- Summer brought little of the calm markets had priced in before it began. Three tensions we had been tracking escalated through the season, surfacing as fresh flashpoints by early September: a fresh energy shock out of the Middle East, a bond sell-off that has spread from the US to Japan to Europe (and drawn erratic policy responses from central banks and governments alike), and renewed doubts about the AI boom – talk of chip and power shortages, bubble concerns and underwhelming returns on frontier-model spend.
- But none of this has yet broken the real economy. Q2 growth surprised to the upside, carried by net exports and Germany's stimulus in Europe, and by robust AI-related investment in the US and Asia. Forward-looking indicators such as European manufacturing PMIs at four-year highs and strong US Q3 nowcasts show little sign of deterioration, albeit with marked divergence across sectors.
- The Great Unwind: the case for an orderly path forward. Our baseline continues to be that these three tension points will ease rather than escalate, but that call rests on specific mechanisms. In the report we also model three downside scenarios and their economic and capital market impacts, each based on a disorderly release of the tensions.
- The energy shock risk. We expect the conflict to stay contained and that rerouting, demand destruction and alternative supply do the job energy markets need them to do, easing prices and letting the inflation shock fade. This rests on the assumption that the Saudi pipeline outage does not extend past a month. If the outage runs longer, the shortfall could exceed 8Mb/d and Brent could stay near USD100/bbl, turning what is now a contained shock into a stagflationary shock.
- The bond sell-off risk. We assume inflation moderates without meaningful second-round wage effects, and that sovereign markets can absorb elevated issuance without a disorderly rise in risk premia. The assumption doing the real work here is political: the US midterms don't produce further fiscal slippage or erratic policy interventions and European elections (France, Italy) do not add fiscal uncertainty. Note that the real rate component of the US 10 treasury rates has increased by a full percentage point this past year confirming that bond markets also price more AI growth. In the downside, self-reinforcing yield–fiscal loop can take hold and force central banks to stabilize bond markets.
- The AI capex cycle risk. We think that AI capex decelerates slightly to peak at USD1trn in 2028, with value creation shifting from infrastructure build-out toward adoption and productivity. This rests on the assumption that infrastructure and power bottlenecks stay manageable, financing conditions don't tighten sharply and AI earnings keep validating the spend. A sharper investment retrenchment, given how much of US and Asian growth and equity performance now runs through AI, could spill from technology into corporate investment and trade, and from wealth effects into consumption.
- Our scenario in numbers: Global GDP growth slows from +3.0% in 2025 to +2.6% in 2026, recovers to +2.8% in 2027, and eases again to +2.6% in 2028. US growth moderates from +2.1% (2026) to +1.8% (2028); the Eurozone recovers from +1.0% to +1.3%; Asia stays the main engine at around +4.4%. We forecast Brent to average 87 USD/bbl in Q4 2026 before easing to 76 USD/bbl on average in 2027 as Gulf export flows gradually normalize both in and around Hormuz but refined-product markets are running out of buffers, and prices should remain elevated for the next few quarters. European gas prices should stay elevated through year-end, with TTF averaging around EUR65/MWh in Q4 2026, before easing to about EUR36/MWh in 2027. Inflation stays 1–2pps above target in Europe and the US until H2 2027, which is why we still expect one more Fed hike to 4.00–4.25% and two more ECB hikes to 3.00% by year-end, with cutting cycles starting only once inflation is back at target – earliest end-2027 – and terminal rates of 3.50% (Fed) and 2.25% (ECB) reached in 2028.
- What it means for markets. We see bond yields dropping when markets start pricing in our central bank scenario of a lower terminal rate in both Europe and the US. The 10y Bund is expected to gradually drop from current levels to 3.3% at the end of 2026 and further to 3.0% in 2028 (US 10y from 4.7% end of 2026 to 4.3% in 2028). Credit spreads will slightly widen from current low levels given strong supply from the technology sector. Equity markets will finish the year with double digits again but decelerate to high single digits in 2027 and 2028, with valuations normalizing to historic levels (S&P500 2026-2028: 14%, 8%, 8%, Eurostoxx: 15%, 9%, 8% total return). The euro will gain gradually against the dollar, reaching 1.18 in 2028 as policy uncertainty and a weaker economy in the US will weigh on the greenback. Private markets are entering a more selective phase: Private equity faces a record ~13,300-company exit backlog, while infrastructure benefits from the AI buildout, with data-center investment expected at USD 5–7trn this decade.