Nano Dimension acquires AI company DeepCube for $70M

Nano Dimension, a 3D printer manufacturer with a specialism in additively manufactured electronics (AME), has announced the acquisition of machine learning company DeepCube.

Having already signed a definitive agreement, Nano Dimension is set to pay DeepCube shareholders around $40 million in cash and a further $30 million in American Depositary Shares (ADSs). The ADSs will stand still for varying periods of up to three years. The transaction is expected to close in the coming week, subject to customary closing conditions being met.

Deep learning with DeepCube

At the core of DeepCube’s proprietary technology is a number of deep learning algorithms that the company uses to improve data analysis and deploy complex artificial intelligence systems. The company boasts fast and accurate deep learning model training that can significantly improve inference performance and real-time metrics. As a bonus, DeepCube’s framework can reportedly be deployed on top of any existing hardware in both datacenters and edge devices, resulting in benefits such as speed improvement and memory use reduction.

The company’s technical staff – ML experts, defense force veterans, homeland security professionals, governmental big data agencies, and academics – are all expected to join Nano Dimension once the deal closes.

The DeepCube team. Photo via DeepCube.
The DeepCube team. Photo via DeepCube.

Towards distributed digital fabrication

So what does Nano Dimension want with a deep learning company? It starts with the ongoing semiconductor chip shortage being faced worldwide. According to Yoav Stern, CEO of Nano Dimension, similar supply chain issues are also hovering over the printed circuit board market, with PCB manufacturers being subjected to even greater margin pressures due to the prevalence of low labor costs in China.

Nano Dimension wants to tackle the issue with the development of an ML-based ‘Distributed Electronic Fabrication’ network. The platform will enable digital control of the entire supply chain for AMEs and 3D printed high-performance electronic devices (Hi-PEDs) produced using Nano Dimension’s 3D printers.

Stern explains, “The core of this solution will be DeepCube’s AI/ML/DL brain that is expected to manage a neural network of edge devices. They will self-learn and self-improve their efficacies, and self-manage and maximize yield throughout the total network. Nano Dimension machines shipping today, as well next generation devices which are under development, will be edge devices in this digital-fabrication-neural-network solution.”

Much like how personal computers went from standalone devices to nodes in local areas networks (and eventually the internet), Nano Dimension 3D printers will serve as contact points to a wider web. Nano Dimension customers will be able to 3D print as many electronic devices as they need, where they need them, with ML-optimized yields and price points.

Dr. Eli David, CTO and co-founder of DeepCube, concludes, “Over the last 18 months, DeepCube has been negotiating with a few of the world-leading integrated circuits and central processing unit (CPU) manufacturers which are still hoping to adopt its unique AI technology for dramatically improving the efficiency of deep learning deployments in the real-world. Those opportunities will be pursued as a part of Nano Dimension’s strategy.”

Nano Dimension's DragonFly LDM 3D electronics printer. Photo via Nano Dimension.
Nano Dimension’s DragonFly LDM 3D electronics printer. Photo via Nano Dimension.

Machine learning in 3D printing

Complex machine learning models have a whole host of applicatications in 3D printing, beyond just distributed workflow optimization. Metal and composite 3D printer OEM Markforged recently launched its own AI-based Blacksmith software for use with the X7 3D printer. Leveraging the X7’s own existing laser micrometer and a patented scanning algorithm, Blacksmith can be used for on-the-fly part inspection mid-print.

Elsewhere, at the University of Tübingen, researchers recently 3D printed a robotic arm capable of mimicking the movements of an elephant’s trunk using machine learning algorithms. Equipped with a gripper on the tip, the FDM-printed robot can roam around and learn to adapt to new tasks, such as picking up small objects and transporting them.

Subscribe to the 3D Printing Industry newsletter for the latest news in additive manufacturing. You can also stay connected by following us on Twitter and liking us on Facebook.

Looking for a career in additive manufacturing? Visit 3D Printing Jobs for a selection of roles in the industry.

Featured image shows the DeepCube team. Photo via DeepCube.