The global economy is absorbing two powerful and opposing shocks: a war-driven energy crisis that raises costs and suppresses demand, and an artificial-intelligence investment boom that is lifting trade, construction and technology spending. For now, the second force is strong enough to offset much of the first. Whether that balance lasts is becoming the central economic question.
The International Monetary Fund expects global output to grow 3.1% in 2026 and 3.4% in 2027. The forecast is modest by historical standards, but it represents an upgrade for 2026 and reflects an economy that has proved more durable than many forecasters expected. The IMF says AI-related investment and trade have become a key source of momentum, while higher energy and commodity prices linked to the Middle East war are acting as a negative supply shock. IMF outlook
Why the economy has held up
The AI boom is not limited to software companies. It is generating demand for advanced chips, data centres, electricity, construction and financial services. In the United States, AI-related capital spending has accounted for just over half of GDP growth so far this year, according to Investec. South Korea and Taiwan are benefiting from semiconductor production, while European surveys show stronger activity in technology equipment and software services. Investec global overview
This investment has helped offset weakness elsewhere. Energy-importing economies are paying more for fuel, while households face pressure from higher transport and food costs. Yet governments have drawn on stockpiles and policy support, limiting the immediate pass-through of energy prices into broader inflation. The result is an uneven expansion rather than a synchronized recovery.
Some countries are positioned to gain from the boom. Economies producing chips, computing equipment or energy can attract investment and improve their trade positions. Others face a double burden: expensive imports and limited fiscal capacity to protect consumers or finance new infrastructure. The IMF describes the outlook as increasingly divided between countries benefiting from AI or favourable terms of trade and lower-income, energy-importing economies facing more difficult conditions. IMF assessment
The vulnerability beneath the headline growth
Resilience does not mean safety. Inflation is expected to rise to 4.7% this year before easing to 4.1% in 2027, with energy costs and AI-related demand contributing to the pressure. Higher prices complicate central banks’ efforts to reduce interest rates, while elevated borrowing costs worsen the strain on heavily indebted governments, businesses and households. CNBC summary of the IMF outlook
Financial markets are also questioning whether the AI expansion can justify its scale. The IMF has warned that expectations could prove excessive, triggering a correction in technology shares and related investment. Neocloud companies, whose business models depend on rapid demand for computing capacity, are already showing signs of strain. A market pullback would not automatically cause a global recession, but it could reduce corporate investment and weaken the confidence supporting the current expansion.
The United States illustrates the tension. Its economy is benefiting disproportionately from AI spending, but its federal deficit has widened by $217 billion to $1.992 trillion, according to reporting summarized by The Asian Banker. Rising Treasury yields could make the fiscal position more difficult and increase financing costs across the economy. The Asian Banker
What comes next
The immediate risk is a prolonged energy shock. Investec reports 32 shipping incidents linked to Iran this month, the highest since the conflict began, while attacks affecting Saudi energy facilities and airports are adding to regional uncertainty. If disruption spreads through major transport routes or production sites, oil and gas prices could rise enough to weaken consumption and revive inflation. Investec reporting
The more optimistic scenario is that AI investment evolves into measurable productivity growth. That would broaden the benefits beyond chipmakers and data-centre operators, allowing stronger output without equivalent inflation. The less favourable scenario is that spending remains concentrated, debt-financed and dependent on optimistic valuations. In that case, a correction would expose how much current growth relies on a narrow group of industries.
Policymakers therefore face a difficult trade-off. Tightening too aggressively could choke productive investment and deepen weakness in energy-importing economies; easing too quickly could entrench inflation and inflate financial risks. The next phase will depend less on the headline growth rate than on whether AI gains spread through wages and productivity, and whether geopolitical shocks remain contained.