Can the AI workhorses carry the world's markets - again?
Big-cap stocks are heading into earnings with optimism creeping back. The question hasn't changed: will a small group of AI-heavyweights keep doing the heavy lifting, or does breadth finally show up?
Forecasts for 3Q have improved across the board, but the strength still funnels through a narrow set of tech leaders. After an easy 2Q beat, the bar is higher this time. We're about to find out if markets are broadening - or still leaning on the same few digital giants.
Relief rally, higher bar
Global earnings and sales estimates have risen over the past three months, helping fuel a sharp rebound. That looks more like relief than conviction: fears of a full trade war cooled, but fundamentals haven't cleared the runway.
Even with upgrades, forecasts sit below where they started the year. 2Q was an easy beat; 3Q isn't. As expectations climb into 4Q25 and 1Q26, repeat-sized beats - and the big price pops that followed - will be harder to land.
AI vs. everyone else: the catch-up is pushed to 2026
At the start of 2025, many thought the AI trade was stretched and non-AI earnings would catch up by year-end. That didn't happen. AI leaders blew past estimates in the first half, while non-AI cuts piled up amid tariffs and softer demand.
Now both AI and non-AI growth are set to cool into late 2025, but AI looks more durable thanks to capex and productivity demand. The convergence story has shifted to 2026, when non-AI growth is expected to accelerate as AI moderates - narrowing the gap, not erasing it.
US edge narrows as global revisions improve
The US 2026 earnings outlook had been stretching ahead, helped early on by tariff dynamics. Recent upgrades abroad narrowed that gap. Back in April, US growth was expected to beat emerging markets by 1.23x (13.5% vs. 11.3%) and other developed markets by 1.57x (8.6%).
After a volatile summer, those ratios widened to 1.34x and 1.9x. They've since tightened to 1.23x and 1.75x. Net: the US is no better positioned versus EM than before tariffs ramped, while developed ex-US remains more exposed if global growth stalls.
Sectors beyond tech step up in 2026
Tech still leads, but the gap narrows. Sector outlooks point to tech earnings up 24.6% in 2025 (after 27% in 2024), easing to 20.5% in 2026 - likely still the fastest, just by less.
Most other major sectors accelerate in 2026. Discretionary flips from a 5.4% contraction in 2025 to 16.8% growth in 2026. Staples, financials, health care, and industrials all pick up. Energy's decline moderates. Materials hold steady after a strong 2025. Communications is the outlier given its mega-cap, tech-adjacent names.
Margins: AI keeps the US out front
Margins remain a US advantage through 2026, helped by high-margin AI exposure that offsets tariff drag. US bellwethers are projected to lift margins 67 bps to 17.8% in 2025 and another 95 bps to 18.8% in 2026.
EM bellwethers could rise 127 bps in 2025 to 16.7%, then slip 13 bps in 2026. Developed ex-US may drop 36 bps to 14.3% in 2025, with a modest rebound to 14.9% in 2026. Within that, AI bellwethers jump 222 bps from 24.7% (2024) to 26.9% (2026), versus just 51 bps for non-AI (14% to 14.5%). That spread is a big piece of the US and EM margin lead.
The bellwether test
Watch a global bellwether basket - 133 of the world's largest companies across regions and sectors - for the earliest read on breadth. If beats and guidance expand beyond the AI core, this rally can broaden. If not, index leadership stays narrow and more fragile.
What to watch this season
- Guidance breadth: Do beats and outlooks extend beyond mega-cap AI to cyclicals and defensives?
- Capex and AI spend: Who's funding the stack (chips, data centers) vs. who's monetizing productivity gains?
- Margins vs. tariffs: Pricing, mix, and efficiency vs. cost pressure; inventory and lead times matter.
- Regional leadership: Are beats concentrated in the US, or do EM and developed ex-US close the gap?
Practical positioning
- Expect smaller beat rates and more single-name volatility as the bar rises.
- Keep core AI exposure; layer in quality non-AI cyclicals that re-accelerate into 2026 (industrials, financials, select staples and health care).
- Track discretionary's 2026 turn (from -5.4% to +16.8%): staged entries beat all-or-nothing bets.
- Re-check your US overweight. Select EM and developed ex-US can work where revisions inflect and FX isn't a headwind.
- Focus on margin durability: automation, unit economics, and pricing power over pure top-line hopes.
For a macro backdrop on growth assumptions and trade's drag or lift, see the IMF's latest growth outlook.
If you're in finance and building practical AI leverage, this curated set of tools is a fast filter: AI tools for finance.
Your membership also unlocks:
