A new report from Greenpeace Australia Pacific and independent expert Ketan Joshi reveals how the frenzied rollout of AI data centres in Australia is set to derail the renewable energy transition, entrench gas and turbocharge climate pollution, prompting calls for an urgent moratorium on data centre approvals until appropriate guardrails are in place.

The frenzied rollout of AI data centres in Australia is rushing through massive new projects, which will derail Australia’s energy transition unless the government urgently intervenes.

Key findings

• The frenzied rollout of AI data centres in Australia is rushing through massive new projects, which will derail Australia’s energy transition unless the government urgently intervenes. Our conservative assumptions mean this impact is understated, in this analysis.
• Australia’s biggest proposed data centre, the 1GW Mamre Road Data Centre Campus in Western Sydney, will generate peak annual grid emissions equivalent to that produced by 560,000 petrol cars for a year or all domestic flights within NSW in 2023.

• Data centres already fail to cover their own emissions with new renewables and their rollout will dramatically hold back Australia’s energy transition.
• No data centre operator analysed in this report adequately proves their claim of driving Australia’s renewable energy growth. Claims they are doing this through truly “additional” new power purchasing agreements for renewable energy are unsubstantiated.
• There are early signs of a data centre-fuelled gas boom in Australia, which will come with massive, nationally significant climate costs. For example, the Tamboran proposal for the Northern Territory would effectively double the state’s emissions. In NSW, Cloud Carrier’s proposed gas-fired project would wipe out NSW’s entire projected 2028 emissions cuts.

• Even if only 1 in 4 new Australian data centres were powered by new on-site gas, it would result in 2.8x higher total emissions compared to using grid power.
• New analysis shows that on-site gas for data centres globally could fuel emissions that exceed Brazil’s total power grid emissions by 2030.
• Fossil fuel corporations are quietly joining the data centre lobby group as members, and sponsoring and attending technology industry conferences. The two industries are reinforcing each other’s talking points and PR spin.

• Data centre operators do not disclose the customers of an individual facility, the purpose of the computations performed there, or site-specific energy consumption, despite the industry’s defense of its ‘critical infrastructure’ status or claims of transparency. It is a matter of public record that AI is being used for abuse, war and other human rights violations.
• Data centres can be ‘right sized’ through community ownership schemes, well-deployed AI software and strict moratoria to allow for democratic governance of this industry.

This report recommends:

• An urgent moratorium on data centre development until safeguards are legislated
• Binding, legislated standards for AI development, including substantiated claims of additional renewable energy
• Full disclosure of services delivered, emissions, finances and energy use, per project
• Full assessment of compliance with human rights frameworks

Lead author: Ketan Joshi is an independent climate, environment and sustainability expert. He was the lead author on “The AI Climate Hoax”, published with several corporate accountability and environmental groups in 2026, and previously wrote “Windfall: Unlocking a Fossil Free Future” with the University of New South Wales Press. He worked for eight years in Australia’s renewable energy sector (corporate and government), and has worked with European NGOs working on climate communications and corporate accountability.

Energy Vampires: the AI data centres draining Australia

Valley Link would connect a potential nuclear reactor and fossil fueled-powered plants to serve suburban data centers.

By Charles Paullin

GOOCHLAND, Va.—Deborah Blackburn leaned on her cane in a line to enter the Central High Cultural and Educational Complex, angst-ridden over a giant transmission line proposal for reasons that are common refrains here: It’s all to benefit data centers in Northern Virginia, and it will disrupt the rural character here outside Richmond.

Residents Wrangle Over Transmission Line Proposal for Rural Virginia

A major change in the way that China measures its core climate goal has effectively halved the growth in the country’s carbon dioxide (CO2) emissions over the past five years.

The revised measure of “carbon intensity”, the amount of CO2 per unit of economic output, implies that China’s emissions have only gone up by 7% from 2020-2025.

This is just half of the 14% rise indicated by previous official statistics.

