Hey, I'm a climate and energy writer / data analyst who focuses on corporate + govt accountability
Creator + curator of the Greensky feed: https://ketanjoshi.co/greensky/
Based in Oslo but write about US, Europe, Aus too - [email protected]
@ketanjoshi.co PinnedSo.......how the hell did the New York Times get Australia's AI data centre story so wildly wrong?
Albanese's new AI rules **WORSEN** their position by allowing more room for gas
Check out my new piece at @thepointau.bsky.social for all the lovely, gory details
@reneweconomy.com.au
![4. Discussion
This study makes three contributions. First, it extends the emerging concern around datacenter
scale and proximity to urban populations by showing that the location of the electricity generation
serving these facilities is itself a public-health consideration [1,6,31].
Second, the two-scenario comparison distinguishes the implications of electricity-supply pathway
from those of datacenter size: in the hypothetical matched-demand scenario, the modeled grid-supply
pathway produced greater additional exposure to PM2.5 and consequently greater excess annual
mortality that affects a large geographical area, whereas the BTM pathway concentrated mortality
closer to the datacenter. Moreover, in the real-world scenario, C2’s substantially greater capacity
produced greater absolute regional and near-site excess exposure to PM2.5 and excess mortality under
the specified operating assumptions. The generation-weighted sensitivity substantially reduced the
estimated grid-related mortality burden but did not change the matched-demand ordering or the much
greater near-site concentration of the BTM burden (Sec 3.4). Thus, the magnitude of the grid-related
burden is specific to how the electricity-system response is represented and, for the primary analysis,
to the 2023 TVA marginal generation response. It should not be generalized to regions with different
marginal generation mixes.
Third, the equity analysis shows that these spatial differences can translate into different population distributions of burden, with the largest BTM-attributable mortality share occurring in the
low-income, high-people-of-color stratum. These findings show why electricity demand or emissions
intensity alone cannot describe the health implications of datacenter power](https://cdn.bsky.app/img/feed_fullsize/plain/did:plc:66lbtw2porscqpmair6mir37/bafkreici5prebpmhzqj343bp7uxt2iwahuyh4txbzunkosbt75e42gbtpe)
