AMD’s Lisa Su just addressed one of the most important questions in AI infrastructure right now. Asked about readiness for roughly 3 GW of 2027 deployments, she was clear: “We have good visibility into land, power, shells, all of those, the data center infrastructure. We know where the GPUs are going in terms of which data centers and which locations.” That level of specificity matters. Power and site readiness have become real constraints on AI buildouts, and AMD is signaling it has line-of-sight into the physical infrastructure, not just the chips. Looking further out, Su said AMD is selectively investing alongside key partners to identify and secure data center capacity multiple years ahead of new silicon. Planning horizons that used to be 1–2 years are now stretching to 3–5 years. This comes on the back of a record quarter: Data Center revenue nearly doubled (+97% YoY), Q4 guidance implies strong sequential growth, and management is now guiding to >80% revenue growth in 2027 with Data Center expected to exceed 80% of total sales. The message is consistent: demand is outrunning supply, AMD is securing more capacity across the stack, and the company is working to de-risk the physical infrastructure side of the equation for the next wave of deployments. Power, land, and shells are becoming as strategic as wafers and packaging. $AMD