On paper, the revision creates a gap of 700m tonnes of CO2 (MtCO2) per year, equivalent to the total emissions of Germany or South Korea.

While China has never officially defined how it measures carbon intensity, it has now made what appears to be a retrospective change, with the effect of making targets easier to meet.

The shift means that China officially came close to meeting its carbon-intensity target for 2020-2025, whereas official statistics had previously pointed towards falling well short.

The new definition of carbon intensity has not been made public, but plausible approaches to calculating the metric do not seem to be sufficient to explain the Germany-sized gap.

The apparent gaps or inconsistencies in China’s new carbon accounting also mean that China could meet its international climate pledges for 2030, even if its emissions go up, whereas the previous measure would have required them to fall.

This article explains how the metric appears to have shifted, what changes might potentially explain the revision and what the revised measure implies for China’s climate goals.

Measuring carbon intensity

Reducing carbon intensity – CO2 emissions per unit of GDP – has been China’s key climate commitment since the Copenhagen climate conference in 2009.

At that time, the country pledged to cut its carbon intensity to 48% below 2005 levels by 2020. This was followed up by a 2030 target of a 60-65% reduction, announced in 2014, which was then upgraded to more than 65% in 2021.

Since carbon intensity was made a key progress indicator in China’s 14th five-year plan for 2021-25, the country has reported reductions in carbon intensity every year in its statistical communique, issued at the end of February.

Neither China’s international climate pledges (its nationally determined contributions, NDCs) nor other official documents have ever set out a definition of carbon intensity, despite it being a cornerstone of the country’s climate commitments.

However, until this year, it was possible to closely reproduce the reported numbers, based on a straightforward interpretation of what carbon intensity means.

But the types of emissions that are included in the carbon-intensity metric have now changed.

Previously, it was possible to reproduce the reported carbon-intensity data by combining official GDP data with estimates of emissions from the use of fossil fuels. The latter could be estimated based on the officially reported consumption of coal, oil and gas, multiplied by China’s official emissions factors for the CO2 per unit of energy from each fuel.

The previous carbon-intensity measure apparently included emissions from the use of fossil fuels to generate energy, as well as their use as chemical feedstocks, so-called “non-energy uses”. However, it did not include non-fossil fuel CO2 emissions from industrial processes, such as the production of cement, as shown by the “old scope” in the figure below left.

Based on the annually reported progress against this old scope, China’s carbon intensity had fallen by a total of 12.4% from 2020-2025.

This was well short of the 18% target set for these years under the 14th five-year plan.

In September 2025, Huang Runqiu, head of the Ministry of Ecology and Environment, acknowledged this gap, saying that meeting China’s carbon-intensity targets had become “more challenging” due to the effects of the Covid-19 pandemic and trade tensions.

Yet the 15th five-year plan, published in March 2026, reported that China had cut its carbon intensity by 17.7% over the same period – just shy of the 18% target.

As such, it is clear that there has been a major shift in the way that China measures its carbon intensity, specifically in terms of which types of emissions are included.

Moreover, the revised numbers imply that – rather than missing it by a large margin – China officially came close to meeting its carbon-intensity target for the 14th five-year plan.

A footnote in China’s latest statistical communique offers a brief description of carbon intensity as relating to the CO2 emissions from “energy activities and industrial production”.

This indicates that the carbon-intensity calculation now includes industrial process emissions and excludes non-energy uses of fossil fuels, shown by the “new scope” in the figure above.

In comments sought by Carbon Brief, Ryna Cui, associate research professor at the University of Maryland School of Public Policy, who was not involved in the analysis, agrees that the changes to the carbon-intensity methodology are “unclear”. However, she notes that “limited data” makes it challenging to fully verify the nature and impact of the changes.

The revision mirrors a recent change made to the way that China measures its “energy intensity”, the energy use per unit of economic output. In 2024, energy intensity was changed to exclude non-energy use of fossil fuels and energy use from non-fossil fuels.

