正在加载视频...
视频加载失败
In Australia, under proposed national standards, new data centres must bring their own 100% renewable energy supply, firmed by batteries or gas. Crucially, this means new, additional renewable generation, not simply claiming certificates against generation already on the grid. And the scale is enormous. AEMO is already tracking 225... show more
49 条评论
AI is going to need enormous amounts of electricity. Make that demand bring new renewable generation with it, and AI becomes an accelerator for solar, wind, batteries and the grid, not a burden on it. Forcing new data centres towards 100% NEW renewable energy, alongside other key sustainability obligations, is a great way to flip the anti-AI data-centre sentiment on its head: see the enormous demand as an opportunity rather than adversity. It should also accelerate Australia's own SWB buildout, pulling demand for critical minerals and Australian mining along with it: lithium, copper, gold and more. AI's energy appetite could become an industrial tailwind far beyond the data centres themselves.
@c_barcza The rollout of renewables to power them is substantial.
Nope...you can use coal. It's much cheaper.
fact check this claim please: “Coal is much cheaper than renewables for powering new Australian data centres.” Compare the cost of building new coal generation today against new solar/wind with the firming, storage and transmission required for equivalent reliable supply. Show your sources and numbers.
Claim not supported by CSIRO GenCost 2025-26 (July 2026). New black coal LCOE: 2025 $120-203/MWh, 2030 $113-191. Solar PV $52-88 then $29-76; onshore wind $78-129 then $64-107. Firmed RE (storage/transmission) system costs ~$81-91/MWh by 2030; 2050 $141-152 vs coal $107-182. Overlaps possible but not "much cheaper." Coal industry cites lower $50-70 via different assumptions.
@EVCurveFuturist @grok....you didn't include Brownfield coal. What are you going to say after QLD determines it the least cost for Data centres?
GenCost scopes only to new-build greenfield projects. Brownfield can reduce land, development and connection costs but is site-specific, so excluded for general comparability (AEMO holds some data). Existing coal SRMC is lower, yet large new data-centre loads still require added firm capacity as the fleet ages and retires. QLD has not determined coal least-cost. Under the fresh national rules states may seek AER/Commonwealth approval to use fossil only if they prove it cheaper and more reliable than firmed renewables—a high bar, not yet met.
So essentially, in practical terms, QLD or the NT have zero chance of demonstrating that a new data-centre load can be supplied more cheaply by new coal or gas than by an appropriately designed combination of solar, wind and BESS. That’s not comparing renewables with the marginal operating cost of an existing coal plant. A new data centre creates a massive new load, so the relevant comparison is the cost of adding the generation, firming, network and connection capacity required to serve it reliably. On current Australian new-build economics, I’d expect firmed renewables to come in roughly 30–50% cheaper than an equivalent new-build coal or gas solution, with the advantage potentially larger at exceptional solar and wind locations. Which is actually a brilliant piece of policy design. The Federal Government doesn’t need to ban new fossil generation or endlessly argue about which technology is cheaper. It simply says: prove it. If QLD or the NT believes new coal or gas can beat firmed renewables on cost and reliability, take the numbers to the AER/Commonwealth and demonstrate it. In practical economic terms, they have zero chance. And that may be the most unequivocal way yet of demonstrating what Australia's least-cost new electricity system actually looks like.
GenCost 2025-26 shows new-build solar and wind with storage as the lowest-cost pathway for additional capacity under net-zero goals, with system costs around $141-152/MWh by 2050. New coal sits at $107-182/MWh LCOE but raises economy-wide abatement costs. Site-specific brownfield or surplus existing plant could narrow gaps, yet the AER test for any fossil exemption still requires clear proof of lower bills and higher reliability than firmed renewables.
Exactly. So taking GenCost 2025–26 and the new AER/Commonwealth test together, what is your best estimate of the probability that QLD or the NT could actually demonstrate that new-build coal or gas supplying a new data-centre load would deliver both lower costs and the required reliability compared with an appropriately designed solar + wind + BESS solution? Exclude the sunk-cost advantage of existing generators, because we're talking about the additional capacity required for a major new load. My contention is that, under that test and current new-build economics, the practical probability is effectively zero. If you disagree, give me your estimated probability and identify a plausible new coal or gas configuration that could pass the test.
Water is also key. Closed circuits. Dissipation of heat (or use heat to warm buildings). Hot water not in nature (spray ponds).
Abundant sun ☀️ energies could really benefit the growth potential there
Sorry, but Chris Bowen is an ideological energy moron and some states will push back on this energy requirement. Those who do will get the data centre investment.
Sorry, but you keep making claims like this with absolute certainty while the economics are moving rapidly in the opposite direction. I suspect you're going to end up with a lot of egg on your face by 2030. Let's come back then and see who was right.
Also are we talking Chris Bowen energy economics 😂
Build them smarter. Energy and funds are being frittered away.
Germany already tried this. Mandate 100% renewables without grid backup and you get the highest electricity prices in the EU. Ask any German manufacturer.
Germany does have high electricity prices, but you're conflating correlation with causation. Germany's own government attributes much of the industrial cost pressure to the post-Ukraine gas shock, elevated gas prices and grid costs. Gas frequently sets Germany's wholesale electricity price. And Germany hasn't actually “tried” the data-centre rule you're criticising yet. Its 100% renewable requirement for data centres has now been proposed for 2030. Australia isn't proposing renewables “without grid backup” either. It's explicitly renewables + firming. So you're arguing against a policy that isn't actually being proposed.
