A 0.4% Dip, a Near-Empty Budget: Reading the Global Carbon Budget 2026
Every November, just before the world's climate negotiators gather, a large international team of scientists publishes the accounts of the planet's carbon ledger. The Global Carbon Budget is not a forecast model or a corporate pledge tracker. It is a bookkeeping exercise, coordinated by the Global Carbon Project and led for this edition by Pierre Friedlingstein of the University of Exeter, that adds up where carbon dioxide comes from and where it goes. The 2026 edition, circulated as a preprint under review for the journal Earth System Science Data with public discussion opened on 10 November 2026, carries a headline that will be quoted widely: global fossil carbon dioxide emissions are expected to decrease in 2026. This Research Desk analysis looks at that single finding in depth, at what lies underneath a very small number, and at why the authors themselves describe current trends as far from what the Paris Agreement requires.
The headline number: a 0.4% dip
Start with the number. According to the study, global fossil CO2 emissions, counting the small carbon uptake from cement carbonation, are expected to decrease in 2026 by 0.4% relative to 2025, within a range of minus 1.5% to plus 0.6%. That brings fossil emissions to an expected 10.4 billion tonnes of carbon, or 38.2 billion tonnes of carbon dioxide. The comparison point is 2025, when the consolidated data show fossil emissions grew by 1.2% to 10.5 billion tonnes of carbon, plus or minus 0.5, which the study expresses as 38.4 plus or minus 1.8 billion tonnes of carbon dioxide. A footnote explains that all growth rates are adjusted for the extra day in the leap year 2024, so that the comparisons are like for like. The first thing a careful reader should notice is the width of the uncertainty band. A range that runs from a decline of 1.5% to an increase of 0.6% means the authors cannot rule out that 2026 will in fact show growth. The study is explicit that the preliminary estimate is based on the data available for 2026 so far, and that the uncertainty quoted relates only to the estimated growth rate, not to the uncertainty of the 2025 baseline it is applied to.
How the 2026 projection is built
It helps to understand how the projection is built, because the method explains both its strengths and its limits. The authors do not estimate global emissions from one source. They assemble separate projections for China, the United States, the European Union, India, Japan and for all other countries combined, using a different method for each. For China they use the growth rates in the National Bureau of Statistics' 2026 statistical communique, together with cement clinker trade data and estimated changes in lime production. For the United States they take the Energy Information Administration's Short-Term Energy Outlook and combine it with year-to-date cement clinker production. For the European Union they use monthly energy data from Eurostat through July and August, extended with daily gas demand estimates and daily coal-power generation data. India and Japan are projected from monthly estimates through the late summer. Aviation is derived from monthly IATA data, shipping from bunker fuel sales and OECD estimates, and the rest of the world from International Monetary Fund growth projections combined with extrapolated changes in emissions intensity, constrained by global oil and gas totals. The study also cross-checks the result against Carbon Monitor, a separate dataset built from daily proxies such as electricity consumption and travel patterns. Carbon Monitor lands in the same place, with a global decrease of 0.5%, although with a much wider range of minus 3.5% to plus 2.6%. The agreement between two independent approaches is reassuring on direction. It is less informative on size.
The fuel split: oil down, coal up
The more revealing story is in the fuel split. The study projects that oil emissions in 2026 will fall by 1.9%, with a range of minus 2.6% to minus 1.2%, and natural gas emissions will fall by 0.3%, with a wide range of minus 1.6% to plus 0.9%. Cement is expected to fall by 0.6%. Against those declines, coal emissions are expected to grow by 0.6%, with a range of minus 0.6% to plus 1.7%. In the authors' words, the decreases in oil and gas are partly compensated by the growth in coal. Notice that oil is the only fuel whose projected decline sits entirely below zero across its whole range. That makes oil the most robust part of the 2026 story, and the global dip is in large part an oil story. Coal, the most carbon-intensive of the three fossil fuels, is moving the wrong way, even if only slightly, and its range straddles zero. Gas is essentially flat.
