A new way to optimize and prioritize AI projects for the GPU shortage
Summary
The increasing number of AI startups and services has led to high demand for high-end GPUs such as A100s and H100s, thereby overwhelming Nvidia and its manufacturing partner TSMC, both of whom are struggling to meet the supply. At a recent off-the-record meeting in London, OpenAI’s CEO Sam Altman candidly acknowledged that the computer chip shortage is stymieing ChatGPT’s advancement. Altman reportedly lamented that the dearth of computing power has resulted in subpar API availability and has obstructed OpenAI from rolling out larger “context windows” for ChatGPT. On the one hand, product leaders find themselves caught in a relentless push to innovate, facing the expectations to deliver cutting-edge features that leverage the power of gen AI. In my experience, including my early days leading digital transformation at a healthcare company and later while working with various McKinsey clients, this approach has been a game-changer in scenarios where capacity constraints are a critical factor.
Classifications
AskAI Classifications
LabelsLinked Companies
