AMD bets AI PCs cut enterprise cloud token costs

AMD is positioning enterprise AI PCs as a critical layer of hybrid AI infrastructure, as CIOs grow concerned about the rising costs of agentic AI cloud workloads. Senior VP Rahul Tikoo says tokenomics has become a C-suite priority, driving demand for local AI processing. AMD's new Ryzen AI Halo platform, featuring up to 128GB of memory, allows organizations to run large language models and agentic workflows locally, reducing cloud token expenses while improving security and latency. Tikoo says models once limited to frontier cloud systems can now run effectively on endpoint devices.