World Bank upgrades US growth forecast as tariff concerns recede

The World Bank upgraded US growth sharply, raising the annual forecast by 0.6 percentage points to 2.2% and revising a US 2025 projection up to 2.1% (from 1.4%), citing surging AI-related spending and a smaller-than-expected tariff drag. Global growth is modestly weaker — average growth around 2.6% this year — with China forecast to slow to about 4.4% (and 4.2% in 2027) amid weak consumer confidence and a property downturn; the eurozone and Japan show subdued expansion. The bank attributes limited tariff impact to front‑loading of trade, delayed implementation and tariff mitigation strategies, implying continued US outperformance but heightened downside risks from China’s slowdown for markets and EM exposure.

Analysis

Market structure: The World Bank's 0.6ppt US upgrade to 2.2% and revised 2025–26 outlook shows demand concentrated in AI-driven capex and logistics/reshoring. Direct winners are AI compute leaders (NVDA, MSFT, GOOGL), semiconductor-equipment suppliers (LRCX, AMAT, ASML) and logistics/industrial names that capture reshoring volumes; losers are China-exposed cyclicals, commodity exporters tied to Chinese construction, and pockets of EM banking/property. Front‑loaded trade ahead of tariffs implies a one-time boost in manufactured goods flows rather than durable export growth.

Competitive dynamics & cross-asset: AI spending increases pricing power for GPU/cloud providers and extends semicap order books 2–4 quarters, compressing margins for laggards. On macro cross-assets, stronger US growth and capex tilt push UST yields higher (pressure on 5–10y), support a stronger USD (benefit UUP/short EM FX), and lift industrial commodities (copper, semiconductor gases) while reducing safe-haven demand for long-duration bonds.

Risks & timing: Tail risks include tariff re‑escalation, a China hard landing or major property default — each could shave >1pp off global growth in 6–12 months. Immediate (days) drivers are tariff/news headlines and PMIs, short-term (weeks–months) are quarterly earnings/capex announcements, and long-term (2–5 years) is diffusion of AI across sectors. Hidden dependency: much of the upgrade is front‑loaded trade and concentrated capex; if AI spend narrows to a few firms, broader job/income effects will lag.

Contrarian view: Consensus underestimates fading momentum after front-loading; market is likely to rotate from a narrow AI winner list into broader cyclicals only if capex proves broad-based. Historical parallel: past capex-led cycles (semicap booms) produced sharp mid-cycle corrections after inventory rebalancing. Unintended consequence — stronger yields from outsize US growth can derate high-multiple growth stocks even as revenues rise.

Key Decisions for Investors

• Establish a 2–3% portfolio long in NVIDIA (NVDA) via a 3–6 month 10%–15% OTM call spread to capture AI capex upside while capping premium; trim/exit if NVDA total return falls >20% or implied vol spikes >40%.
• Build 1–2% positions in semiconductor-equipment names (Lam Research LRCX and Applied Materials AMAT, 0.5–1.0% each) and a 0.5–1% position in ASML to play extended order books; target 12–18 month hold and take profits on 20–30% gains or order‑book misses.
• Reduce core duration: sell ~50% of long Treasury exposure (or short 7–10yr futures equal to 0.5–1% portfolio risk) and redeploy into 1–2% floating-rate notes (FRN) and 1–2% long USD (UUP). Add to duration short if 10y yield breaks above 4.25%.
• Hedge China/EM risk with a 1–2% short FXI position or 3-month put spread on FXI (buy 5% OTM puts, sell 10% OTM) and scale to 3% if China PMI <48 or a major property developer defaults — lock this hedge for 3–6 months.