Article
AI Infrastructure Where We're Going, We NEED Roads Networking is not the supporting act in AI clusters. It is the system itself. Last week in London, I stepped out of Sainsbury’s with a couple of Jellycats for my...
Article
AMD's challenge is architectural: their "world size" is limited to 8 GPUs per node, while Nvidia can interconnect 72. AMD's solution won't arrive until 2027, leaving Nvidia with a commanding lead for training and large-scale inference.
Media
The Future of Tech Leadership Discover the shocking truths and emerging trends that will define technology leadership in the years ahead The landscape of technology leadership is undergoing a dramatic transformation. As...
Media
Nick Hume interviews Philip Johnston, co-founder of Starcloud, about building large data centers in space. They explore how reduced launch costs and innovative cooling solutions make orbital compute viable, discussing current projects, space challenges, and AI's future beyond Earth.
Article
AMD MI355X Analysis Strong Node-Level Inference, but Not Yet Rack-Scale 288 GB HBM3E Memory 8 TB/s Bandwidth 1.4 kW Power Draw Quick Navigation 1. Overview 2. Performance Analysis 3. Memory Advantage 4. Interconnect...
Article
A recent Synergy Research Group report has revealed a striking statistic that should give every infrastructure professional pause: hyperscalers—AWS, Azure, Google, Meta, and their peers—now account for 44% of the world's total data center capacity. More importantly, this figure is projected to surge
Media
The artificial intelligence revolution is fundamentally transforming the infrastructure that powers our digital economy. In this exclusive Q&A session with Mark from Value3, our U.S. Tech Lead Nick Hume explores what's really driving the AI boom—from infrastructure challenges to strategic opportunit
Media
In this conversation, Mark and Nick Hume discuss the current state and future of the data center industry, particularly focusing on Nvidia's advancements in GPU technology and the implications for AI and data centers. They explore the challenges posed by rapid GPU iteration, the evolution of data ce
Article
During my deployment engineering days at Amazon, we were in the middle of what felt like controlled chaos. New AWS regions launching monthly, services expanding globally, data centers coming online faster than anyone thought possible.
NVIDIA's NVLink Fusion True Innovation or Strategic Lock-in? Earlier this week at Computex, Jensen Huang introduced NVLink Fusion, positioning it as a means to "democratize scale-up" by allowing customers to mix and...
