Advantech, a worldwide leader in IoT intelligent systems and embedded platforms, announced the extension of its global edge AI ecosystem partnership through a strategic collaboration with DEEPX, a South Korean AI semiconductor technology innovator with a focus on energy-efficient neural processing units (NPUs). In the announcement, the two companies also rolled out their first collaborative edge AI acceleration module, EAI-1961 series, using DEEPX´s DX-M1 NPU technology.
EAI-1961 Series represents Advantech’s first edge AI acceleration solution, integrating its new DEEPX NPU. It comes in the industry-standard M.2 form factor and supports up to 25 TOPS of AI inference performance and up to 4 GB of LPDDR5 memory. It also supports thermal performance that is optimized to be energy-efficient.
Joey Hsu, the Director of Advantech’s Embedded Business Sector, pointed out the need to compare and integrate multiple AI chip technologies because DEEPX’s high efficiency, high performance, and low power conversion or cooling capabilities can be instrumental at the edge.
“DEEPX CEO Lokwon Kim pointed out the importance of their technology participating in the global industrial ecosystem and explained that collaboration with Advantech will help them demonstrate their AI chips on the global stage,” he said.
This collaboration also enhances Advantech’s current edge AI ecosystem, which already includes collaborations with companies like Intel, Qualcomm, NVIDIA, Hailo, and Axelera AI, to name a few, making Advantech’s portfolio one of the most robust in the industry for edge intelligence.
Why This Matters for the Semiconductors & IoT Industry
This alliance is an important milestone in edge computing and underlines the larger trend that is emerging in the way semiconductors, IoT connectivity solutions, and AI technology are coming together at the edge.
1. Enhancing Artificial Intelligence Inference on the Edge
With massive amounts of data generated by IoT devices, ranging from cameras and sensors in smart factories, autonomous robotics, and industrial automation technology, it is essential to note that there is a considerable need for doing AI inference computations at the edge as opposed to computing in the cloud. The EAI-1961 is one such edge AI accelerator solution for such a need.
For semiconductor engineers and IoT solution enablers, this alliance offers not only the potential for the commercialization of NPUs with focuses other than general-purpose processors, optimized for embedded intelligence, but also the extension of the industry from general processors to domain-specific processors optimized for the IoT domain of applications.
2. Enhancing the IoT Value Chain with Scalable AI
Various Internet of Things (IoT) solutions have started making increasing use of AI technology in the field of computer vision, quality inspection, anomaly detection, predictive maintenance, and real-time situational awareness. DEEPX technology is being incorporated by the company, Avantech, into its hardware platforms, including industrial motherboards, embedded PCs, and AI systems, making it easier for companies to adopt the technology and have the ability to create production-ready and scalable AI-powered IoT solutions.
Also Read: Assent Unveils AI-Native Compliance Solution to Streamline Supply Chain Sustainability and Risk Management
3. Enriching the Semiconductor Ecosystems to Empower Diverse AIo
The collaboration demonstrates how the semiconductor industry can integrate with other hardware and software to provide end-to-end solutions. For DEEPX, partnership with Advantech ensures its chips are linked to an extensive hardware service and distribution network world-wide. Conversely, Advantech gets to benefit from the cutting-edge NPUs to enhance its edge AI solution suite.
This reflects in turn the overall semiconductor industry shift, in which semiconductor designers are not pure-supply providers anymore but are instead involved in hardware-software co-optimization ecosystems that provide end-to-end IoT and edge AI software stacks from silicon to systems integration and through to cloud connectivity.
How Businesses in These Sectors Can Benefit
Providers Of IoT And Edge Solutions: Companies intending to create smart factories, self-driving robotic systems, logistics solutions, or smart cities have the opportunity to incorporate high-performance AI modules that can enhance existing systems without having to rebuild the hardware design.
Industrial OEMs & Integrators: This is because industry standard M.2 edge AI accelerators have been provided for the purpose of accelerating the whole process of integrating advanced AI capabilities into products and services like vision systems and quality inspection solutions.
Semiconductor Partners and Chip Designers:
Collaborations of this type give credence to the commercial feasibility of specialized AI silicon chips, especially NPUs developed for low power and high performance in inferences.
This also encourages semiconductor firms to develop sophisticated AI-enabled silicon chips at an advanced edge AI level and partner with the necessary ecosystems.
Manufacturing & Automation Companies: Edge AI offering real-time analytics enables increased efficiency of product flow, minimized defects, and predictive maintenance, thereby resulting in decreased operating costs.
Software and Cloud Suppliers: Integration with robust hardware ecosystems drives the market for complementary software solutions ranging from AI model deployment infrastructure to cloud-edge solutions for orchestration and analysis.
Industry-Wide Implications
Adoption Patterns for Edge AI
With edge AI increasingly at the forefront in industries and IoT applications, collaborations such as Advantech + DEEPX are poised to accelerate their adoption by providing ready-to-deploy modules supported by world-class supply chains.
Standardized industry interfaces (such as M.2 form factor) and compatibility with key partners make integration and scalability even faster by encouraging experimentation and deployments of edge AI.
Encouraging Innovation Beyond Disciplines
Edge AI provides a basis for other innovative paradigms such as industrial automation, robotics, AR/VR visual systems, healthcare devices, and self-driving cars because Edge AI helps to enable local processing opposed to doing the processing in the center.
Concerns & Considerations
Ecosystem Integration: The ability to work well with different edge devices, operating systems, and wireless standards is still a challenge. The hardware components and artificial intelligence algorithms need to be scalable.
Power & Thermal Limitations: Although NPUs such as the DX-M1 are efficient, there are power and cooling limitations in edge environments when deployed in rugged industries.
Perhaps,
Fragmentation Risks: With numerous players providing different silicons and platforms, the need to sync standards, especially regarding software frameworks and runtimes, has assumed prominence to prevent fragmentation in the IoT and silicon industries.
Conclusion
The partnership between Advantech and DEEPX, as well as its edge AI ecosystem strategy, are a reflection of the trend in the industry towards intelligent and interconnected systems in computing where the data is generated. Convergence between semiconductors, Internet of things technology, and edge artificial intelligence is projected to change the face of various businesses by enabling them to accelerate their digital journey through edge computing technology itself. This joint effort is a shining example of the value that can be created through ecosystem innovation, which is the integration of special silicon with industry-hardened hardware platforms. It is producing value not only in the technology supply side but also on the user side.