View

AI-centric workload investments drive growing demand for infrastructure, and most organizations continue to have a distinct strategy to address these workloads. High-performance infrastructure designed for AI — with processors, storage and networks optimized for machine-learning (ML) workloads — is required to efficiently take advantage of larger models, invest in lower-latency applications and leverage a greater diversity of datasets. 451 Research's AI & Machine Learning, Infrastructure 2023 study reveals that many organizations are struggling to meet the full scale of demand for AI capabilities. Discussion of this challenge has in recent years focused on graphics processing unit cost and availability, but the bottlenecks for organizations are significantly wider ranging.