Regions and fuels: who is moving which way
The regional picture explains where those fuel movements come from, and it is not a story of one direction. The study projects that fossil emissions will decrease in 2026 in China by 0.9%, in the United States by 1.3%, in the European Union by 1.5% and in Japan by 1.0%, while India is projected to rise by 4.4%, with a range of 1.7% to 7.1%. International aviation and international shipping are each projected to be 1% lower, and the rest of the world in aggregate is expected to fall by 0.7%. Every one of the regional declines comes with an uncertainty range that touches or crosses zero. India's range, by contrast, is entirely positive. Put differently, the evidence for a rise in India is statistically firmer than the evidence for a fall in any of the other major regions.
The size of each emitter gives these percentages their weight. China is projected to emit around 3.4 billion tonnes of carbon, or 12.3 billion tonnes of carbon dioxide. The United States is around 1.4 billion tonnes of carbon, or 5.0 billion tonnes of carbon dioxide. India is around 0.9 billion tonnes of carbon, or 3.3 billion tonnes of carbon dioxide. The European Union is around 0.6 billion tonnes of carbon, or 2.4 billion tonnes of carbon dioxide, Japan around 0.3 billion tonnes of carbon, or 1.0 billion tonnes of carbon dioxide, and international aviation and shipping together around 0.3 billion tonnes of carbon, or 1.2 billion tonnes of carbon dioxide. The rest of the world, a very large group of countries, accounts for around 3.7 billion tonnes of carbon, or 13.6 billion tonnes of carbon dioxide. A modest percentage change in the largest emitters therefore moves more carbon than a dramatic change in a small one, and India's growth, though large in percentage terms, comes from a base well below that of China or the rest of the world combined.
Looking inside each country by fuel shows how differently these economies are moving. In China the study projects coal emissions to rise slightly by 0.3% while oil emissions fall by 4.6%, natural gas by 0.9% and cement by 6.2%. China's decline is therefore driven by oil and cement rather than by a retreat from coal. The United States is close to the opposite pattern. Coal emissions are projected to fall by 6.7% and oil by 1.1%, while natural gas edges up by 0.5% and cement rises by 6.2%. The European Union shows declines across the board: coal down 2.9%, oil down 1.8%, gas down 0.1% and cement down 1.8%. Japan has coal up 1.7%, oil down 2.2%, gas down 4.0% and cement down 3.1%. India is the outlier, with coal emissions up 5.9%, gas up 2.2%, oil down 0.9% and cement up 9.0%. The global coal increase is, in large part, an Indian coal story, combined with a flat Chinese coal line. The global oil decline is heavily a Chinese story, since a 4.6% fall in the oil emissions of the world's largest emitter carries a great deal of weight. The study does not attribute causes to each of these fuel movements in the projection section, and this article does not either, but the pattern is clear enough that readers tracking transport electrification, coal power and construction activity will recognise where to look.

Decarbonising economies, rising demand
To understand why this year's small dip does not amount to a turning point, it helps to step back from 2026 to the decade. The study reports that fossil CO2 emissions decreased, at the statistical threshold the authors apply, in 40 countries that also had growing economies during 2016 to 2025. In the previous decade, 2006 to 2015, the equivalent number was 22 countries. These 40 decarbonising economies contributed 2.8 billion tonnes of carbon, or 10.1 billion tonnes of carbon dioxide, of annual fossil emissions over the last decade in aggregate, which is 26% of the world's fossil CO2 emissions. That is a real and encouraging development, and it speaks directly to the question of whether growth and emissions can be separated. It also means that three-quarters of global emissions come from places that have not shown that pattern.
