IMF's July WEO: AI Economies Pull Ahead as $89 Oil Squeezes Importers

The IMF's July 2026 World Economic Outlook Update projects global growth of 3.0% in 2026 and 3.4% in 2027 — numbers that look almost identical to what the Fund forecast back in April. On the surface, that stability suggests a world economy holding steady. But the aggregate figure is masking a more consequential shift happening underneath it: a split between two distinct groups of economies.

On one side are countries plugged into the AI and technology supply chain, which are benefiting from what the IMF describes as "accelerated demand-driven momentum in the global technology cycle" — investment and adoption tied to artificial intelligence. On the other side are energy-importing economies, contending with an assumed oil price of $89 per barrel and rising food costs that the IMF says are the main drivers pushing headline global inflation up from 4.1% in 2025 to 4.7% in 2026, before it eases to 3.9% in 2027.

The IMF frames the overall picture as a modest slowdown reflecting the economic effects of the Middle East war, offset in part by the AI-driven technology boost. Advanced economies overall are projected to grow 1.7% in 2026 and 1.8% in 2027 — but that average blends economies riding the tech cycle with those absorbing the full weight of costlier energy and food. The headline number held; the story underneath it did not.

The IMF's July update, following a press briefing held July 8, 2026, is notable less for what changed in the topline numbers than for what it reveals about how growth is being distributed. Global output at 3.0% in 2026 and 3.4% in 2027 is essentially a rollover of the April forecast. But that stability is the product of two offsetting forces pulling in opposite directions — and the IMF is explicit that they are not evenly distributed across countries.

The upward force is the technology cycle. Mechanically, this shows up in a few channels: capital expenditure on data centers, semiconductors, and related infrastructure; elevated demand for chips and computing hardware that flows through export- and manufacturing-oriented economies in the supply chain; and productivity effects as firms integrate AI tools into existing operations. Economies with meaningful exposure to this supply chain are capturing investment and demand that economies outside that chain simply are not.

The downward force is energy and food costs. The IMF's forecast is built on an assumed average oil price of $89 per barrel, a level that keeps sustained pressure on any economy that is a net importer of oil. This is a straightforward terms-of-trade mechanism: when a country imports more energy than it exports, a sustained higher oil price acts like a tax on the entire economy, raising input costs across transport, manufacturing, and agriculture, squeezing real incomes, and widening trade deficits.

Put together, this means the "global recovery" implied by a stable 3.0-3.4% growth path is not a shared experience. Advanced economies as a group are projected to grow 1.7% in 2026 and 1.8% in 2027, but that average conceals a split between economies effectively subsidized by AI-linked investment and demand, and economies effectively taxed by energy dependence.

The July 2026 WEO update is best read through the lens of terms-of-trade economics rather than as a simple revision of growth arithmetic. The literature on oil-price shocks has long distinguished between net oil-exporting and net oil-importing economies: a rise in oil prices functions as a positive terms-of-trade shock for exporters and a negative terms-of-trade shock for importers. The IMF's $89/barrel assumption is the mechanical link between the oil market and the 4.7% inflation projection for 2026.

On the other side of the divergence, the AI/technology investment cycle interacts with national accounts primarily through the capital expenditure and productivity channels. Sustained investment in data-center construction, semiconductor fabrication capacity, and related infrastructure shows up directly in gross fixed capital formation, while the demand pull for chips and hardware transmits through export receipts in economies positioned upstream or midstream in that supply chain.

A second-order issue worth flagging is that "tech-integrated" is not a uniform category. Even within economies benefiting from the AI cycle, the gains are plausibly concentrated in specific firms, regions, and skill segments — a winner-take-most dynamic consistent with broader patterns in digital and platform-driven growth.

The policy implications diverge accordingly. For energy-importing economies, the persistent $89/barrel assumption argues for fiscal buffers, targeted subsidy or transfer mechanisms, and diversification of energy sourcing where feasible. For tech-exposed economies, the policy challenge is managing overheating risk in the sectors absorbing AI-linked investment, monitoring valuation and leverage in AI-related capital markets for bubble dynamics, and ensuring that investment-driven growth translates into broader productivity gains rather than remaining narrowly concentrated.