabled it to continue developing and commercializing its neuromorphic hardware.
Sources2 sources
Cite this profile: Cite the canonical profile. /brand-wiki/brainchip-holdings-ltd · Editorial policy · How profiles are compiled