The Global Economy Survived the Oil Shock. Not Everyone Did.

The world economy took the largest oil-market disruption in decades and kept growing. That is the headline. The bill landed somewhere else.

The International Monetary Fund's July 2026 World Economic Outlook Update projects global growth of 3.0 percent this year and 3.4 percent in 2027. Output does not fall off a cliff. The cumulative forecast is close to April's. Yet the same report expects global inflation to rise from 4.1 percent in 2025 to 4.7 percent in 2026, and it describes a world pulled in opposite directions by war and technology.

The aggregate survived. The damage did not disappear. It moved through fuel bills, food prices, currencies, interest rates and public budgets, hitting countries with the least room to absorb it. At the same time, an investment boom in artificial intelligence, semiconductors and data centres lifted a much smaller group of economies.

Global resilience is an average. In 2026, that average hides an economic split.

Two shocks, one number

The energy shock is real. The IMF's full July report assumes oil prices will average about $89 a barrel in 2026, 32 percent above 2025. It projects natural-gas prices up 22 percent, fertiliser up 26 percent and food up 8 percent. Those figures sit inside the same forecast that leaves cumulative global growth broadly unchanged from April.

Several shock absorbers bought time. Producers outside the Gulf raised output. Governments and companies drew down inventories. Demand fell. Renewable power has a larger share of the energy mix, and many economies use less energy for each unit of output than they once did.

That protection is temporary. An IMF oil-market analysis estimates that more than 1.1 billion barrels of crude had failed to reach the market by the end of May. Stocks and spare production prevented a far larger price spike, but those buffers are now thinner. Survival came partly from using the insurance.

Prices also travel unevenly. Since the war began, retail petrol prices rose about 30 percent in emerging Asia and 15 percent in Latin America, according to the IMF. Asian liquefied-natural-gas prices climbed about 50 percent, compared with 25 percent in Europe and roughly 10 percent for US Henry Hub gas. Geography matters. So do contracts, reserves, taxes, subsidies and the currency used to pay for imports.

One global oil benchmark cannot describe those different bills. One global growth rate cannot describe what follows.

The technology lifeboat has a short passenger list

The strongest offset came from countries already wired into the AI hardware cycle.

The IMF found that the four leading net exporters of AI-related hardware — Taiwan, South Korea, Thailand and Malaysia — beat its first-quarter growth expectations by an average 4.4 percentage points on an annualised basis. The surprise for the rest of the world was negative 0.3 point. South Korea grew at a 7.5 percent annualised rate in the quarter, more than four times the IMF's April projection, powered mainly by semiconductors and AI hardware.

That split continued into the annual forecast. The Fund expects South Korea to grow 2.6 percent in 2026 as chip demand outweighs its dependence on imported Middle Eastern energy. Malaysia benefits from data-centre activity and the technology upturn. Thailand and Vietnam receive upgrades tied partly to technology exports and investment. The United States gets support from technology-related business investment while its status as a net energy exporter limits the war's direct hit.

These countries did not win merely by “adopting AI.” They entered the boom through factories, chip packaging, export networks, power systems, capital markets and firms able to spend at enormous scale. The distinction matters because most countries cannot copy those assets on demand.

The International Energy Agency says global data-centre investment nearly doubled between 2022 and 2024, reaching about $500 billion. The United States held 45 percent of global data-centre electricity use in 2024, China 25 percent and Europe 15 percent. Emerging and developing economies outside China contain half of the world's internet users but less than 10 percent of its data-centre capacity.

Reliable electricity is an entry ticket. Grid connections, skills and affordable capital are others. The IEA estimates that one-fifth of planned data-centre projects could face delays if grid constraints are not fixed. UN Trade and Development warns that investment is clustering in AI, semiconductors and other strategic industries while roughly 75 percent of foreign direct investment to developing economies goes to only ten countries.

The boom is global in valuation and demand. Its construction jobs, export income, tax base and infrastructure are far more local.

