Reports
AI-generated structured vendor updates
AMD and Liquid AI Discuss Efficient AI Architecture from Silicon to Systems
AMD's CTO and Liquid AI's CEO discuss the evolution of AI architecture, emphasizing efficiency as key to extending AI from the cloud to edge and endpoint devices. They argue that co-design from silicon to systems enables low-power, responsive AI inference, supporting always-on agents and multi-model orchestration.
Arm Launches Performix Performance Toolkit, Targeting AI Agent Era Optimization
Arm launched Performix, a free performance analysis toolkit designed to provide unified performance insights and optimization across the Arm platform for AI agent development. Integrated into mainstream AI dev environments via the Arm MCP Server, it turns runtime hardware data into actionable optimization guidance, with support from ecosystem partners like Microsoft and MongoDB.
AMD Extends Edge AI Architecture to Space, Defining Orbital Computing Paradigm
AMD's CTO proposes applying the core principles of 'performance-per-watt' and 'mission-critical reliability' from terrestrial edge AI to space computing. The company is providing a repeatable platform foundation for in-orbit satellite intelligence and future orbital data centers through heterogeneous computing, open software stacks, and modular system design.
AMD Highlights AI PC as Critical Infrastructure for Enterprise Agentic AI in IDC White Paper
AMD released an IDC white paper indicating that over 80% of enterprises are planning, piloting, or deploying AI PCs to support scaled Agentic AI. The report highlights high-performance NPUs and on-device AI processing as critical for enabling real-time, secure workflows, signaling a shift in enterprise AI infrastructure from cloud to endpoint.
Microsoft Launches Hosted AI Agent Infrastructure, Treating Agents as Independent Compute Entities
Microsoft introduces "Hosted agents" in its Foundry platform, providing each AI agent with an isolated, enterprise-grade sandbox featuring durable state, built-in identity, and governance. This move aims to standardize the runtime infrastructure for AI agents, lowering the barrier to enterprise deployment, though comments note it shifts the control point from the application layer to the infrastructure layer.
Cisco Launches AI Agent Security Scanner, Shifting Security Control Point to IDEs
Cisco has launched an AI Agent Security Scanner IDE extension designed to identify and mitigate new attack surfaces in the AI development toolchain. The tool provides local, multi-layered protection by statically scanning MCP server configurations and agent skill definitions, embedding secure coding rules during code generation, and continuously monitoring file integrity at runtime.
Google Cloud Next '26: Agent Gateway Seizes Control Plane, TPU 8i Locks Inference
Google Cloud Next '26 announces 8th-gen TPUs (8t for training, 8i for inference), Agent Platform with Agent Gateway, Agent Identity, Agent-to-Agent Orchestration, Agentic Data Cloud, and Agentic Defense integrating Wiz. The move shifts control from infrastructure to agent orchestration, locking enterprises into a vertically integrated stack.
Cisco and NVIDIA Elevate Network to AI Media Processing Control Plane
Cisco and NVIDIA deepen collaboration with a validated design based on the open-standard Media Exchange Layer (MXL). This integration merges Cisco's IP media fabric with NVIDIA's Holoscan platform, transforming the network from a transport layer into an active processing layer that supports real-time AI inference, enabling low-latency, multilingual AI-driven live media production for broadcasters.
Anthropic Launches Claude Opus 4.7 with Cyber Safeguards
Anthropic has launched Claude Opus 4.7, showing notable gains in advanced software engineering, multimodal understanding, and long-horizon reasoning. This release introduces automated safeguards to detect and block prohibited high-risk cybersecurity uses, alongside a Cyber Verification Program for legitimate research, aiming to inform the safe future release of more powerful models like Mythos.
NVIDIA Shifts AI Infrastructure Metric from FLOPS to Cost Per Token
NVIDIA advocates for "cost per token" as the primary economic metric for AI infrastructure, replacing "FLOPS per dollar." This shift moves the focus from computational inputs to business outputs, requiring full-stack optimization across hardware, software, and networking to lower enterprise AI inference TCO.
