The Gas Consumption Pipeline Isn’t Easily Broken
Illustration by Paul Ward
Natural gas production and consumption is enmeshed in the global economy at every level, from the macroeconomic to the hyperlocal. The fundamental dynamics of the industry are shaped by the wide dispersion of gas production, by the physical constraints on its transport, and by many countries’ heavy reliance on gas for essential winter heating or powering industrial production. With the world’s economic and political outlook more uncertain than ever, how should we assess the near-term prospects of the gas market?
Forecasters from the Swift Centre for Applied Forecasting were asked to predict global annual natural gas consumption in 2025 and 2026. They were further asked to make two conditional forecasts about major exogenous events that could affect the natural gas market – 1) the probability of an economic recession in the US or EU by December 31, 2026, and 2) the probability of a substantial military conflict in the Middle East, involving at least one of Qatar or Iran, by August 31, 2026 – and then to assess how either of these events coming to pass could affect gas consumption in 2026.
Key takeaways
The Swift Centre forecasters thought that gas consumption growth is very likely to remain at historical average rates (around 2%) in both 2025 and 2026, with limited tail risk on either side.
The Swift Centre forecasters generally expect modest economic growth in the US and EU over the next two years; while US growth is likely to be resilient, some saw a non-trivial risk of a mild recession in the EU.
A garden-variety recession, however, was not seen as a significant threat to the trend of consistent gas consumption growth in the forecast period. The Swift Centre forecasters thought that it would probably take a major economic downturn – which they saw as unlikely – for gas consumption to decline outright.
The Swift Centre forecasters saw a roughly one-in-four chance of a “substantial” military conflict in the Middle East involving Qatar or Iran by mid-2026, but such a conflict would be unlikely to seriously affect global gas consumption unless it was substantial enough to involve the US or other major powers – and maybe not even then.
When considering these questions, most of the forecasters grounded their reasoning on a few basic observations. The first and foremost is that gas consumption trends have not been particularly volatile in recent years. Annual consumption growth has hovered around 2% for much of the past decade. Recessions have dampened growth, but only in major economic downturns – such as those caused by the Global Financial Crisis or the COVID-19 pandemic – has consumption declined outright. Underlying this resilience is the fact that gas consumption tends to be inelastic: it is relatively difficult for consumers to switch away from gas because its primary use is for the heating of homes and businesses. At the same time, the production of natural gas has become widely distributed geographically, so when supply is interrupted in one region, other regional centres are often able to take up the slack. Localised supply interruptions might cause swings in gas prices, but consumption trends are less liable to change.
The forecasters noted consistent patterns in the numbers released by two of the leading sources of data on the gas market, the IEA and the Energy Institute. The Energy Institute’s numbers for 2023 and 2024 “are consistently lower than IEA’s by 2%,” one forecaster calculated, and several others made the point that the “EIA tends to undershoot its forecasts more frequently than it overshoots them.” Overall, the forecasters saw these divergences in data from different sources as consistent and predictable, and thus easily corrected for. The IEA’s forecast error, for example, “generally does not exceed 1.5% in either direction.” And in the forecasters’ eyes, both major data sources pointed to the same overall conclusion: stable growth is by far the most likely outcome, especially in the near term.
Resilient consumption, resilient economies
What will global natural gas consumption be in 2025?
Median estimate: 4,280 billion cubic metres (80% confidence window 4,181-4,355 bcm)
For 2025, the forecasters converged tightly around a median forecast of global growth of 4,280 billion cubic metres (bcm), which would mark a 2% increase over the IEA’s 2024 figure of 4,210 bcm. The very narrow prediction range (no forecaster had their median growth below 4,226 bcm or above 4,328 bcm) was indicative of a strong consensus around a prediction of historically stable growth in 2025 – and not only because the data for the first two quarters of this year has already been released. As one forecaster put it, 2025 just “feels like a tight band year – stable fundamentals, not a lot of drama.”
Last year, according to the Energy Institute, gas consumption grew by 2.8%. One forecaster observed that “this seems to have been a return to approximately the rate of growth you’d expect in ‘normal’ times.” Another said that they “see little reason to depart from the longer-term average calculation [of] 2.0%. I consider this growth rate reasonable because a major economic interruption (like a pandemic or severe economic recession) is highly improbable for 2025.”
