At Salesforce Inc.’s Dreamforce event in San Francisco, tech titans debated whether AI deployment should be slowed, while about an hour’s drive south at the AI Infra Summit in Santa Clara, experts from Amazon Web Services, Oracle, Broadcom, Qualcomm and d-Matrix described building AI infrastructure as fast as possible. Qualcomm’s Tony Pialis said “Tokens per watt has become the new key metric in this AI war” and that the industry cannot stay on the current trend. AWS senior vice president Peter DeSantis said “The majority of compute is going to be serving inference,” calling inference a massive workload; AWS’ Graviton family of 64-bit Arm-based CPUs, including the Gaviton5 CPU launched in June, supports real-time AI reasoning and multistep task orchestration. Memory is increasingly critical: a typical AI server uses roughly eight times more memory than a traditional server, and AI server memory spend is projected to jump from $35 billion in 2025 to between $175 billion and $190 billion by 2027, roughly a fivefold increase.