Technology

Beyond the Bottleneck: AI Cluster Networking Report 2025

Source

SemiEngineering

Published

Sep 24, 2025

TL;DR

AI Generated

The article discusses the increasing demand on data center networks due to the growing complexity of artificial intelligence (AI) technology. It highlights the importance for organizations to carefully design, test, and scale their infrastructure as AI becomes a central part of enterprise strategies. The report, based on a global survey by Heavy Reading and Keysight Technologies, delves into the challenges, technology choices, and investment priorities shaping AI cluster networking in 2025. It emphasizes the need for a scalable foundation in AI infrastructure, focusing on more than just speed.

Read Full Article

AI adoption rate is declining among large companies — US Census Bureau claims fewer businesses are using AI tools

A survey by the U.S. Census Bureau indicates a decline in AI adoption among large companies, with usage dropping from 13.5% to 12% between June and August. The survey, conducted biweekly, tracks AI adoption trends and is considered reliable with data from 1.2 million firms. The decline in AI adoption is also observed among smaller businesses, raising questions about the impact on employment. While some industry leaders like AMD remain optimistic about AI's future, recent trends suggest a potential slowdown in AI adoption across various business sizes.

Tom's Hardware•

8 months ago

SemiEngineering

Scalable I/O Virtualization: A Deep Dive Into PCIe’s Next Gen Virtualization

The article delves into Scalable I/O Virtualization (SIOV) as a solution to the limitations of Single Root I/O Virtualization (SR-IOV) in high-density, multi-tenant cloud environments. SIOV introduces a software-centric approach, utilizing Scalable Device Interfaces (SDIs) managed by the Host OS and hypervisor to enhance resource allocation and isolation. Key aspects of SIOV include thousands of lightweight virtual interfaces, seamless integration with trusted computing, and the importance of SDI reset mechanisms for secure fault recovery. The article emphasizes the need for rigorous verification to ensure secure and scalable cloud platforms, focusing on aspects like RID isolation, page-level BAR security, interrupt integrity, and SDI reset handling.

SemiEngineering•

9 months ago

MIT Technology Review

Finding value from AI agents from day one

Agentic AI, capable of autonomous decision-making, is gaining traction in corporate settings for tasks like procurement and cybersecurity. Despite excitement, experts caution against rushing deployment to avoid complexity and cost. Prioritizing interoperability early on can enhance the value of AI agents, with multi-agent systems expected to play a crucial role in the future. Organizations are advised to focus on incremental use cases and simplicity to maximize the benefits of agentic AI. Third-party developers are already offering multi-agent capabilities, but larger organizations may need to develop their own API architecture for long-term success.

MIT Technology Review•

10 months ago

Dr. L.C. Lu on TSMC Advanced Technology Design Solutions

Dr. L.C. Lu, a key figure at TSMC, focuses on design-technology co-optimization, packaging innovations, and AI-driven methodologies for next-gen semiconductor systems. TSMC emphasizes DTCO and DDCL innovations for scaling from N5 to A14 nodes, with NanoFlex and NanoFlex Pro architectures offering efficiency gains. N2P and N2U nodes incorporate advanced DTCO and power delivery optimizations, with hybrid dual-rail architectures achieving significant energy savings. TSMC collaborates with EDA partners for AI integration, enhancing productivity and design quality. Advanced packaging technologies like CoWoS and SoIC play a crucial role in enabling AI scaling, with memory bandwidth and interconnect performance scaling aggressively. TSMC addresses power delivery and thermal management challenges in AI systems through advanced solutions. TSMC's advancements in design methodologies and AI-driven automation promise improved productivity and scalability in chip-package co-design.

SemiWiki•

5 hours ago