The carbon intensity of energy, meaning the emissions released per unit of energy used, has consistently declined over the past decade, by 0.4% a year globally. The summary reports declines in China, the European Union and the United States, and no decline in India, while the rest of the world combined also declined by 0.3% a year. The discussion section is more cautious about why this has not translated into falling emissions. It says the global decarbonisation trends have not been sufficient to offset the growth in global energy demand, and that energy demand has been growing faster than expected because the decline in energy use per unit of GDP has weakened, in China, in the United States and globally. In China, the study notes, emissions growth slowed from 5.7% a year in 2006 to 2015 to 2.7% a year in 2016 to 2025. It attributes that to weaker economic growth, worth about four percentage points, offset by a considerable weakening in the rate of improvement in energy per unit of GDP, worth about two percentage points, alongside sustained improvements in emissions per unit of energy. In India, strong economic growth is not offset by declines in either energy per GDP or emissions per unit of energy, which drives emissions up. In the rest of the world, economic growth has slowed considerably, but carbon intensity has increased. The study adds that, despite high deployment of renewables in some countries, fossil sources continue to grow to meet rising energy demand.
The decade view gives a more honest sense of momentum than a single year. Total anthropogenic emissions, fossil plus land use, have grown by 0.3% a year over 2016 to 2025, against 1.7% a year over 2006 to 2015. In 2025 total anthropogenic emissions were 11.7 plus or minus 0.9 billion tonnes of carbon, or 42.8 plus or minus 3.2 billion tonnes of carbon dioxide, with a preliminary estimate of 11.8 billion tonnes of carbon, or 43.3 billion tonnes of carbon dioxide, for 2026. In other words, even while fossil emissions are projected to dip, the total, which includes net emissions from land-use change, is preliminarily estimated to be slightly higher in 2026 than in 2025. The study attributes the slower growth of fossil emissions over the last decade in part to the emergence of climate policy and technological change, including a shift from coal to gas, growth in renewable energy and reduced expansion and utilisation of coal capacity in some countries. Its overall verdict is blunt: current global fossil CO2 emissions growth is far from the reductions needed to meet the ambitious climate goals of the Paris Agreement.
The 1.5°C budget: under four years
That verdict becomes concrete when the emissions are set against the remaining carbon budget, the amount of carbon dioxide that can still be emitted before a given temperature threshold is likely to be crossed. The study reports that human activities have caused warming of 1.37°C in 2025. From the start of 2027, the remaining budget for a 50% likelihood of limiting warming to 1.5°C is estimated at 40 billion tonnes of carbon, or 150 billion tonnes of carbon dioxide. At the 2026 level of emissions, that is less than four years of emissions according to the executive summary, and about three years according to the discussion section, which describes the budgets as corresponding to about 3, 12 and 24 years for the 1.5°C, 1.7°C and 2°C limits. For 1.7°C the remaining budget is 140 billion tonnes of carbon, or 510 billion tonnes of carbon dioxide, about 12 years. For 2°C it is 280 billion tonnes of carbon, or 1,035 billion tonnes of carbon dioxide, about 24 years.
These figures are the product of a deliberate judgement, and the study is open about how contested the underlying numbers are. The authors update the remaining budget using two sources. One is the IPCC's Sixth Assessment Report estimate, updated for emissions since 2020, which gives 55, 155 and 290 billion tonnes of carbon for the three thresholds. The other is a newer study, Forster and colleagues in 2026, which proposes a significantly lower budget, mainly because of an updated climate emulator for the warming caused by non-CO2 agents and because of the emissions that occurred between 2020 and 2025, and which gives 25, 125 and 275 billion tonnes of carbon. The Global Carbon Budget takes the average of the two, producing the 40, 140 and 280 figures. The authors emphasise the large uncertainties, particularly close to the 1.5°C limit, and note an additional uncertainty of plus or minus 220 billion tonnes of carbon dioxide arising from alternative non-CO2 emission scenarios. The two sets of estimates overlap once all uncertainties are considered. The practical reading is that the 1.5°C budget is somewhere between nearly gone and very small, and that the debate is about how many years remain rather than whether the margin is comfortable.
