Part 1 of this series explored Meta’s existing climate footprint and data centres plans. Read it here
The “Greenhouse gas protocol” (GHGP) is a global, voluntary system of emissions measurement that was created several decades ago by a coalition of high-emitting companies (including Monsanto, British Petroleum and General Motors), in partnership with the World Resources Institute, eager to seize the narrative on climate.
This collection of eye-watering technical guidelines covers the direct emissions of companies, the emissions associated with power consumption, and the indirect emissions of whatever’s bought or sold by the company in the going about of its normal business (think: the steel that goes into a car they build, or the emissions released when a lump of sold coal is burned). All of these are undergoing a massive, lengthy and rare review, giving companies a new chance to push things in their desired direction.
For the climate impact of drawing electrical power, there is a welcome recognition that the current state of affairs (as I wrote about earlier this week) is pretty grim. The worst underreporting of climate impacts has come from the companies with the most significant expansionist agenda.
Related Article Block Placeholder
Article ID: 1223108
There are two broad camps here. One camp, generally led by Google, is proposing a shift towards the higher-resolution measurement of grid emissions, and the matching of a company’s power consumption to grid intensity with significantly more granularity. Known as “hourly matching”, this features some strange logic that whoever helped the renewables get built gets to claim its power, but it does so in a way that’s less bad than the current system of claiming renewables from any project, anywhere in the world.
Independent. Irreverent. In your inbox
Get the headlines they don’t want you to read. Sign up to Crikey’s free newsletters for fearless reporting, sharp analysis, and a touch of chaos
By continuing, you agree to our Terms & Conditions and Privacy Policy.
The other camp is led by a coalition of companies helmed by Amazon and Meta, known as the “Emissions First Partnership”. Instead of proposing a small improvement to a bad system, they are pushing a material worsening of a bad system. Under their framework, they calculate the emissions avoided by the renewable energy project, and subtract that from their measured greenhouse gas emissions.
This isn’t the “weird cousin” of carbon offsetting. This is carbon offsetting. And it includes the single most catastrophic feature of carbon offsets: the hazy, fuzzy assumption of a counterfactual, which can be tweaked and fudged to claim some absurd doomsday scenario was avoided thanks to an action (and therefore, a lot of “climate action” can be claimed). While companies have moved away from explicit carbon offsetting, Meta wants to bake it so deep into the measurement philosophy that there is no clear way for the time-poor reader to tell what is happening. It is far, far cheaper to juice counterfactuals, where you are limited only by your imagination, than to have to procure real power purchasing deals with clean power sites (already slowing down in the US), or match your power demand with output at a high resolution.
In Australia, you can see the echoes of this global debate in little-known changes to the domestic renewable certification system. The new “guarantees of origin” come stamped with more information, such as when and where they were generated, allowing companies to choose to only buy ones matched to where their demand occurred. But nobody has to do this; it’s completely optional. As I reported here at Crikey, Australia’s domestic data centre industry is presenting itself as the cure to the growing ailment of being wildly off track for a target of 82% clean power by 2030, primarily through direct power purchasing agreements. But any changes to the GHGP will affect companies directly, particularly Amazon, whose electrical presence in Australia is growing fast.
Big tech is being pulled in a few directions at once. In the US, where most tech companies are headquartered, cosying up to a fascist, climate-denying government means making hollow, technosolutionist noises about carbon capture and nuclear fusion. But for the rest of the world, and particularly in Europe, climate change still exists. Now that these companies have become existentially dependent on data centre expansionism, required to power generative systems that require far more power to deliver worse results, they have to refresh their greenwashing techniques, and that means eroding measured reality from the ground up.
It is not clear what any of this is meant to actually be for. It was recently revealed that Meta explicitly allows its chatbots to develop “sensual” relationships with children, and offer fabricated medical misinformation. Meta illegally downloaded a torrent of millions of books (including mine), to help train its models. Private chats with chatbots were inadvertently made public, revealing sensitive information. Meta’s chatbots are designed to present as real humans, resulting in the rapid spread of mental health crises and even some tragic deaths.
I am trying to think of a single, substantiated and significant societal benefit of Meta’s generative AI investment, and I am pretty confident that it simply does not exist.
A Financial Times analysis of regulatory filings recently found that “most of the anticipated benefits [of generative AI], such as increased productivity, were vaguely stated and harder to categorise than the risks”. The only companies with any real upside were energy companies: specifically, Entergy, the supplier of Meta’s massive new fossil-fuelled power station. The FT also found this nugget in Meta’s 10k form in 2024: “We may not be successful in our artificial intelligence initiatives, which could adversely affect our business, reputation, or financial results”.
If there is a neat through-line in the history of climate activism, it is the ongoing fight to connect human society to physical reality. First, the messy struggle to nurture the acceptance of the basic science of the problem, and more recently, the struggle to get everyone who says they accept the science to at least fucking act like it. With big tech, that struggle continues in the new arena of how we measure and report on who’s committing the worst climate sins.