This exclusion also created a major incentive for expanding the chemical industry and the non-energy use of fossil fuels.

As for the change in carbon-intensity metric, this follows the highly energy-intensive pattern of economic growth during and after the Covid-19 pandemic and China’s “zero-Covid” policy.

Germany-sized gap

The shift in the way that China is measuring its carbon intensity has implications for estimates of the country’s emissions, which are only reported officially some years later.

Changes in carbon intensity and GDP are reported far more quickly – and can be used to estimate changes in China’s CO2 emissions.

China’s total emissions from energy and industrial processes were 11.2bn tonnes of CO2 (GtCO2) in 2020. Based on the originally reported changes in carbon intensity and GDP, its fossil-fuel CO2 emissions had grown 14% by 2024, an increase of 1,430m tonnes (MtCO2).

In contrast, the newly reported carbon-intensity figures imply that China’s CO2 emissions only grew by 7% between 2020 and 2025, up just 690MtCO2, as shown by the figure below.

The gap between these figures amounts to 730m tonnes of CO2 (MtCO2), equivalent to the annual emissions of Germany or South Korea.

On paper, therefore, the change in the carbon-intensity metric effectively halves the rate of growth in China’s CO2 emissions over the past five years.

Decoding the new carbon-intensity methodology

The change in the carbon-intensity metric could have other significant implications, explored below, making it important to understand how it is being calculated.

Yet, while there are some indications of what the new approach entails, these changes do not seem to account for the magnitude of the revision.

The new scope includes industrial-process emissions. One of the largest sources of these emissions, the cement industry, has been contracting due to a slowdown in real estate and infrastructure construction.

This reduction in emissions is one reason why China’s carbon intensity has improved more quickly under the new scope than under the old one.

In addition, the new scope excludes non-energy use of fossil fuels – largely relating to the chemicals industry – where there has been rapid growth over the past five years.

This is another factor in carbon intensity improving faster under the new scope.

Indeed, China’s chemicals industry drove more than half of the growth in its total fossil-fuel use in the past five years, including 40% of coal use and all of oil use. As a result, non-energy use reached 13% of the total consumption of fossil fuels in 2025, up from 7% in 2020, after growing at an average annual rate of 13%.

The figure below illustrates the impact of these changes in scope. It shows the change in China’s emissions from 2020-2025 due to the use of fossil fuels for energy, its industrial-process emissions and non-energy use of fossil fuels.

The first few rows show changes based on the consumption of fossil fuels overall, amounting to a combined 1,430MtCO2 rise in emissions.

This compares with the 690MtCO2 rise implied by the new carbon-intensity metric, leaving that Germany-sized 730MtcO2 gap in emissions. The new scope explains some of this gap.

In terms of industrial processes, the 30% fall in cement production could account for a 300MtCO2 fall in China’s CO2 emissions. In addition, the amount of carbon stored in products, such as plastics, asphalt and rubber, could account for an estimated 100MtCO2 fall in emissions.

On the other hand, emissions from the incineration of plastics increased by an estimated 40% and from metals industry processes by 10%, with aluminium production having expanded by 21%. Together, these would have increased emissions by an estimated 60MtCO2.

In total, the changes in emissions from fossil-fuel use, industrial processes, carbon retained in products and waste incineration add up to a combined 1,070MtCO2 rise from 2020-2025, shown in the penultimate row of the figure below.

Again, this revised total – based on the change in scope of the carbon-intensity metric – goes some way to explaining the Germany-sized gap in China’s CO2 emissions.

However, the new carbon-intensity figures imply that China’s CO2 emissions only increased by 690MtCO2, as shown in the final row of the figure below. This leaves a residual gap of around 380MtCO2, which does not appear to be accounted for by the data available.

One way to make the numbers add up would be to assume that the amount of carbon embedded in chemical-industry products has increased by the equivalent of 500MtCO2.

