The Global Economy at Halftime 2026: Three Shocks, Two Reversals, One Fragile Recovery

Today is the first day of the second half of 2026. The first half is now a data set, not a forecast, and the data tells a story that no one predicted in January. At the start of the year, the consensus was benign: the 2025 rate-cutting cycle would continue, inflation would converge toward 2%, and global growth would hold around 3%. Two of those three predictions were wrong before February ended.

The Bank for International Settlements titled its 2026 annual report “Progress and Peril.” That may be the most accurate two-word summary of the first half. Progress: the global economy did not collapse despite an oil shock, tariff escalation, and monetary whiplash. Peril: the growth that survived is narrower, more concentrated, and more dependent on a handful of companies, sectors, and policy choices than at any point since the pandemic.

Here is the halftime scorecard.

Shock 1: Hormuz Closed, Oil Spiked, Inflation Returned

The defining event of H1 2026 was the closure of the Strait of Hormuz in late February. Approximately 21% of global oil supply and 25% of LNG shipments pass through the strait. When military conflict disrupted transit, Brent crude spiked from $78 to $111 per barrel within weeks. The consequences cascaded through the global economy with a speed that surprised even those who had modelled energy supply shocks.

US headline CPI rose to 4.2% by May. Eurozone HICP hit 3.2%. Urea fertiliser prices surged 46%, feeding into food price inflation that was disproportionately severe in low-income countries where households spend 40–60% of income on food. The Iranian economy contracted an estimated 6.1% as sanctions tightened and oil exports were physically disrupted. Iraq, Qatar, and Bahrain all entered contraction.

A peace deal in mid-June brought Brent back to approximately $83 — but the damage was done. Three to six months of elevated energy costs were already embedded in supply chains, input prices, and inflation expectations. Central banks that had been cutting rates in 2025 found themselves forced to reverse course, and the reversal defined the rest of the half.

Shock 2: The Rate-Cutting Cycle Died in Under Six Months

In 2025, nine out of ten major central banks cut rates, delivering 32 total cuts and 850 basis points of cumulative easing — the most since the 2008–2009 financial crisis. The Fed cut 75bp. The ECB cut 100bp. The Bank of England cut 100bp. The Swiss National Bank cut 75bp, reaching 0%. Among major banks, only the BOJ hiked in 2025 — by 50bp, beginning its exit from decades of zero rates.

By mid-2026, the cycle had reversed. The ECB hiked to 2.25% on June 11. The BOJ raised to 1.00% on June 16. The RBA added 25bp to reach 4.35%. The Fed held at 3.50–3.75% but 9 of 18 FOMC members signalled at least one hike before year-end. KPMG’s June 2026 Global Economic Outlook declared the “rate-cutting cycle over.”

The reversal was faster than any peacetime precedent. It took 18 months to build the easing cycle and fewer than six to unwind it. The trigger was Hormuz-driven energy inflation, but the underlying vulnerability was that the “last mile” of the post-pandemic disinflation had never been completed. Core services and shelter inflation in most advanced economies never fell to 2% during the cutting cycle. When energy prices spiked, headline inflation surged past targets that had only been approached, never reached.

Shock 3: AI Investment Masked a Consumer Recession in the United States

The US Q1 GDP third estimate was revised upward to 2.1% from 1.6%, which looked reassuring until you examined the composition. Consumer spending — 70% of the American economy — was revised downward to 0.5%, the weakest since Q2 2023. What saved the headline was business equipment investment (+15.8%) and intellectual property products (+13.8%), both driven overwhelmingly by AI data centre construction and chip purchases.

Pantheon Macroeconomics calculated that without AI-related spending, US corporate equipment investment would be negative. Five hyperscalers — Amazon ($200B), Google ($175–185B), Microsoft (~$150B), Meta ($115–135B), and Oracle (~$65B) — are spending approximately $725 billion in combined capex. Morgan Stanley estimates total hyperscaler capital expenditure at $805 billion. Apollo calculated that hyperscaler capex alone equals roughly 2% of US GDP.

The concentration is historically unprecedented. Five companies are driving approximately 60% of US nonresidential fixed investment growth. The consumer is flagging: the Conference Board expectations index fell to 53.3 in June (lowest since January 2014), University of Michigan sentiment has been below 60 for 24 consecutive months, and the savings rate has risen to 4.0% as households shift toward precautionary behaviour. The US economy is not in recession, but it is running on one engine.

The H1 Scorecard: Winners and Losers

The pattern is striking. The economies that outperformed in H1 share one or both of two characteristics: they are plugged into the AI semiconductor supply chain (South Korea, Vietnam, Taiwan), or they are large domestic markets with strong services growth and limited energy import dependence (India, Spain). The economies that underperformed are either direct casualties of the Hormuz crisis (Iran, Gulf states) or industrialised economies hit by the combination of higher energy costs, tighter monetary policy, and structural manufacturing decline (Germany, France, Canada).