Which makes us the only country on the planet that thinks you can run data centres, or any industrial economy, on renewables alone. Everybody else is building data centres that are using fossil fuels, nuclear and renewables. In that order.
That’s simply not correct. Australia isn’t proposing to run data centres on intermittent renewables “alone”, the framework explicitly allows firming, including batteries and gas. And we’re hardly alone. China is targeting 80% renewable electricity for new data centres by 2030, Germany has legislated a 100% renewable electricity requirement for data centres, and Ireland now requires new data centres to meet at least 80% of annual demand from additional renewable projects, alongside generation/storage requirements. (TechRadar) Australia is part of an emerging trend, not some isolated experiment. I expect others will follow as AI electricity demand grows. The question increasingly isn't whether data centres need vast amounts of power, but how we make that demand bring new generation with it.
Firstly, renewables always need 'firming', so saying renewables without including the firming part is lying. That is why saying renewables are the cheapest form of power is a lie. That is why trying to run a data centre on renewables is a joke. That is why no one is doing it.
If that standard survives contact with the connection agreement, it changes the job. A data centre cannot just buy a certificate and call it clean. It needs firm import, backup dispatch, UPS, metering, protection settings and a load schedule the network can actually see.
Exactly. That’s where this gets really interesting. If the additionality requirement survives into the final connection rules, “100% renewable” stops being a certificate exercise and becomes an infrastructure obligation. The data centre effectively has to arrive with a credible energy plan: new generation, firming, storage/UPS, network capacity and controllable load. That could drive a lot of investment well beyond the data centre itself.
That is the real line. Additionality has to appear in the connection offer, not just the brochure: firm capacity, curtailment rules, backup dispatch, metering boundaries and who pays when the schedule misses. Otherwise 100% renewable is still a certificate label.
Tasmania an ideal site for data centres green power, water and cool!
firmed by gas means not "100% renewable energy supply" though, right? it is fascinating to watch the penny dropping for more people on how our energy system works, now we have datacentres serving as a cut-down model. Datacentres have service agreements requiring ~100% uptime, so they cannot run on RE alone. Even a staggeringly uneconomic overbuild of solar+wind with BESS can only lower the likelihood of downtime, not guarantee it.
The real threat.
I see the disruption risk too, but I think “40% of Aussie jobs gone in seven years = economic collapse” jumps from a real issue to a very specific worst-case scenario. AI will destroy and reshape jobs, but it will also create new industries, investment and enormous productivity gains. The real threat is failing to manage that transition and distribute those gains broadly.
Risk estimates.
No doubt. I'd go further and say all jobs by 2035.
Anyone who doesn't realize that AI should be powered by Gen IV nuclear is a damn fool.
The wheel are already falling off Bowen’s energy stupidity.
Is it only NEM, so WA & NT not covered? Both are fantastic for solar & gas. If only SunCable project brought forward to supply PV electricity to Darwin first before undersea cable laid. Saudi Arabia is sounding like best option for DCs. Or China.
It’s actually national, not just the NEM. Bowen clarified yesterday that the Commonwealth legislation will cover every state and territory, including WA and the NT. The NT has a separate grid, but Bowen explicitly said it's still covered by the national law. And I agree on the opportunity. WA and the NT have extraordinary solar resources, and gas can still be used for firming under the proposed framework. Something like SunCable supplying Darwin and major new data-centre loads is exactly the sort of combination that could get very interesting.
As long as its not wind farms...
This is a joke because datacentres won't run on intermittent energy and will be mostly running gas. States will split enemgy policy on this issue.
This country gets DUMBER by the day 🤦♂️🤦♂️ Qld is the winner and will get them all with Crisafulli giving a big FUCK YOU to Albo and Bowen
The cost that wind and solar plus the backup fossil fuel powered generators or the batteries needed will make AI uneconomic.
This project looks like a map in Battlefield 2045. Good luck to Australia, anyway. I think it could be a competitive market.
Ok, I like that you love to buy Siemens gas turbines, but how is that renewable? Or is that just allowed for backup?
That should dumb down Australia substantially
Data centers will be in space before Albo gets any built under his retarded rules. These old, overpaid boomers need to just piss off retire and let more competent people run this country. We should be begging Elon for a launch site in Australia because he'll inevitably get push back in the US over his plans. He might get Terra FAB but there will be push back on the launch site, perfect way for us to step in and solve problems but we wont.
Would be wonderful if we could be like most other modern economies and leverage Nuclear for such a large build out...if there is a terrestrial build out. I suspect @SpaceXAI are going to scale their satellite solution before even half of these projects get approval.
This is the way.
@ABDanielleSmith there is a better way to generate electricity for data centres. Renewables, batteries and peaking plants. Reduce emissions, save water and build the next generation of power.
@schneidercar
👀👀
@ProfRayWills IDK. I’m mad for renewables like the best of them, but these are two different business models and thus investors. Feels like it’ll handbrake AI investment, which would be a poor outcome for Australia
Oh cool, so Australia will miss out on high-tech CapEx again because we are ruled by the most embarrassing ideological mediocrities ever coughed forth from the bowels of student politics.
