KI-Chips: Bedenken hinsichtlich des Energieverbrauchs großer KI-Modelle, GPU-Technologien und die Umstellung auf kundenspezifisches Silizium

Frontier artificial intelligence (AI) attracts global attention, with governments and hyperscalers racing to lead in domains like drug discovery and autonomous infrastructure. GPUs and other AI chips have been instrumental in driving this growth, providing the compute needed for deep learning within data centers and cloud infrastructure. However, the capacity of global data centers is expected to reach hundreds of GWs in the coming years, and investments into AI data centers are within the hundreds of billions of US dollars, which has put concerns about the energy efficiency and costs of current hardware into the spotlight.

 

This webinar will highlight the key challenges faced within the AI chips market. IDTechEx will explore graphics processing units (GPUs), the leading technology used across data centers and cloud infrastructure for training and inference of AI, specifically highlighting key trends that facilitate their dominance. This will be followed by a discussion on other AI chips, with this webinar focusing primarily on custom AI application-specific integrated circuits (ASICs), which have been adopted by hyperscalers and cloud service providers (CSPs) alongside GPUs. Attendees will gain insights into hardware and market trends, which underpin the AI chips market for data centers and cloud computing.

 

Discussion points in this webinar include:

 

This article shares some of the research from the new IDTechEx report, "AI Chips for Data Centers and Cloud 2025-2035: Technologies, Market, Forecasts".