The eurozone divergence is now the widest since the introduction of the euro. Spain is growing at 2.5% while Germany is barely above zero and France is flat. The ECB’s single interest rate must serve both — a structural impossibility that explains why the eurozone aggregate of 1.1% growth is analytically useless. It conceals two entirely different economic realities sharing a currency.

The China Paradox: Growing and Stagnating Simultaneously

China does not fit neatly into the winner/loser framework because it is both. Q1 GDP came in at exactly 5.0%, powered by exports that surged 53.2% in May (headline) on the back of AI chip and semiconductor demand. The June PMI data confirmed the bifurcation: high-tech manufacturing at 53.5 (strong expansion) while employment contracted for a fifth month at 48.4. Property prices fell 8.3% in Beijing. Consumer confidence dropped to 89. Youth unemployment sat at 15.6%.

China’s export economy is booming. China’s domestic economy is deflating. The aggregate 5.0% growth rate, like the eurozone’s 1.1%, conceals more than it reveals. The structural question — whether AI-driven high-tech manufacturing can absorb workers shed from property, construction, and consumer-facing sectors — will not be answered in H2. It may not be answered this decade.

The Numbers That Didn’t Make Headlines

Three data points from H1 deserve more attention than they received:

The yen hit a 40-year low. Japan’s currency touched 161.95 per dollar on June 30, its weakest since December 1986. The Ministry of Finance spent a record $73.6 billion on intervention in a single month. It did not work. The carry trade — estimated at $300–500 billion in outstanding positions — continues to weaken the yen despite BOJ rate hikes. J.P. Morgan projects 164 by Q4. A disorderly unwind remains the single largest tail risk in global markets.

23 African countries entered debt distress. Interest payments now consume 27.5% of government revenue across the continent, up from 19% in 2019. The interest payment crisis is not limited to Africa, but it is most acute there. The G20 Common Framework, designed to coordinate sovereign debt restructuring, has delivered results for Zambia and Ghana but at glacial speed — Zambia’s restructuring took four years. With $12.1 trillion in non-OECD sovereign bonds outstanding and 24 countries facing 50%+ debt maturity by 2027, the refinancing cliff is approaching faster than the institutional machinery can process.

The semiconductor market is ahead of schedule. WSTS forecasts $1.5 trillion in semiconductor revenue in 2026 — reaching the $1 trillion milestone four years earlier than consensus expected. AI chips account for roughly half of revenue but less than 0.2% of unit volume. The supercycle has made entire national economies — South Korea, Taiwan, Malaysia — dependent on a single supply chain. The concentration risk is not theoretical. If AI capex retrenchment occurs, it will be felt from Suwon to Penang to Taipei to Austin simultaneously.

What H2 Depends On: Three Dates

The second half of 2026 will be shaped by events that are already scheduled. Three dates matter more than others:

July 22: IMF World Economic Outlook update. The April WEO projected global growth of 3.1%. The Hormuz peace deal, the rate reversals, and Q1 GDP data from major economies will all feed into revised forecasts. A downgrade below 3.0% would be the weakest non-crisis IMF projection on record. The EM growth downgrade from 4.2% to 3.9% — already signalled — will confirm what the World Bank described as emerging markets’ weakest per capita income growth since the pandemic.

July 30–31: Back-to-back Fed and BOJ meetings. The Fed meets July 30. The BOJ meets July 31. Together, these two decisions will determine whether the 275bp rate gap that is driving the yen carry trade widens, narrows, or holds. If both surprise hawkish, the yen could recover 5–10% and trigger partial carry trade unwinding. If both hold, the yen weakness persists and the carry trade accumulates further.

Mid-July onwards: Q2 GDP data. Advanced economies will begin reporting Q2 GDP in the second and third weeks of July. This is the first quarter that fully captures the Hormuz spike, the rate reversals, and the tariff effects of Liberation Day. If Q2 data shows the eurozone contracting, Canada deepening its recession, or more countries entering recession, the narrative will shift from “resilient global economy” to “synchronised slowdown” — the framing that moves bond markets and policy expectations.

The Halftime Verdict

The global economy at halftime 2026 is not in crisis. Global growth at 2.5–3.1% is positive. Unemployment in most advanced economies remains historically low. The Hormuz peace deal has removed the most acute supply-side risk. Corporate earnings in Q1 were strong, with over 80% of S&P 500 companies beating estimates.

But the growth that exists is fragile in ways that January’s forecasts did not anticipate. It is concentrated in a handful of AI-linked sectors and companies. It is unevenly distributed, with India and Southeast Asia accelerating while Europe stagnates and the Middle East contracts. It depends on a carry trade that could unwind violently, on AI investment that five companies could retrench, and on a peace deal whose durability is untested.

The BIS called it “progress and peril.” The progress is real. So is the peril. The second half will determine which one defines 2026.

Explore the full dataset across 218 countries on our global rankings page, or compare specific economies using the country comparison tool. For GDP data by country, see our largest economies and fastest-growing economies pages.