Energy importers pay through every door

An energy importer first pays through the trade balance. More dollars leave to buy the same fuel. The currency can weaken, making the next shipment more expensive. Transport and fertiliser costs then move into food prices. Central banks face higher inflation and may keep rates high. Governments can subsidise fuel or food, but countries with thin budgets borrow that relief from another priority.

Wealthier states have reserves, deeper capital markets and more credible currencies. Poorer importers often have none of the three. The IMF expects some developing economies without stocks to face shortages while competing with richer buyers for available cargoes. It also warns that higher food and energy prices can widen external deficits, trigger capital outflows and intensify debt refinancing pressure.

The mechanism reaches households quickly. A World Bank analysis of African transport notes that road transport consumes roughly 80 percent of imported fuel across the continent and that petrol and diesel imports cost a median 3.6 percent of GDP. In economies where households spend close to 40 percent of their income on food, higher freight and fertiliser costs do not look like a modest adjustment in a forecast table. They mean fewer meals, less travel and less money for school or medicine.

Even the country categories resist easy celebration. Gulf producers closest to damaged infrastructure lost export volume despite higher prices. The IMF expects Iraq, Kuwait and Qatar to contract sharply in 2026. China gains from high-tech manufacturing but still absorbs dearer imported oil and weak domestic consumption. India remains fast-growing because services and consumption are strong, not because the energy shock vanished.

There are no clean camps. Exposure and protection stack differently in every economy.

GDP can hold while people lose

Gross domestic product measures output. It does not tell us who can still afford food, who lost hours at work or which clinic lost funding when a government protected fuel prices.

The IMF projects sub-Saharan Africa to grow 4.3 percent in 2026, the same as its regional forecast in April. Inside that stable number, oil-importing economies are hit harder, official development assistance is declining, and higher prices for essentials are expected to aggravate poverty and food insecurity in Nigeria. A regional average can stay upright while vulnerable households fall.

Labour-market scenarios make the transmission visible. In May, the International Labour Organization modelled an illustrative case in which oil prices remain about 50 percent above their early-2026 average. It estimated a loss of working hours equivalent to 14 million full-time jobs in 2026 and 38 million in 2027, with real labour income falling 1.1 percent and 3 percent. Those are conditional estimates, not observed losses and not the IMF baseline. They show how an energy price can become a wage shock even when total world output keeps rising.

Hunger models carry the same warning. The World Food Programme estimated that 45 million more people across 53 countries could face acute hunger if the conflict continued through the second quarter and oil stayed above $100. The IMF's July baseline uses a lower average oil price, so the WFP figure must not be presented as an outcome. It marks the human scale of the downside.

Global growth answers one question: how much did the world produce? It cannot answer whether the people with the smallest buffers absorbed the largest loss.

What breaks if chips stop carrying the average

The IMF baseline assumes the AI-driven technology cycle will moderate without delivering an extra productivity boost. Even that restrained assumption leaves technology investment doing important work against the energy drag.

A sharper reversal would remove the offset. If investors cut their expectations for AI profits, technology spending could retreat, chip exports could weaken and concentrated equity markets could correct. The IMF traces the next steps through lower household wealth, cross-border portfolios, capital flows and tighter financial conditions. The damage would spread beyond technology.

Keep the energy crisis running at the same time and the split becomes a global slowdown. Import-dependent economies would still face expensive fuel and food. Technology exporters would lose orders. Governments that already spent fiscal room cushioning the first shock would have less capacity for the second. Countries outside both the energy-export and technology-production clubs would receive no windfall at all.

This is a risk scenario, not a forecast. It matters because the 3.0 percent baseline depends on assumptions that can fail: shipping through the Strait of Hormuz begins to normalise, inventories keep doing their job, financial conditions remain supportive and the technology cycle cools rather than collapses.

The global economy survived the first oil shock on those terms. The next test is whether governments mistake the average for a shared recovery. Countries that own chips, compute, power and financial buffers can catch the upside. Countries buying fuel in dollars with crowded budgets carry the downside. People living inside those economies experience the divide long before it appears in the global total.