Enregistré dans:
Détails bibliographiques
Auteurs principaux: Cottier, Ben, Rahman, Robi, Fattorini, Loredana, Maslej, Nestor, Besiroglu, Tamay, Owen, David
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:https://arxiv.org/abs/2405.21015
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915142454738944
author Cottier, Ben
Rahman, Robi
Fattorini, Loredana
Maslej, Nestor
Besiroglu, Tamay
Owen, David
author_facet Cottier, Ben
Rahman, Robi
Fattorini, Loredana
Maslej, Nestor
Besiroglu, Tamay
Owen, David
contents The costs of training frontier AI models have grown dramatically in recent years, but there is limited public data on the magnitude and growth of these expenses. This paper develops a detailed cost model to address this gap, estimating training costs using three approaches that account for hardware, energy, cloud rental, and staff expenses. The analysis reveals that the amortized cost to train the most compute-intensive models has grown precipitously at a rate of 2.4x per year since 2016 (90% CI: 2.0x to 2.9x). For key frontier models, such as GPT-4 and Gemini, the most significant expenses are AI accelerator chips and staff costs, each costing tens of millions of dollars. Other notable costs include server components (15-22%), cluster-level interconnect (9-13%), and energy consumption (2-6%). If the trend of growing development costs continues, the largest training runs will cost more than a billion dollars by 2027, meaning that only the most well-funded organizations will be able to finance frontier AI models.
format Preprint
id arxiv_https___arxiv_org_abs_2405_21015
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle The rising costs of training frontier AI models
Cottier, Ben
Rahman, Robi
Fattorini, Loredana
Maslej, Nestor
Besiroglu, Tamay
Owen, David
Computers and Society
The costs of training frontier AI models have grown dramatically in recent years, but there is limited public data on the magnitude and growth of these expenses. This paper develops a detailed cost model to address this gap, estimating training costs using three approaches that account for hardware, energy, cloud rental, and staff expenses. The analysis reveals that the amortized cost to train the most compute-intensive models has grown precipitously at a rate of 2.4x per year since 2016 (90% CI: 2.0x to 2.9x). For key frontier models, such as GPT-4 and Gemini, the most significant expenses are AI accelerator chips and staff costs, each costing tens of millions of dollars. Other notable costs include server components (15-22%), cluster-level interconnect (9-13%), and energy consumption (2-6%). If the trend of growing development costs continues, the largest training runs will cost more than a billion dollars by 2027, meaning that only the most well-funded organizations will be able to finance frontier AI models.
title The rising costs of training frontier AI models
topic Computers and Society
url https://arxiv.org/abs/2405.21015