Ama­zon is the largest emit­ter of the big five tech com­pa­nies by a mile – the emis­sions of the sec­ond-largest emit­ter, Apple, were less than half of Amazon’s in 2022. How­ev­er, Ama­zon has been kept out of the cal­cu­la­tion above because its dif­fer­ing busi­ness mod­el makes it dif­fi­cult to iso­late data cen­ter-spe­cif­ic emis­sions fig­ures for the company.

As ener­gy demands for these data cen­ters grow, many are wor­ried that car­bon emis­sions will, too. The Inter­na­tion­al Ener­gy Agency stat­ed that data cen­ters already account­ed for 1% to 1.5% of glob­al elec­tric­i­ty con­sump­tion in 2022 – and that was before the AI boom began with ChatGPT’s launch at the end of that year.

AI is far more ener­gy-inten­sive on data cen­ters than typ­i­cal cloud-based appli­ca­tions. Accord­ing to Gold­man Sachs, a Chat­G­PT query needs near­ly 10 times as much elec­tric­i­ty to process as a Google search, and data cen­ter pow­er demand will grow 160% by 2030. Gold­man com­peti­tor Mor­gan Stan­ley’s research has made sim­i­lar find­ings, pro­ject­ing data cen­ter emis­sions glob­al­ly to accu­mu­late to 2.5bn met­ric tons of CO2 equiv­a­lent by 2030.