However, the reported output of major chemical-industry products cannot account for this level of embedded carbon. The figure below shows that the increase in output of major chemical products only explains around a 110MtCO2 increase in retained carbon.

Much of the increase in the production of plastics was cancelled out by a contraction in the use of bitumen for asphalt, due to lower road-building activity.

Furthermore, the 14th five-year plan for 2021-25 had a target of raising the share of waste incineration to 65% of urban residential waste treatment capacity, up from 45% in 2020.

So, while plastics production did go up, resulting in increased amounts of retained carbon, a larger share of this retained carbon was being incinerated, meaning its carbon would quickly be released back into the atmosphere.

One reason why carbon retained in products has grown more slowly than the amount of fossil fuels used in chemicals production is that the fastest growth has been in the coal-based chemicals industry.

Coal-based processes have a much lower conversion efficiency than oil- and gas-based production, with process emissions that are typically multiple times as high.

For example, these emissions are 10 times as high for the production of olefins – a key plastics feedstock – from coal as compared with oil or gas. The process is reported to require 3.75 tonnes of standard coal per tonne of product. This implies that only 30% of the carbon in the coal is retained in the product, with the other 70% being emitted in the process.

There are also chemical processes that use fossil fuels as a feedstock, but where the end product does not contain carbon. One example is ammonia, a key building block for fertiliser, where production grew by 52% from 2020 to 2025.

Neither the change in scope of the carbon-intensity calculation, nor the change in the amount of carbon retained in products, is sufficient to explain the size of the revision in the newly reported numbers. There must be another explanation.

There are two options. Either the new scope broadly aligns with what is outlined above, but also excludes a subset of the CO2 emissions. Or the scope does not exclude any of the CO2, but there are gaps in the monitoring of some energy or industrial-process emissions.

Either explanation would mean that China is not accounting for some of its CO2 emissions. It would also mean that the improvement in carbon intensity for 2020-2025 is over-reported.

China’s latest officially reported emissions inventories reinforce the second of the two options above, namely, that there are gaps in emissions reporting from the chemical industry.

From 2018 to 2021, the latest year for which China has reported on its emissions, the CO2 output of chemical-industry processes only increased by 13%. Over the same period, non-energy use of fossil fuels increased by 29%, according to data reported to the International Energy Agency by the Chinese government.

One factor in these apparent gaps could be that China’s National Bureau of Statistics (NBS) is required to publish data on carbon intensity very quickly, since it is a key indicator in the country’s five-year plans.

On the other hand, detailed greenhouse gas emissions inventories and energy statistics are only published years later, by the environment ministry and NBS, respectively.

What the change means for China’s targets

The change in the definition of carbon intensity has the effect of weakening China’s climate targets and introducing more uncertainty into tracking progress.

On the basis of China’s new numbers, it will require less effort to hit the 2030 target for a 65% reduction in carbon intensity on 2005 levels, as per China’s Paris pledge.

This target can now be met even if CO2 emissions go up between 2025 and 2030, whereas the previous metric would have required a reduction.

It will also require less effort to hit the 17% target in the 15th five-year plan.

The apparent gaps in the CO2 emissions numbers for 2025 could affect the delivery of China’s other key climate pledges, such as the commitment to peak CO2 emissions before 2030. They could also allow the chemical industry’s CO2 emissions to continue climbing rapidly, while still officially meeting the 2030 goals for CO2 intensity.

Moreover, the apparent gaps or inconsistencies in China’s new carbon accounting also mean that China would be able to officially meet its target to peak its CO2 emissions by 2030, even if its overall CO2 emissions do not actually reach a peak.

The apparent gaps could also affect the delivery of China’s newer target to cut its greenhouse gas emissions to 7-10% below peak levels by 2035 and beyond.

Nevertheless, researchers and analysts can still monitor progress by calculating China’s CO2 emissions independently.

China’s reporting on fossil-fuel consumption, the output of plastics and other carbon-containing products, as well as manufacturing of commodities with substantial process emissions, provides a basis for tracking emissions under the new scope.

