Why the desperate gas industry is suddenly obsessed with AI slop

You’d be hard to find an unhappy gas company CEO right now. The share prices of carbon bomb sellers like Woodside and Santos shot up after America and Israel attacked Iran in February this year. They and most other companies have seen windfall profits from the first quarter of this year and it’s certain to continue.

That will help buy another mansion this year, but what happens in the long term?

The second global fossil fuel crisis of this decade has sharpened the mind of those regions most reliant on these products, towards figuring out how to ditch reliance on them entirely. It isn’t just high costs: scarcity means Asian countries in particular are suffering from actual shortages of oil and gas. 

As Louise Morris covered here at The Point last week, Asia’s appetite for Australian gas is beginning to show signs of weakness. A new report from think tank Ember shows, for instance, that Japan’s demand for gas for its power sector has consistently fallen for 7 years, now. Of course it has: why would you rely on an expensive fuel so vulnerable to supply disruptions and cost spikes, that you have to source from another country? 

China, another of Australia’s main gas destinations, has spent 1.5 years in a plateau of gas-fired power generation. Once wind and solar growth become greater than demand growth in China, that’ll mean both coal and gas power generation start to fall. 

Surely Australia’s gas companies can turn to the domestic market: where people like Anthony Albanese insist gas is “needed” for Australia’s renewable energy revolution? Well, no. Not only did we already know that Australia’s gas extraction volumes are obscenely higher than what is used in the power grid, we now know that battery power is murdering Australia gas power far more than most of us ever expected. 

As the Grattan Institute highlighted in a detailed report recently, “as Australian households and businesses search for cheaper, cleaner, and more efficient fuels, they are using less gas”. It is depicted in dramatic form: 

As Grattan correctly state, “less demand for gas means more demand for electricity. Without integrated planning, consumers and taxpayers are exposed to the risk of over-investing in gas and under-investing in electricity infrastructure”.

This is where the latest glimmer in the eye of the gas industry comes into play. As I detailed in a recent report I led for Greenpeace Australia Pacific, the gas industry sees a real saviour in the prospect of the tech industry, and oversized data centres being massive demand sinks for all of their oversupplied fossil gas. 

In Canada, the UK, Ireland, parts of Europe and the US, there is a massive, concerted push for the direct use of gas to power the impossibly gargantuan data centres being built to power the generative AI slop boom. It’s tempted to the tech companies because it gets them online quicker, they don’t have to deal with the messy wait times and power price politics of the grid, and there are growing relationships between fossil fuel companies and the tech industry. 

In Australia, I found a wide range of obscene examples. The “Beetaloo compute zone” (seriously) for instance, is a proposed five gigawatt data centre (that’s 5x bigger than the largest formal application in Australia) that I estimated would have annual emissions of around 25 megatonnes of carbon dioxide equivalent. So if it became operational, it would double the emissions of the Northern Territory. 

 

View this post on Instagram

 

A post shared by Ketan Joshi (@ketanjoshi19)

I also dug into the “Cloud Carrier” project in New South Wales, which is proposing a gas plant the same size as the recently built Kurri Kurri plant, a site that would use so much gas that it would be “25% more than NSW’s FY24 residential gas demand of 31.3 [petajoules, PJ], and 74% of the Narrabri Gas Project’s anticipated annual production of 53 PJ”.

The head of Australia’s data centre lobby group featured as a guest on a panel at the Australian Energy Producers conference last month – a panel sponsored by Chevron. That event also featured a presentation titled: “Australia’s AI future depends on gas”. There are many more examples in the full report, here.

Both the tech industry and the fossil fuel industry are enacting a sort of loud, public wish-fulfilment here. But declaring these wishes loudly can be influential. Before the LNG boom, there were loud calls for Australia to rush into LNG exports to avoid missing its chance, and those calls worked – Australia suffered massive local impacts while all the benefits flowed to private operators. The end of that boom is tied inextricably to the start of this one.

Hopefully, that hope is hollow. I think data centre growth should be restricted, aggressively, by governments that work to protect people from the private interests of corporations we know, for sure, do not care for their local or aggregate impacts. There are calls to manage unchecked AI data centre growth; such as through a renewable energy mandate or increased taxation. These are good ideas, but if they don’t come with direct restrictions on data centre hype they’ll just scrape around the edges of the core problem: two corrosive industries hoping to exploit Australia a second time around.