Google has released its 2024 Environmental Report, an 80-plus page document detailing all the company’s major efforts to use technology and reduce its contribution to environmental issues. But it completely dodges the question of how much energy AI is using — probably because the answer is “more than we’d like to say.”
You can read the full report here (PDF)And honestly, there’s a lot of interesting stuff in there. It’s easy to forget just how many plates a big company like Google keeps spinning, and there’s some really remarkable work here.
For example, this one is working on Water Replenishment ProgramIt hopes to compensate for the water it uses in its facilities and operations, ultimately leading to a net positive outcome. This is done by identifying and funding watershed restoration, irrigation management and other actions in the area, with dozens of such projects around the world funded at least in part by Google. In this way it has reclaimed 18% of its water use (by whatever definition of that term is used here) and is improving every year.
The company also takes great care to advance the potential benefits of AI in climate, such as optimizing water systems, creating more fuel-efficient routes for cars and boats, and predicting floods. We’ve already highlighted some of these in our AI coverage, and they could actually be quite helpful in a number of areas. Google doesn’t need to do all of this, and many larger companies don’t either. So credit must be given where credit is due.
But then we get to the “responsibly managing AI’s resource consumption” part. Here Google, which was so confident about every statistic and estimate until now, suddenly throws up its hands and shrugs. How much energy does AI use? Can a person use AI’s resources? In fact Are you sure?
Yet this must be worse because the company first underestimates the entire data center energy market, saying it accounts for just 1.3% of global energy use, and the amount of energy used by Google is only 10% of that at most – so according to the report, only 0.1% of the world’s total energy is powering its servers. A minor point!
Notably, in 2021, it set out that it wants to reach net zero emissions by 2030, though the company admits there is a lot of “uncertainty,” as it likes to call it, about how that will actually happen. Especially because its emissions have increased every year since 2020.
In 2023, our total GHG [greenhouse gas] emissions were 14.3 million tCO2e, representing An increase of 13% year-on-year 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 integrate more AI into our products, reducing emissions may be challenging due to the more rapidly growing energy demands of AI compute and emissions associated with expected growth in our technology infrastructure investments.
(The emphasis in this and the quote below is mine.)
Still, the development of AI is lost amid the above uncertainties. Google has the following excuse for why the company is not specific about the contribution of AI workloads to its typical data center energy bill:
Predicting the environmental impact of AI in the future is complex and evolving, and our historical trends may not fully reflect the future trajectory of AI. As we integrate AI deeper into our product portfolio, The distinction between AI and other workloads will not be meaningful. Therefore, we are focusing on data center-wide metrics Because these include the overall resource consumption (and therefore, environmental impact) of AI.
“Complex and evolving”; “trends may not be fully captured”; “differences will not be meaningful”: this is the kind of language you hear when someone knows something but doesn’t really want to tell you.
Does anyone really believe that Google doesn’t know how much AI training and inference adds to its energy costs? Isn’t being able to break down those figures with such precision part of the company’s core competency in cloud computing and data center management? It has all these other statements about how efficient its custom AI server units are, how it’s doing all this to reduce the energy needed to train AI models by a factor of 100, and so on.
I have no doubt that there are a lot of great green efforts going on at Google, and you can read about them in the report. But it’s important to highlight what it refuses to do: the huge and growing energy costs of AI systems. The company may not be the primary driver of global warming, but despite its potential, Google doesn’t appear to be in a net positive position just yet.
Google has every incentive to understate and obfuscate these figures, which, even in its low, highly efficient state, can hardly be good. We’ll be asking Google for more specifics before we find out if they get even worse in the 2025 report.
googles-environmental-report-pointedly-avoids-ais-actual-energy-cost-techcrunch
