On February 17, 2026, SoftBank and Ampere Computing launched a joint project aimed at enhancing the efficiency of running small AI models on CPUs, with plans to create a low-latency, high-efficiency inference environment as a key technology for next-generation AI infrastructure.
