The global economy is entering 2027 with an unusual split personality: growth remains resilient, but the forces supporting it are becoming narrower and more vulnerable. The latest forecasts from the OECD and International Monetary Fund point to continued expansion, while warning that artificial-intelligence investment, energy shocks and geopolitical fragmentation are reshaping who benefits—and who absorbs the risks.
The OECD projects global growth of 2.9% in 2026 and 3.0% in 2027, below the 3.4% recorded in 2025. The IMF is more optimistic, forecasting 3.1% growth this year and 3.4% next year. The difference reflects contrasting assumptions, but both institutions describe an economy that is holding together rather than accelerating broadly.
AI has become the main growth engine
Investment linked to AI is now doing more than transforming technology companies. It is supporting manufacturing, trade and business spending across several major economies. OECD data cited in an October statement show computer and electronics production in the 12 months to July grew 18% in the European Union, 11% in the United States, 19% in China and 12% in other major Asian economies.
That concentration is significant. The world economy is not being lifted evenly by household consumption or public investment; it is being pulled forward by data centres, advanced chips, network equipment and the companies financing them. The IMF says stronger AI investment has helped sustain demand and may eventually produce productivity gains. In that view, today’s capital spending could become tomorrow’s broader improvement in wages and output.
There is a less reassuring interpretation. The OECD warns that rapid AI investment is increasingly reliant on external financing. If projected returns fail to materialise, a market correction could spread beyond technology shares into credit markets and corporate investment. The IMF has likewise highlighted the risk of a reassessment of AI expectations.
“Robust AI-related activity” is supporting investment, production and trade, while external financing could amplify a correction if expected returns do not materialise, the OECD says.
Energy and geopolitics expose the limits
The expansion has also survived a major energy shock linked to conflict in the Middle East more effectively than many expected. The OECD attributes part of that resilience to sizeable oil inventories, supply from outside Gulf producers and discretionary government support. Those buffers reduced the immediate damage to consumers and businesses.
They do not eliminate the underlying vulnerability. Higher energy prices raise transport and production costs, weaken real incomes and complicate central-bank decisions. The OECD expects G20 inflation to rise to 4.1% in 2026 before easing to 3.6% in 2027. The IMF similarly projects global headline inflation at 4.7% this year and 4.1% next year.
For policymakers, the challenge is therefore asymmetric. Supporting households against energy costs can prevent a sharp downturn, but prolonged fiscal support may reinforce inflation. Higher interest rates can restrain prices, yet they also make AI infrastructure, housing and government debt more expensive. The result is a narrower margin for error than the headline growth figures suggest.
A divided recovery
Countries with technology exports, favourable trade positions or the capacity to subsidise energy have more room to adapt. Others face higher import bills, weaker demand and less fiscal space. The IMF describes this as a “split global economy,” in which technology-related investment cushions some economies while geopolitical and commodity shocks deepen pressure on others.
That divide also complicates international cooperation. Governments want access to advanced chips, data infrastructure and domestic production, but efforts to secure supply chains can increase costs and reinforce trade barriers. The OECD says technology-related trade has remained strong, while its broader outlook warns that tariffs, production-chain adjustments and uncertainty could weigh on growth.
Businesses face a similar choice. Delaying AI investment could mean falling behind competitors, particularly as productivity gains begin to appear. Moving too aggressively risks building capacity on assumptions about demand and efficiency that have not yet been tested at scale. Investors, meanwhile, must distinguish between companies selling essential infrastructure and those valued mainly on future promises.
What comes next
The central question for 2027 is whether AI investment broadens into durable productivity growth or remains a concentrated capital cycle. A benign outcome would combine easing energy prices, lower inflation and measurable gains in output per worker. That would validate the current spending boom and allow central banks to reduce pressure on borrowing.
A harsher outcome would involve renewed energy disruption, persistent inflation or a sharp reassessment of technology valuations. In that scenario, the same concentration that currently supports growth could magnify the downturn. Governments would then face pressure to protect strategic industries and households without deepening trade fragmentation.
The forecasts do not describe an economy in crisis. They describe one increasingly dependent on a small number of favourable conditions. Growth is holding up, but the next phase will test whether AI can become a broad productivity story—or whether it is mainly postponing the consequences of weaker, more uneven foundations.
Sources: OECD Economic Outlook, September 2026; IMF World Economic Outlook, October 2026; OECD statement to the IMF International Monetary and Financial Committee.