The forecasters were virtually unanimous in deeming a severe economic downturn – or, conversely, rapid growth – unlikely for the remainder of this year, all but ruling out one of the main historical causes of deviations from the gas consumption trend. “The first half of 2025 saw enormous economic uncertainty because of the Trump administration’s tariffs,” one observed, but “while some of that uncertainty remains, there is increasing clarity, and the world is adjusting to probable new realities.” This forecaster expects “a bit of a rebound” in the second half of the year, but not enough to “fully make up for the slower growth seen in H1 2025.” All in all, modest economic growth was expected to translate to gas consumption staying on-trend.
Among the potential drivers of faster gas consumption growth, several of the forecasters highlighted rising energy demand from AI data centres and new demand created by industrial and infrastructure development in Asia as particularly important – but at the same time, they generally did not anticipate either factor pushing consumption higher than the current trend until 2026 at the very earliest. Data centres today “consume large amounts of energy, but not on a global scale,” one wrote. “The construction of the Medog dam by China in Tibet is starting and will use vast quantities of energy-intensive cement, but not in 2025.” Some reported that gas storage levels in Europe are somewhat lower than usual for this time of year, a situation which could spur somewhat higher consumption as stocks are replenished before next winter – currently projected to be colder than normal due to a weakening polar vortex.
More of the same (gas) in 2026 – with the usual tail risks
What will global natural gas consumption be in 2026?
Median estimate: 4,378 bcm (80% confidence window 4,212-4,511 bcm)
The forecasters saw the 2026 gas consumption outlook as essentially the as that of 2025: the median (4,378 bcm) and 80% confidence range (4,212-4,511 bcm) were almost identical to those of the first forecast, with the minordivergence simply reflecting the fact that tail risks had to be slightly higher on both sides because “an additional year increases uncertainty regarding the global economy.”
Accordingly, forecasters’ assessments of the balance of risks for the global economy in 2026 diverged more as the forecasted period stretched further into the future. Six forecasters put the risk that the EU or US would undergo an economic recession in 2026 at between 8% and 15%, but two others pegged it at a much higher 45%.
Will the US or EU experience an economic recession by 31st December 2026?
Probability: 19% (range 8%-45%)
2026 gas consumption in the event of a recession
Median estimate: 4,322 bcm (80% confidence window 4,164-4,460 bcm)
On the more optimistic side, one forecaster argued that “the recent tariff deal between the US and the EU to tax most incoming goods at 15% is an indicator that the policies of Trump are … not disastrous”; they also saw “the rapid return to peace in both the Pakistan-India conflict and the Cambodia-Thailand conflict” as indicative of Trump’s and Xi’s mutual inclination to help “quickly put an end to potentially devastating wars.” Several predicted US interest rates being left unchanged until the end of the current Fed Chair’s term in May 2026, helping to contain tariff-related inflation, with the successor Chair being likely to cut rates – potentially sparking inflation but also fueling above-trend growth for most of 2026. “I don’t think the US is going to experience a recession before 2027,” one wrote. More pessimistic was the forecaster who saw “a reasonably high chance of a recession in either the US or the EU by the end of 2026, for several reasons”; in the US, these potential risk factors included the economic toll of Trump’s tariffs, the overvaluation of both the housing and stock markets, and AI-related disruptions to employment.
Most of the forecasters assessed the EU’s economic prospects in 2026 as “shakier” than those of the US, with several singling out Germany as a source of concern, but most judged the combination of monetary stimulus (recent rate cuts) and continued military spending as enough to stave off recession in Europe. “I think they probably muddle through,” one wrote of the EU. Another was more pointed in coming to the same conclusion: “The EU is, I think, unlikely to experience a strict recession, although it will have a per-capita one.”
