OCP stories
Oxford optical AI spin-out Lumai names ex-Palo Alto Networks marketer René Bonvanie as board chair to spearhead its push into commercialisation.
AI chipmakers can cut die-to-die link area, power and I/O by up to 50% as multi-die designs become harder to scale.
The shift could help AI data centre operators add denser racks without rebuilding sites, as NVIDIA moves to cut power-conversion losses.
Rising electricity demand is emerging as a bottleneck for AI deployment, with some training sites approaching 1 GW and facing delays.
European firms needing secure, low-carbon compute gain a bigger home as the deal expands cloud, AI and dedicated HPC services under one roof.
Data centres and research labs could cram larger AI models and simulations in memory, with Dell's new rack scaling to 144 GPUs per rack.
The rack-ready system targets organisations needing denser, liquid-cooled infrastructure as AI and scientific computing demands surge.
Demand is being lifted by edge and AI workloads, with the market forecast to more than double to USD $4.32 billion by 2030.
The move should cut AI inference costs for Zoho while giving the software group tighter control over data, power use and its infrastructure stack.
The in-house platform is meant to lower AI inference bills by 20-30% and trim data-centre power use as Zoho tightens control over its stack.
Demand for AI clusters is pushing networking hardware to the limit, and Marvell says its new chip could cut rack power use and latency.
The move could help ease bottlenecks as huge AI clusters struggle to keep hundreds of thousands of GPUs synchronised during network faults.
Rising AI power and cooling demands are pushing operators towards open hardware as Legrand adds rack, power and thermal gear for dense sites.
Siemens and Rittal forge IEC-focused alliance to standardise high-density AI data centre power, debuting modular sidecar rack systems.
HPE boosts its AI and supercomputing line-up with new NVIDIA-powered Cray blades and AI Factory systems for trillion-parameter models.
AMD and HPE will launch the Helios AI platform in 2026, combining CPUs, GPUs and advanced networking to power large-scale data centres worldwide.
Dell launches modular AI rack systems with 60% cooling savings, featuring PowerCool eRDHx and open standards for efficient enterprise deployments.
AMD unveils Helios, a rack-scale AI platform built to Meta's new Open Rack Wide standard, designed for high-performance AI and HPC data centres.
Dell launches the IR9000 rack and PowerCool eRDHx cooling system to enhance AI efficiency with up to 60% reduced cooling costs and open-standard design.
Lower power use and total cost of ownership are central to Zoho's new server, aimed at cutting AI inference costs and tightening data control.
.webp)
