CPU stories
As AI workloads shift from training to inference, firms risk stranded kit unless they build systems that can adapt over five to seven years.
The update lets Windows users run bigger AI models on-device, as AMD bids to make local processing a rival to cloud-based services.
The chip aims to ease power and cooling constraints for operators seeking AI systems with tighter supply chain control and domestic manufacturing.
AI training is exposing network bottlenecks that can idle GPUs, raise costs and force operators to redesign data centres around lossless fabrics.
Local AI features could run faster, stay private and work offline as chipmakers push laptops with dedicated neural processors.
Data centre operators could cut energy use and rack space as Fujitsu's new 2nm chip targets AI inference with air cooling.
The top-end iPhone range gains AI features and a variable-aperture camera, as Apple targets better photos, battery life and performance.
The deployment will help AWS meet surging demand for AI computing as the cloud giant deepens reliance on NVIDIA hardware and robotics.
Faster on-device AI and new connectivity are now central to Apple's smallest desktop, with models aimed at creators and developers.
Mac users tackling AI and graphics-heavy work will get more local muscle, as Apple's latest desktop silicon pushes memory and compute higher.
It lets cloud operators pair custom XPUs with NVIDIA's rack-scale systems, aiming to cut delays and integration costs in AI build-outs.
Enterprises could cut GPU use and power costs as the open-source bundle lets Akida neuromorphic chips share IBM's scheduler for inference tasks.
Linux gamers will get smoother access to streamed titles, as Nvidia moves GeForce NOW out of beta and trims cloud gaming lag.
The pay-per-minute service could cut testing costs for teams that need to check apps across many handset types and screen sizes.
Organisations running Java in production will get faster fixes for serious flaws as Azul moves to monthly security-only updates from August 2026.
The deal gives Qualcomm a stronger software layer for developers as AI workloads spread from edge devices into data centres.
Retailers and venue operators can now deploy Philips screens without an external media player, as PPDS adds BrightSignOS and built-in AI support.
Enterprise buyers in Saudi Arabia will get a locally hosted AI PC from September 2026, alongside Adobe's shift to sovereign workloads.
The deal underscores how scarce power and land have become for AI expansion, with OpenAI planning a huge Ohio facility over 20 years.
Australian buyers are weighing convenience against control as desktop PC demand rises, with local warranties and faster access swaying choices.