Cisco Details How AI Agentic Frameworks Reshape Network Operations Architecture
Cisco's blog details the application of AI Agentic frameworks in network engineering, outlining an evolution from chatbots to multi-step workflow orchestration. The core involves encoding human expertise into 'skill' files, connecting to infrastructure APIs via the MCP protocol, and setting human-in-the-loop gates, shifting the engineer's role from task executor to orchestrator.
Cisco Shares Enterprise AI Assistant Patterns, Emphasizing Deterministic Security and Guided Interaction
Based on 18 months of production experience with its Customer Experience AI Assistant, Cisco identifies non-obvious patterns critical for enterprise AI success. Key insights include enforcing RBAC via deterministic code (not LLM prompts), proactively disambiguating enterprise acronyms, minimizing clarification loops, and providing guided follow-up questions grounded in actual system capabilities.
Arm Partners with Monash University Malaysia to Advance Semiconductor Talent for AI Era
Arm announced a collaboration with Monash University Malaysia's School of Engineering, donating IC design development boards and appointing an executive as a guest lecturer. The initiative aims to cultivate semiconductor talent with hands-on Arm architecture and modern system design experience for the AI era.
Anthropic Partners with Mozilla, AI Models Independently Discover High-Severity Firefox Vulnerabilities
Anthropic's Claude Opus 4.6 model discovered 22 vulnerabilities in Mozilla Firefox over two weeks, with 14 classified as high-severity. This demonstrates AI's ability to independently identify unknown vulnerabilities in complex software and its nascent capability to generate exploits, signaling a new phase in AI-powered cybersecurity offense and defense.
ARM Optimizes Gemma 4 On-Device AI Performance with Google
ARM's SME2 technology in Armv9 architecture accelerates Google's Gemma 4 model on mobile devices, achieving 5.5x prefill speedup and 1.6x faster decoding. The collaboration enables developers to access optimizations without code changes, shifting on-device AI toward default mobile app architecture.
Google Launches Gemma 4 Open Models, Targeting Edge Inference and AI Agent Architecture
Google introduces the Gemma 4 open model family, with four sizes from 2B to 31B parameters, emphasizing breakthrough intelligence-per-parameter and native support for agentic workflows, multimodality, and long context. The small models are engineered for edge devices, aiming to bring frontier reasoning to mobile and IoT scenarios.
Google Launches Gemma 4 Open Model Family
Google introduces Gemma 4 open model family with four size variants, optimized for edge and mobile devices. The series supports multimodal processing, long context windows and 140+ languages under Apache 2.0 license.
AMD Announces Breakthrough MLPerf Inference 6.0 Results, Showcasing Multinode Scaling and Multimodal Capabilities
AMD's MLPerf Inference 6.0 submission, powered by Instinct MI355X GPUs, surpassed 1 million tokens per second for the first time on models like Llama 2 70B and GPT-OSS-120B. The results highlight efficient multinode scaling, rapid enablement of new workloads (e.g., text-to-video model Wan-2.2-t2v), and reproducible performance across a broad partner ecosystem.
Cisco Discloses Memory Poisoning Attack Method in AI Coding Assistants
Cisco's security team discovered and validated a persistent memory poisoning attack method targeting AI coding assistants like Claude Code, demonstrating how tampering with MEMORY.md system files can persistently manipulate AI behavior. This vulnerability prompted Anthropic to remove user memory files' system prompt privileges in v2.1.50.
Intel Demonstrates AI Performance with Xeon 6 and Arc Pro GPUs in MLPerf Inference
Intel showcased the performance of its Xeon 6 CPUs and Arc Pro B-Series GPUs in the MLPerf Inference v6.0 benchmarks, particularly in handling large language models (LLMs). The results indicate that a system with four Arc Pro B70 GPUs can process 120B parameter models, delivering up to 1.8x higher inference performance in multi-GPU setups.