Technology

Can Edge AI Keep Up?

Source

SemiEngineering

Published

Apr 23, 2026

TL;DR

AI Generated

Experts discuss the challenges of keeping edge AI architectures adaptable while maintaining power, performance, and area targets. The cadence for model updates varies by application, with some industries experiencing rapid changes while others remain more static. Heterogeneous architectures and robust software/compiler toolchains are crucial for balancing adaptability with efficiency. The discussion includes insights from industry leaders at Arm, Cadence, Expedera, Mixel, Quadric, Rambus, Siemens EDA, and Synopsys on the evolving landscape of AI model development and hardware design.

Read Full Article

The Great AM Reset: Why Applications Will Decide Who Survives

The article discusses the importance of applications in determining which companies will thrive in the wake of the Great AM Reset. It emphasizes that the ability of applications to adapt and evolve will be crucial for businesses to survive in the changing landscape. The article suggests that companies need to prioritize the development and optimization of their applications to remain competitive. It also highlights the role of technology in enabling businesses to innovate and meet the evolving needs of consumers.

3DPrint.com•

3 months ago

2026 Outlook with Howard Pakosh of Tekstart

Howard Pakosh, founder of TekStart Group, discusses the company's journey in developing AI-focused semiconductor products, with the upcoming launch of their latest chip, Cognitum, being a significant milestone in 2025. Despite facing various challenges including market uncertainty and new tariffs, the team's resilience and focus on engineering discipline helped navigate through the obstacles. Pakosh emphasizes the importance of Edge AI as a key growth area for 2026, highlighting the need to push intelligence closer to where data is generated for faster decision-making. TekStart's Cognitum chip is designed to address this demand for edge reasoning by enabling autonomous reasoning directly on the chip, reducing cloud inference costs and making large-scale deployments more practical. The company plans to continue active participation in conferences in 2026 to engage with customers and share insights on their technological advancements.

SemiWiki•

4 months ago

Edge AI Safety: Agentic AI Architecture That Leverages 3D To Integrate A Dedicated Safety Layer (Princeton, HKUST, NC State Univ.)

Researchers from Princeton University, Hong Kong University of Science and Technology, and North Carolina State University published a technical paper titled “3D Guard-Layer: An Integrated Agentic AI Safety System for Edge Artificial Intelligence.” The paper proposes an AI safety architecture leveraging 3D integration to enhance safety in edge AI systems. By dynamically learning and mitigating attacks against AI systems, this architecture aims to improve system reliability and performance while minimizing costs. The system utilizes the advantages of co-location with edge computing hardware to monitor, detect, and proactively address threats to AI systems.

SemiEngineering•

6 months ago

SemiEngineering

LPDDR: A Versatile Memory Powering The Next Wave Of Mobile, Edge & Endpoint Computing

The article discusses the evolution of memory technology, highlighting LPDDR as a key solution for mobile, edge, and endpoint platforms due to its balance of performance and power efficiency. LPDDR has advanced through generations, with LPDDR6 offering features like dual sub-channels and enhanced reliability for AI inference workloads. Compared to HBM and GDDR, LPDDR excels in energy efficiency and idle power savings, making it ideal for devices where battery life and thermal limits are crucial. LPDDR is increasingly integrated into smartphones, tablets, laptops, automotive systems, and AI accelerators, offering unique features like fine-grained concurrency and reliability enhancements. The article emphasizes the importance of LPDDR in the evolving landscape of mobile, automotive, and edge AI platforms.