While under the UN’s climate framework China is free to use any definition it wants to meet its own nationally determined climate pledges, retrospective changes to methodology or inconsistent accounting could erode the value of the country’s commitments.

Moreover, it will, ultimately, have to close any gaps in its emissions data and reporting, under the transparency rules of the Paris Agreement.

China’s next transparency report to the UN, due by the end of this year, should also provide more clarity on the methodology and data underlying the revised numbers.

This underscores the importance of monitoring, reporting and verification for industrial process emissions. “Mass balances” based on fossil-fuel consumption and product output could be used as a check on CO2 emissions reporting. Finally, China’s emissions data could also be made more granular and clearly defined.

Carbon Brief has approached the National Bureau of Statistics and Ministry of Ecology and Environment for comment.

The University of Maryland’s Cui tells Carbon Brief that in general, China’s climate goals are “improv[ing]” in terms of their coverage and scope. However, she adds:

“The issue is…the ambiguity and inconsistency in the coverage, definition and method between target setting and progress tracking, which can lead to large uncertainties and room for manipulation. It highlights the importance of transparency in national climate targets, following the UNFCCC’s international transparency framework, which should also be applied as best practices for domestic targets.”

About the data

The calculations in this analysis are based on China’s total coal, oil and gas consumption from energy statistical yearbooks covering the years until 2023, with data for 2024 and 2025 taken from the latest statistical communiques.

“Originally reported” CO2 emissions were back-calculated from carbon-intensity reductions and GDP growth given in annual statistical communiques. The revised emissions for 2020, 2024 and 2025 are similarly back-calculated from the reductions in carbon intensity from 2020 to 2025 and from 2024 to 2025, as reported in the 15th five-year plan outline and the 2025 statistical communique, respectively, combined with annually reported GDP growth.

Cement process emissions up to 2024 are from Robbie Andrews’ estimates, scaled to 2025 based on year-on-year change in total cement output.

Process emissions from the metals industry are based on calculating emissions for aluminium, silicon, lead, zinc and crude steel from the bottom-up, using industrial output data and IPCC default emission factors scaled to the reported total in 2021. For steel, the calculations are based on typical quicklime use in basic-oxygen and electric-arc furnaces.

Emissions from the incineration of plastics are based on a peer-reviewed estimate of plastics incineration in 2022, combined with growth rates in the overall power generation from waste-to-energy plants. The analysis assumes that the share of plastics in the energy content of the incinerated waste stayed constant over this period, which is a conservative assumption given the rapid rise in plastics production.

Total non-energy use of fossil fuels in 2020, 2024 and 2025 is available from an NEA data release, with data for 2021-2023 found in the China energy statistical yearbook 2025.

The mix of coal, oil and gas within non-energy use is based on the energy statistical yearbook data up to 2023, with the increase in coal in 2024 and 2025 based on Wind Financial Terminal data on coal consumption in the chemical industry. Gas use, which is relatively minor, is assumed to have grown on trend and oil is calculated as the residual.

Primary plastics, rubber, and urea output data are from NBS industrial statistics. The production of solvents, lubricants and waxes, as well as the use of bitumen in construction, is from energy statistical yearbooks. The analysis assumes no change in output from 2023 to 2025, given the lack of clear trends.

Guest post: How changes to coal mining have affected China’s methane emissions

China energy

|

19.03.26

Q&A: What does China’s 15th ‘five-year plan’ mean for climate change?

China energy

|

06.03.26

Ma Jun: ‘No business interest’ in Chinese coal power due to cheaper renewables 

China energy

|

19.02.26

Analysis: China’s CO2 emissions have now been ‘flat or falling’ for 21 months

China energy

|

12.02.26

The post Analysis: China’s new carbon metric leaves Germany-sized gap in its emissions appeared first on Carbon Brief.

Analysis: China’s new carbon metric leaves Germany-sized gap in its emissions