AMD takes on Nvidia with its Helios AI rack scale system
THE SO WHAT
Rack-scale systems like AMD’s Helios shift the conversation from chips to full-stack clusters—power, networking, and software tuned as a unit. If you’re building serious training capacity, vendor selection now means picking an ecosystem and reference architecture, not just a GPU SKU.
READ THE SOURCE
MORE FROM THE WIRE
Applied AIChatGPT Health Rolls Out to Everyone While OpenAI Stares Down Major Lawsuits
Rolling out a health-branded assistant while facing lawsuits over allegedly fatal advice moves AI triage from a UX question to a liability and governance question. If you’re deploying AI into any regulated or safety-adjacent domain, you need explicit guardrails, escalation paths, and insurance in place before usage scales, not after.
Applied AIAgentic coding goes hands free as OpenAI brings GPT-Live's full duplex voice control to Codex and ChatGPT on the desktop
Full-duplex voice in desktop ChatGPT and Codex turns coding and ops into a continuous spoken interaction loop—closer to pairing with a junior engineer than issuing discrete prompts. Teams should start experimenting with voice-driven workflows for debugging, code review, and incident response where keyboard friction is currently the bottleneck.
Applied AIIntel Forecast Shatters Estimates on Data Center Growth
Intel’s upside forecast on the back of data center demand shows AI workloads are lifting even lagging incumbents. If you’re multi-sourcing compute, this is a window to extract better terms from non-Nvidia vendors before capacity tightens further.
Applied AIAMD unveils Helios rack and MI455X accelerator in bid to close the gap with Nvidia
AMD’s MI455X plus Helios rack is a bid to sell you an entire AI datacenter lane—72-accelerator racks tuned for training and inference, not just chips. If you’re locked into Nvidia by default, rerun your TCO and performance models with rack-scale AMD in the mix; competition at this level can move your 3–5 year capex curve.