The study also offers a simple piece of arithmetic that shows the scale of the task. Reaching net-zero CO2 emissions by 2050 would require cutting total anthropogenic emissions by about 0.5 billion tonnes of carbon, or 1.8 billion tonnes of carbon dioxide, every year on average, which is about 4% of 2026 emissions and comparable to the fall in fossil emissions observed in 2020 during the COVID-19 pandemic. Even then, a straight-line cut from now to 2050 would produce cumulative emissions of 140 billion tonnes of carbon, which is well above the 40 billion tonnes of carbon remaining for 1.5°C, though still compatible with the 140 billion tonnes remaining for 1.7°C. Set that beside this year's projected 0.4% reduction, and the gap between what is happening and what a 1.5°C pathway needs is plain. A 4% annual cut, sustained for a quarter of a century, is a different order of change from a dip of 0.4% that the authors say could just as easily be a rise of 0.6%.
What the atmosphere shows
The atmosphere is already reflecting this. Global average atmospheric CO2 concentration reached 425.6 plus or minus 0.1 parts per million in 2025 and is projected to rise by 2.2 parts per million to 427.7 parts per million in 2026, which the study says is 54% above the pre-industrial level of around 278 parts per million in 1750. A small fall in emissions does not reduce concentration. It only slows the rate at which concentration rises, because the carbon already emitted stays in the atmosphere, ocean and land system for a very long time.

Reading the numbers responsibly
Two further points of caution belong in any responsible reading. First, this is a preprint that is still under review, and the public comment period has only just begun. Numbers may be refined before final publication, and the projection for 2026 is by its nature provisional. The study notes that a table in its supplement compares each year's projection since 2015 with the actual emissions assessed in the following year's report, which is the right place to look for readers who want to judge how reliable these early-November estimates have been historically. Second, the headline mixes different quality evidence. The oil decline is statistically firm, India's rise is statistically firm, and nearly everything else has a range that crosses zero. A responsible communicator should say that emissions are projected to be roughly flat, with a slight lean downward, rather than that they are falling decisively.
Finally, the study is a reminder of what kind of evidence this is. The Global Carbon Project is a research consortium of universities and public research institutes from many countries, drawing on official statistics, energy agencies and observation networks, and publishing its data openly. The authors state that the data are available at doi.org/10.18160/GCP-2026, and the preprint is released under an open licence. That openness is part of the value. Anyone, including companies, regulators, journalists and civil society, can check the fuel-by-fuel and country-by-country numbers discussed here against the source. For organisations that report on climate performance, including those preparing ESG disclosures, this is the macro context in which any corporate target sits: a world that is bending the curve of emissions growth but not yet bending the curve of emissions, and a 1.5°C budget measured in single-digit years.
WHY IT MATTERS
A carbon budget is a deadline expressed in tonnes. It does what a bank statement does: it turns a debate about intentions into arithmetic. A 0.4% fall in fossil CO2 in 2026 is welcome news, but set against a 1.5°C budget of about 40 billion tonnes of carbon, which the study puts at under four years of emissions at the 2026 level, it is a small step on a very short runway rather than a change of direction.
For businesses, investors and anyone reading corporate climate targets, the lesson is about pace. The study shows that 40 countries have cut fossil emissions while their economies grew, which proves that growth and emissions can be separated. It also shows that this is happening in places that account for 26% of the world's fossil CO2, and that the gains are being absorbed by rising energy demand elsewhere. A target is only as credible as the speed at which it is delivered, and the budget says that speed has to be far greater than anything seen so far.
For readers following the energy transition, the fuel-by-fuel and country-by-country detail is the useful part. Oil is falling, coal is not, and India's rise is the firmest signal in the data. That is where policy, finance and technology can still move the number. Research Desk will keep reading studies like this one from public-interest bodies, because the people deciding how to spend the remaining budget deserve to see the underlying figures, not just the headline.

Comments