The world economy is being pulled in opposite directions. An investment surge around artificial intelligence is supporting growth, while the war in the Middle East is raising energy costs, inflation and financial risks. The result is not a conventional global downturn, but a more uneven and fragile expansion.
The International Monetary Fund projects global growth of 3.1% in 2026 and 3.4% in 2027, while global headline inflation is expected to reach 4.7% this year before easing to 4.1% next year. The figures suggest resilience, but they also mask a widening gap between economies benefiting from technology investment and those absorbing higher energy and borrowing costs. The IMF’s October outlook describes a recovery that is increasingly uneven.
Why AI is carrying the expansion
AI infrastructure has become a major source of demand. Spending on data centers, advanced chips, electricity networks and related equipment is lifting investment in the United States and in manufacturing economies that supply the technology. The IMF says countries including South Korea, Malaysia, Mexico and Thailand have received upgrades partly because of their role in the AI supply chain. Reuters reported that strong AI investment helped offset the drag from higher oil prices.
There is a credible upside case. If companies adopt AI quickly and productivity improves faster than expected, global output could be 0.6% higher by 2029, with much of the benefit accruing to the United States and China. That would make the current investment cycle more than a speculative boom: it would become a mechanism for raising productivity and expanding potential growth.
Yet investment is not the same as productivity. Companies are committing capital before the long-term commercial returns are fully established. That leaves markets vulnerable to a reassessment if AI systems fail to produce sufficiently large gains in revenue, efficiency or labor productivity.
The risk of an AI correction
The IMF has modeled a downside scenario in which disappointment over AI profitability causes investment and asset prices to fall sharply. Under that scenario, global GDP would be 1.1% lower by 2028, while US output would decline by 1.8% relative to the baseline. Another estimate cited by Reuters puts the direct US hit from weaker AI productivity at 0.8%, with spillovers reaching Asian technology exporters.
This does not prove that an AI crash is imminent. It shows instead how concentrated the current growth engine has become. A smaller group of companies and countries is driving a large share of capital spending, making the global economy more sensitive to expectations in the technology sector. The faster investment rises, the greater the potential disruption if expectations turn.
“The global economy remains resilient,” IMF Managing Director Kristalina Georgieva said, but “this resilience is reaching its limits.”
The optimistic view is that such risks are manageable because AI investment is backed by genuine technological progress and strong corporate balance sheets. The more cautious view is that markets may be pricing benefits that will take years to materialize, if they materialize at all.
War turns growth into a distributional problem
The Middle East conflict adds a different kind of pressure. Higher oil prices act as a negative supply shock: they raise transport and production costs while reducing consumers’ purchasing power. The IMF said oil prices have remained above $100 a barrel in a more prolonged conflict scenario, complicating efforts by central banks to bring inflation down.
Energy exporters and countries positioned in the AI supply chain may gain from the shock or withstand it better. Import-dependent, lower-income economies face the opposite combination: more expensive fuel and food, weaker currencies, higher debt-service costs and less fiscal space to protect households. The IMF says advanced-economy bond yields have risen to multi-decade highs, reflecting inflation concerns, fiscal pressures and expectations of stronger future growth linked to AI. The IMF’s annual-meetings briefing warned that these pressures are converging.
This unevenness matters politically. Governments may be forced to choose between subsidizing energy, maintaining public investment and reassuring bond markets. Countries with limited borrowing capacity have fewer options, increasing the risk that the benefits of global growth remain concentrated while the costs are broadly shared.
What comes next
The immediate test is whether energy prices stabilize without a deeper escalation of the war. If they do, AI spending could continue to support activity and inflation could gradually ease. If the conflict intensifies, the energy shock could overwhelm the investment boost, forcing central banks to keep policy tighter for longer.
The second test is evidence. Investors and policymakers will look for measurable productivity gains rather than announcements about future applications. Governments can help by expanding electricity and digital infrastructure, improving competition and supporting worker retraining, while avoiding subsidies that lock in unproductive projects.
The central issue is therefore not whether AI is beneficial or harmful. It is whether productivity gains arrive quickly enough, and broadly enough, to offset the costs of war, debt and higher energy prices. For now, the global economy is growing. But its apparent stability rests on two unstable forces moving in opposite directions.