Published: 20 February 2024
Summary
Chinese enterprises prefer to adopt generative AI on-premises instead of via public cloud, due to regulation requirements and data privacy and security concerns. Because current infrastructure environments can’t support GenAI initiatives, I&O leaders in China must prepare for data center impacts.
Included in Full Research
Overview
Impacts
Deploying generative AI (GenAI) on-premises will force I&O leaders to change the way hosting environments are designed because of the limited availability of high-end graphics processing units (GPUs) in the Chinese market.
Deploying large GPU clusters on-premises will disrupt existing data centers because training the GenAI model requires high-throughput, low-latency and lossless infrastructure. It also requires renovating facilities to support upgraded infrastructure.
Recommendations
CIOs in China and their I&O leaders preparing infrastructure resources for GenAI adoption should:
Clients can log in to view the entire
document.