Ex-Googler joins Crusoe Cloud, neocloud of Stargate fame

The rapid expansion of artificial intelligence necessitates a fundamental shift from general-purpose cloud infrastructure to purpose-built "neocloud" environments. Erwan Menard, Senior Vice President of Product Management at Crusoe Cloud, highlights that AI workloads demand significantly higher power density—up to 120 kilowatts per rack—and specialized networking to manage latency. Unlike hyperscalers, neoclouds integrate energy sourcing, data center design, and software optimization to reduce costs and improve time-to-market for production-scale AI. As inference replaces training as the primary AI workload, efficiency in token throughput and energy consumption becomes critical. Addressing the global energy bottleneck requires innovative strategies, such as modular data centers and partnerships with sustainable energy providers, to support the massive power requirements of modern AI factories. This vertical integration allows for greater control over the infrastructure stack, enabling enterprises to scale AI agents effectively while managing the underlying hardware and energy constraints.