The global economy is entering a more unstable phase in which two powerful forces are pulling in opposite directions. Artificial-intelligence investment is supporting demand, productivity expectations and financial markets, while conflict-driven energy disruption is pushing up inflation and squeezing households. The result is not yet a global recession, but a recovery increasingly dependent on whether the AI boom produces real economic gains before higher borrowing costs expose its vulnerabilities.
The International Monetary Fund raised its 2026 global growth forecast to 3.1%, up from its July estimate of 3.0%, but down from 3.5% in 2025. The upgrade reflects stronger investment and fiscal support, particularly in AI-related infrastructure. At the same time, the Fund warned that economic resilience may be approaching its limits, with energy costs and public debt narrowing governments’ room to respond. The IMF said the economy was facing a “negative energy supply shock” alongside a “positive AI demand shock.”
Why growth has held up
AI spending has moved beyond software demonstrations into large-scale construction of data centres, semiconductor capacity and electricity networks. That investment is creating demand across technology, engineering, construction and energy markets. The IMF expects India to grow by 7.0% in both 2026 and 2027, supported by domestic demand, services and AI-linked investment. China’s third-quarter growth reached 4.5%, according to reported official data, offering evidence that industrial activity remains stronger than some forecasters expected.
Supporters of the current expansion argue that the investment cycle could eventually raise productivity well beyond the companies directly producing AI systems. Faster logistics, automated customer service, scientific research and industrial design could reduce costs across the economy. A survey cited by the Financial Times found economists expecting AI’s benefits to extend beyond the United States and China.
That optimism, however, rests on a difficult transition. The infrastructure spending is happening now; the broad productivity gains may take years to appear. Companies are committing enormous sums before it is clear which business models will generate durable returns. That gap makes markets sensitive to any sign that demand, financing or technological progress is weakening.
The cost of the energy shock
War-related disruption in the Middle East has added a second pressure. Higher oil prices are feeding into transport, manufacturing, food production and household energy bills. US consumer prices rose 0.6% in September, lifting annual inflation to 3.7%, while eurozone inflation reportedly reached 3.8%, its highest level in three years. Those increases complicate central banks’ efforts to lower interest rates.
Energy inflation is especially damaging because it reduces purchasing power while raising companies’ costs. Governments can cushion the impact through subsidies or transfers, but those measures add to public borrowing. Countries already carrying high debt therefore face an uncomfortable choice: accept weaker demand, spend more to protect consumers, or tolerate inflation for longer.
Lower-income economies are more exposed. They typically import a larger share of their energy and food, have less fiscal capacity and can suffer from capital outflows when investors move toward safer assets. A G7 pledge of $2.5 billion to address food insecurity reflects concern that energy and fertiliser costs could spread the shock beyond financial markets.
The financial fault line
The most immediate risk is not simply that AI investment slows, but that its financing structure amplifies a downturn. Investors have begun questioning debt-funded data-centre expansion and complex links between chipmakers, cloud providers and AI start-ups. A sharp sell-off in AI-related shares was intensified by higher US inflation data, while the Philadelphia Semiconductor Index recorded seven consecutive daily declines through mid-October, Reuters reported.
The IMF has modeled a harsher scenario in which expected AI productivity gains fail to materialise. Under that scenario, US equities could fall 20%, global growth could be 1.1 percentage points below its baseline by 2028, and Asian growth could be 1.7 points lower. These are not forecasts of what will happen, but they show how much current valuations depend on future technological success.
There are competing interpretations. Bulls see volatility as a normal repricing after an extraordinary rally and argue that the underlying investment is building infrastructure with long-term value. Skeptics see circular financing, concentrated market leadership and optimistic revenue assumptions as signs of a bubble. Both views can be true: productive technology can generate excessive prices when investors overestimate how quickly its benefits will arrive.
What comes next
Central banks will face the hardest policy environment since the post-pandemic inflation surge. Cutting rates too quickly could entrench energy-driven price increases; keeping them high could weaken housing, business investment and the AI buildout. Governments will also confront pressure to subsidise energy, protect food supplies and support strategic technology sectors without worsening debt risks.
The decisive test will be whether AI investment begins to produce measurable productivity growth and broad wage or output gains. If it does, the current expansion may prove more durable than markets fear. If not, expensive infrastructure, elevated rates and weaker consumer purchasing power could turn today’s investment engine into tomorrow’s financial drag.
For now, the global economy is neither collapsing nor secure. Its apparent resilience depends on a race between technological gains and the costs of war, inflation and debt. The outcome will shape not only markets, but also the political tolerance for another period of unequal growth.