Hyundai Launches MobED: An AI-Powered Autonomous Robot Platform
At the International Robot Exhibition (iREX) 2025 in Tokyo, Hyundai Motor Group has officially launched MobED (Mobile Eccentric Droid), its first mass-produced autonomous mobility robot. First revealed as a concept at CES 2022, MobED has evolved into a production-ready platform powered by artificial intelligence. The robot is engineered for a wide range of industrial and everyday applications, featuring AI-based autonomous navigation, LiDAR-camera fusion sensors, and an eccentric posture control mechanism for stable movement across diverse terrains. An intuitive 3D UI/UX on a wide touchscreen controller enables easy operation, including self-mapping and autonomous driving. Mass production of MobED is set to begin in the first half of 2026.
Sources:
Drone Trial to Enhance Autonomous Vehicle Safety on 5G Networks
A new trial under the Scale project is exploring the use of “drone in a box” systems to improve the safety and performance of autonomous ground vehicles operating on 5G networks at the National Exhibition Centre. The project will evaluate how data from drones can support passenger assurance, venue safety, automated inspections, and emergency response planning. The trial will also examine the integration of these drone systems with autonomous mobility operations over 5G and assess the network’s performance under various crowd conditions. Project partners include Skyfarer for drone integration, Ohmio as the autonomous vehicle provider, and Solihull and Coventry councils for public sector innovation.
Mistral AI Releases ‘Mistral 3’ Family of Open-Source Models
Mistral AI has launched its new ‘Mistral 3’ family of open-source, multimodal, and multilingual AI models. The release features three small, dense models (14B, 8B, and 3B) and a larger mixture-of-experts model named ‘Mistral Large 3’. The company asserts that these models deliver results comparable to closed-source competitors while providing the transparency of open-source development. According to Mistral, ‘Large 3’ surpasses other leading models like DeepSeek V3.1 and Kimi K2 in general and multilingual prompts. The new models have already secured support from major industry players, with HSBC and Nvidia announcing their adoption. HSBC intends to use the models to expand its use of generative AI, while Nvidia has confirmed the models are optimized for its high-performance chips. The ‘Mistral 3’ family is accessible on platforms including Mistral AI Studio, Amazon Bedrock, and Azure Foundry.
Sources:
AWS Introduces AI-Powered DevOps Agent for Faster Incident Response
Amazon Web Services (AWS) has released a public preview of the AWS DevOps Agent, an AI tool engineered to accelerate incident response and enhance system reliability. This new agent autonomously analyzes data from various monitoring tools, correlates it with recent code deployments, and pinpoints the likely root causes of production incidents. By offering targeted mitigation recommendations, the DevOps Agent aims to significantly reduce the mean time to resolution (MTTR) for on-call engineers. In a test with Commonwealth Bank of Australia, the agent identified a problem’s source in under 15 minutes, a task that would typically take a senior engineer hours. Beyond real-time response, the agent also analyzes past incidents to recommend systemic improvements and prevent future issues. The tool integrates with third-party services like Datadog and Dynatrace and communicates via dedicated Slack channels.
Sources:
AWS Unveils New Security Products and Enhancements at re:Invent 2025
At its re:Invent 2025 conference, Amazon Web Services (AWS) announced several new security products, including a preview of the AWS Security Agent. This agent is designed to proactively secure applications throughout the development lifecycle by performing automated security reviews and context-aware penetration testing. AWS also announced the general availability of the AWS Security Hub, a centralized dashboard for viewing and prioritizing security risks, now updated with features for historical trends and exposure summaries. Additionally, Amazon GuardDuty Extended Threat Detection has been enhanced with new findings for EC2 instances and ECS tasks, improving visibility into multi-stage attacks across virtual machine and container environments. Another key introduction was the IAM Policy Autopilot, an open-source tool that helps AI coding assistants generate appropriate AWS IAM policies by analyzing application code.
Sources:
SEALSQ Invests in EeroQ to Advance US Quantum Chip Development
SEALSQ Corp has made a strategic investment in EeroQ, a U.S.-based quantum chip design company. This marks the first U.S. investment under SEALSQ’s ‘Quantum Made in USA’ strategy, which aims to strengthen American and European leadership in quantum technologies. EeroQ is developing a unique quantum computer based on electrons trapped on superfluid helium, a technology enabling ultra-compact processors manufactured with standard CMOS processes. The investment is part of a broader SEALSQ initiative to invest up to $35 million in quantum computing startups. EeroQ recently expanded its operations, opening a new 9,600-square-foot R&D facility in Chicago.
Sources:
UTokyo and IBM Develop ‘Krylov’ Quantum Simulation Algorithm
Researchers from the University of Tokyo and IBM have co-developed a new algorithm for ‘Krylov quantum diagonalization’ (KQD). This technique is designed to find the ground state (the lowest-energy state) of a complex quantum system, a task that challenges even the most powerful supercomputers. The KQD algorithm offers a promising alternative to variational methods and is seen as a foundational element for future quantum computing research. In a recent demonstration published in Nature Communications, the team successfully used the algorithm on an IBM Heron quantum processor to simulate a many-body system of up to 56 sites—one of the largest simulations of its kind on a quantum processor to date. This achievement highlights the potential for current quantum computers to address complex problems in condensed matter physics and quantum chemistry without requiring full fault tolerance.
Sources:
Helm 4 Released with Major Updates for Kubernetes Package Management
The Cloud Native Computing Foundation (CNCF) has announced the release of Helm 4, the first major update to the Kubernetes package manager in six years, timed with the project’s 10th anniversary. This new version introduces significant enhancements to improve scalability, security, and developer workflows for managing applications on Kubernetes. Key features in Helm 4 include a redesigned plugin system with support for WebAssembly (WASM) based plugins, native support for server-side apply, and improved resource watching. The release also provides an updated SDK with modern Go logging interfaces and the ability to embed Helm commands directly into other applications. While introducing some breaking changes, the update aims to maintain compatibility for most workflows from Helm v3 and address the modern complexities of CI/CD, security, and multi-cluster operations.
Sources: