Summary Insights: Infrastructure Considerations for Generative AI

RFG Perspective: Unlike the "cloud first" directives, which were viewed as saving operational costs, a "generative AI first" directive will add to one’s data center operation's Capex and OpEx. Generative AI (genAI) models are computationally intensive, necessitating robust processors, large datasets, and sophisticated data handling. Those AI “engines” may be located on-prem, or at a colocation or cloud provider site, but they are there, and will not be inexpensive. The recent versions of ChatGPT and other genAI models have advanced exponentially over prior attempts. They now offer the potential to become a force multiplier for productivity – playing a variety of roles by supporting work being done by different personas within the company.

Summary Insights: Infrastructure Considerations for Generative AI