Google has published its 2024 Environmental Report, an 80-page document that explains everything the giant is doing to apply its technology to environmental issues and mitigate its own contributions. But it completely avoids the question of how much energy AI uses. Perhaps because the answer is “much more than we want to say.”
The full report can be read here (PDF)There’s a lot of interesting stuff out there, honestly. It’s easy to forget how many plates a company as big as Google has to spin, and there’s some really remarkable work here.
For example, it is working on ~ Water replenishment programThis is done by identifying and funding watershed restoration, irrigation management, and other work in the area, and dozens of such projects around the world are at least partially funded by Google. In this way, we have reclaimed 18% of our water usage (as defined here) and are improving every year.
The company also does a lot to proactively provide AI’s potential benefits in climate, such as optimizing irrigation systems, creating more fuel-efficient routes for cars and boats, and predicting floods. We’ve already highlighted a few in our AI coverage, and they can actually be very useful in many areas. Google doesn’t have to do this, and many big companies don’t either. So give credit where credit is due.
But then we get to the section called “Responsibly Managing AI’s Resource Consumption,” where Google, which had been confident in all its statistics and estimates up to this point, suddenly throws up its hands and shrugs its shoulders. How much energy does AI use? Anyone? really Are you sure?
But that can only be a bad thing, because the first thing the company does is shrink the entire data center energy market. That is, it is only 1.3% of global energy use, and Google uses only up to 10% of that. So according to the report, only 0.1% of the world’s energy powers servers. That’s a small thing!
In particular, while the company committed in 2021 to achieving net-zero emissions by 2030, the company acknowledged there was “a lot of uncertainty” about how that would actually happen, especially since emissions have increased every year since 2020.
Our total GHG in 2023 [greenhouse gas] Emissions amounted to 14.3 million tCO2.2e, representative A 13% increase over the previous year and a 48% increase over the 2019 target year-on-year. This result is primarily due to increased data center energy consumption and supply chain emissions. As we further integrate AI into our products, the intensity of AI computing will increase, increasing energy demand and emissions associated with anticipated investments in technology infrastructure, making it difficult to reduce emissions.
(The quote below and this part is my emphasis.)
But AI’s growth has been marred by the aforementioned uncertainty. Google offers the following excuse for why it doesn’t specifically say how much AI workloads contribute to typical data center energy costs:
Predicting the future environmental impact of AI is complex and evolving, and our historical trends are unlikely to fully capture the future trajectory of AI. As we deeply integrate AI across our product portfolio, The distinction between AI and other workloads is meaningless, so we focus on data center-wide metrics. Because this includes the AI’s overall resource consumption (and therefore environmental impact).
“It’s complicated and evolving”; “It’s unlikely that trends will ever be fully captured”; “The distinction would be meaningless”: this is the language used when someone knows something but doesn’t really want to say it.
Does anyone believe that Google doesn’t know, down to the penny, how much AI training and inference add to their energy bills? Isn’t it part of the company’s core competency in cloud computing and data center management to be able to analyze those numbers so accurately? There are all these other statements about the efficiency of their custom AI server units, how they do all this to reduce the energy required to train AI models by a factor of 100, etc.
There is no doubt that Google is making great green efforts, and you can read all about them in the report. But it is important to highlight what Google is denying: the enormous and growing energy costs of its AI systems. The company is not a major contributor to global warming, but despite its potential, Google does not appear to have made a net profit yet.
Google has every incentive to downplay and hide this number, which is never a good number even when it’s downplayed and very efficient. Let’s ask Google more specifically before we figure out if it gets any worse in the 2025 report.