Despite any differences on the economic outlook, the forecasters were once again strongly aligned on the likely stability of gas consumption growth in the absence of major, black/grey-swan external factors. Even “if there is a recession in the US or the EU before the end of 2026,” one forecaster wrote, “I would expect it to be mild … and [to] at most do little more than flatten gas consumption.” The risk would be higher in the event of correlated recessions in both the EU and US, which one forecaster did call a “non-trivial tail risk…. If both regions stall out or the downturn coincides with energy price volatility or another shock, we could see demand fall off more sharply.” In any significant recession, declines in global gas consumption would probably be more directly linked to lower Asian industrial production than to lower economic activity in the US or EU per se. Germany’s total gas consumption is less than 2% of the global total, and the UK’s share is lower still.
More-recession-resistant drivers of gas demand include the continued build-out of AI data centres and the above mentioned Medog dam project in China. The latter, in particular, will engender massive energy demand via the production of cement and steel – but even in 2026, this project will still be in its early stages. This is another illustration of a general theme: almost everywhere they looked, the forecasters tended to see offsetting factors that pushed their predictions back toward the median.
Will there be a substantial military conflict in the Middle East, involving at least Qatar or Iran, by 31st August 2026?
Probability: 24% (range 14%-35%)
2026 gas consumption in the event of such a war
Median estimate: 4,338 bcm (80% confidence window 4,123-4,490 bcm)
The second conditional question yielded comparable probabilities and a similar takeaway: the odds of a substantial war in the Middle East greatly affecting global gas consumption were seen as quite low. On the probability of such a war, the forecasted odds ranged from 14% to 35%. Several forecasters noted the ambiguity of the “substantial” criterion in this question; at least one adopted a working definition of “substantial” as “enough to affect gas exports from the region”. This definition matters because the probability of military conflict is endemically high for the region but arguably relatively low for the two countries in question: “Conflict in the Middle East is chronic,” as one forecaster wrote, “and the probability of that is well above 50%. But not so much when we are talking about Qatar and Iran.”
The forecasters were generally dubious about the odds of a war involving Qatar itself within the forecast period. “Qatar has never been to war,” one pointed out. Another argued that “if Qatar gets involved in a substantial military conflict, it would most likely be because the US is in a major conflict” and is compelled to guard its Al Udeid airbase there.
By contrast, the likelihood of Iran being involved in hostilities was seen as fairly high. “War with Israel could break out again quickly,” said one forecaster (estimate 35%). Another forecaster (estimate 14%) believes that “Israel and/or the US is likely to attack Iran again if they have evidence that Iran is making progress towards a nuclear weapon or vastly improving their long range missile capabilities” – but added the caveat that “these are more likely to be more limited strikes or covert operations.” There was considerable disagreement on the likelihood of civil war within Iran. One forecaster (estimate 33%) stated that “I think there is a substantial chance of a military conflict involving Iran, including the possibility of a civil war. The Iranian regime is spent.” Another (estimate 14%) read the situation very differently: “The regime of the Islamic Republic of Iran is currently very weakened, but they still control a very tough security services. Dissent is not permitted and mass purges will make it very difficult for an organised internal opposition to form.”
A key point on which the forecasters agreed was that a war involving Iran (or Qatar) would probably not have a major impact on the global gas market. “Most war scenarios would only have a very limited impact on overall consumption,” one forecaster wrote – though he thought that a “much larger war” that could be more disruptive of gas supply does qualify as “a real tail risk”. Several saw a closure of the Strait of Hormuz as the only scenario in which global gas consumption would be seriously affected by a Middle East war, but they generally saw this prospect as highly unlikely: “I don’t think Iran or any other country would be likely to close the Strait of Hormuz,” said one, “because it would also damage that country’s own bottom line.”
Conclusions
In a global environment of nominally high uncertainty, both economically and politically, the Swift Centre forecasters’ convergence around historical average growth rates for gas consumption is quite striking. With the exception of some of the subsidiary forecasts, the probability distributions were neither fat-tailed nor asymmetrical. Tail risks were noted, particularly later in the forecast period, but these risks were hardly obscure: the greatest threats to gas consumption growth were judged to be major recessions and major wars, and the most notable new drivers of consumption were the AI buildout and infrastructure development in China. The forward path for natural gas consumption appears to be one of the few straight-and-level roads in very bumpy terrain.
