The headline is bogus and most of the comments are responding to the headline. Google's emissions increased 13% since last year, "primarily due to increases in data center energy consumption and supply chain emissions." It's unclear how much is due to AI. The supposed surge is a 48% increase compared to *2019*, consisting of moderate increases every year since 2020, not a nearly 50% surge due to AI.
Your comment is a little dismissive. Firstly, the headline doesn't mention "versus last year", and the 48% figure is straight from the document you mentioned and is repeated often within that document. It's directly from Google, and is significant. Moreover that document itself states (page 31) "As we further integrate AI into our products, reducing emissions may be challenging due to increasing energy demands". It's clear that AI advancements are the biggest hurdle in terms of energy use and therefore CO2 emissions for Google.
The article does a good job of clearly stating that the 48% is compared to 2019, but words like "surge" and "spike" do not closely match that factual basis and imho are misleading.
Attributing the 48% to AI specifically is largely baseless though. From my skim, the Google report does not make any specific claims about the increase in datacenter emission coming from AI. The closest claims are on page 12 where "the rapid advancement of AI has brought necessary increased attention to its energy consumption and resource demands" is juxtaposed with "total data center electricity consumption grew 17%"
In particular the third "key point" seems highly misleading.
> The company attributed the emissions spike to an increase in data center energy consumption and supply chain emissions driven by rapid advancements in and demand for AI.
The word "spike" does not occur in the document and the 48% number is never close to a mention of AI. While the 17% "spike" may have been attributed to AI by Google, I think it is clear the document does not attribute the 48% to AI.
Claiming that emissions surged 50% due exclusively to AI (as in the headline) is unsupported by the Google report, which is the article's singular source.
Again, looking at the graph of emissions at Google on pg 31, it's clear that the increase is linear after a dip 2019-2020 and the report identifies supply chain issues as a major source of emissions in addition to datacenter electricity costs—again, notably not specifically calling out their AI training/inference costs as a reason for their increase. The report does identify AI as a challenge going forwards, however that's not the same as saying it's exclusively responsible for the last 5 years of emissions.
It's indeed pretty spun. Some quick googling shows that real GDP is up 11.5% from 2019-2024, which makes up a bunch of the "effect" right there (obviously datacenter power isn't directly tied to economic activity as a first principles thing, but like everything else at scale it's going to be pretty darn close).
Of the remainder, I get an average yearly increase of... 5.8% over the time period in question. I mean, that's an effect, but it's certainly not a big one.
> The supposed surge is a 48% increase compared to 2019*
That's a smoking gun, alright. If you compare today to just before the pandemic, you're going to get all kinds of weird results that can mostly be attributed to the world working in emergency mode for 2020-2022.
Let's not forget what happened in 2020: the world went into lockdown over COVID-19. Everyone was trying to do emergency gear-shift into remote work, and the demand for online collaboration services - and supporting infrastructure - exploded.
That IMO can easily account for majority of the increase since 2019, and it seems that this is another example of why you have to be suspicious of any average that includes pandemic years in its range. 2020-2022 was a unique period in all kinds of ways.
EDIT: Google's document even acknowledges that, in footnote 103 linked to from page 32. Quoting:
Although 2020 was the most recent emissions inventory available at the time the target was set, 2020 was deemed to not be representative of a typical year, because operations were impacted by the COVID-19 pandemic. The next most recent year with representative data, 2019, was selected as the base year.
"In 2023, our total GHG emissions were 14.3 million tCO2e [metric tons of carbon dioxide], representing a 13% year-over-year increase and a 48% increase compared to our 2019 target base year. This result was primarily due to increases in data center energy consumption and supply chain emissions. As we further integrate AI into our products, reducing emissions may be challenging due to increasing energy demands from the greater intensity of AI compute, and the emissions associated with the expected increases in our technical infrastructure investment."
"...reducing emissions may be challenging..."
The supposed justification for this level of pollution and water use [see page 42] is debatable. The expected profits are highly speculative.
What seems more clear is that unless things change it is going to to nigh impossible for young people to rise to challenges of climate change while simultaneously being addicted to mobile/Wifi internet access. Disconnecting a kid today from the internet is like separating a junkie from their drugs.
Gebru's estimates of AI energy consumption were adopted from Strubell. Strubell's estimates of AI energy consumption are ridiculous and according to David Patterson range from 88x to 10000x high. My humble opinion on the affair is that the conflict between Gebru and the rest of the company arose from Gebru's refusal to use the best and most comprehensive information on AI energy consumption, that is what they had access to firsthand as a privilege of being a Google insider.
Google's document is at: https://www.gstatic.com/gumdrop/sustainability/google-2024-e... See pdf page 8 / document page 7 for details, as well as the graph on page 32/31.