Intel and AMD jointly announced new AI-optimized chipsets aimed at doubling inference speed and cutting power consumption by 35% in data-center workloads. The releases, part of their Q4 2025 roadmap, strengthen competition against NVIDIA’s Hopper line and cater to hyperscalers deploying large-scale AI models globally.