Global macro outlook: coexistence of higher equity prices and Treasury yields amid AI investment, fiscal pressure and geopolitical risk: Equities and long-end yields can rise together, supported by AI capex and resilient earnings—for now

Overview

This global macro conference summary presents the view that equity gains and higher Treasury yields can coexist, because resilient earnings and AI-led investment have become less sensitive to rates. It also sets out the conditions that could disrupt that coexistence: a disorderly rates backup, persistent inflation, fiscal stress, geopolitical escalation, and AI infrastructure or monetization failures.

Core views

The central thesis is that two seemingly conflicting outcomes can coexist: the S&P 500 can continue rising while long-end Treasury yields rise. Speakers attributed equity support to resilient US growth, corporate earnings and AI-led capital expenditure, while fiscal deficits, heavier sovereign issuance and higher term premia sustain upward pressure on yields. The report argues that the speed and volatility of a rates move matter more than its level: an orderly 100–200bp increase may be absorbed, whereas a rapid rise could pressure risk assets and prompt capex reassessments. Speakers placed the equity market’s discomfort zone for the 10-year Treasury around 5.5%–6.0%, versus a prior feared threshold near 5%, while noting that historically strong earnings growth can support equities at higher yields. JPMorgan equity strategists cited an S&P 500 year-end target of 8,000 and strong 3Q26 earnings expectations. The bond sell-off is characterized as global and structural rather than principally an AI or US inflation story. Higher yields across Europe, persistent fiscal deficits and competition for capital support that reading. The report cites an estimated US term premium of roughly 125bp, compared with negative territory a decade earlier, alongside approximately $40tn of gross US debt. It argues that fiscal consolidation is likely needed for a durable reduction in term premia, while duration supply and future liquidity-rule changes remain key tests for sovereign markets. At the same time, speakers considered real yields attractive: a roughly 6% income yield against around 2.5% inflation was described as achievable, with an estimated 26% of securities yielding more than 6%. Inflation is expected to remain above target but not become destabilizing. The report sees a broad developed-market hiking cycle beginning after the September FOMC meeting, with eight of nine tracked developed-market central banks projected to hike by year-end and the three largest each hiking twice. A balanced Taylor-rule assessment points to developed-market policy rates roughly 100bp higher, while market curves similarly price around 100bp of tightening at about 25bp per quarter. Yet speakers argued that labor is not the underlying inflation source: unit-cost inflation averaged 2% over four years and 1.5% over the past year; wage inflation is easing and productivity is improving. AI is nevertheless considered inflationary in the short to medium term through wealth effects, data-center investment and demand for chips and software. AI capex is presented as unusually large, durable and relatively insulated from higher rates. Guidance for the five largest US hyperscalers exceeds $750bn in 2026, is expected to exceed $1.1tn in 2027, and total AI capex is cited at $5.5tn through 2030. Consensus estimates cited by the report put AI capex near $900bn by year-end and above $1.2tn by the end of next year, with hyperscalers accounting for about 87% of total spending in 2026 and 2027. Financing is broadening through investment-grade bonds, private credit, infrastructure equity, structured finance and bank lending; the report estimates that high-grade corporate markets could provide more than $2.1tn of data-center financing over five years, with leveraged-finance markets contributing about $350bn cumulatively through 2030. The constructive case depends on revenue monetization ultimately justifying high capex and valuations. The bear case is that value capture lags spending as open-source and non-US model competition erodes pricing power. The report stresses that AI’s binding constraint is increasingly execution rather than demand. Power availability, transmission, permitting, labor, chip supply and local opposition can delay or cancel projects. Grid planning designed for roughly 100bp annual load growth is confronting around 300bp growth; transmission is the tighter current constraint, and one roughly 1,000-mile interstate high-voltage line took 18 years to permit. Large campuses require multi-year planning, with lead times of three to 10 years. Data-center underwriting also requires selectivity: 98% of new capacity by megawatts is single-tenant, making tenant credit and construction quality critical. Explicit warning signs include termination-for-convenience clauses, unprotected tariff or commodity exposure, non-investment-grade tenants and speculative locations. The report notes that two buildings may require about 40,000 tons of copper and 100,000 tons of steel, making unlocked input costs consequential. Productivity is a potential longer-run gain but remains uncertain in timing and distribution. The report cites an estimate that continued hyperscaler capex could add about 0.5% to labor-productivity growth over the next one to two years; some estimates place the AI-capex contribution to US GDP growth at roughly 60bp to 1 percentage point. Speakers expect visible AI productivity effects within 12–24 months and suggested US productivity could approach 3% by 2029–30. But they also note softer labor demand, especially among early-career workers and outside AI-related sectors, and emphasize that broad macro forces remain the main explanation for the US job stall. Enterprise adoption, workflow integration, revenue per employee and AI-related earnings evidence are identified as more meaningful tests than technology capability alone. AI creates countervailing safety, cyber, fraud and political risks. For regulated institutions, especially banks, cybersecurity spending may be defensive rather than revenue-generating, even as AI development tools are estimated to deliver 30%–50% productivity gains. Local communities are increasingly challenging data-center projects over electricity, pollution and affordability. The report expects regulatory scrutiny to intensify even if a comprehensive federal AI bill is unlikely near term. It also frames national security and supply-chain resilience as increasingly important alongside cost in investment decisions, especially across frontier technology, defense, energy, industrial capacity and supply chains. Geopolitical risk remains an important inflation and commodity channel. The report sees the Strait of Hormuz as a critical chokepoint and argues that oil-flow normalization would not necessarily mean geopolitical normalization. In a prolonged conflict scenario, JPMorgan commodity strategy estimates Brent could average $87 in 2027, versus $64 in a baseline scenario in which the world enters 2027 at peace. Speakers expect geopolitical risk premia to remain elevated through at least 2027, while strategic inventories and diversified energy supply become more important. Conference survey results reinforce the broad stance: 43% expected the S&P 500 to produce the strongest return through year-end, versus 24% for gold and 22% for oil; geopolitics and inflation were tied as the largest year-end macro risks at about 35% each, while fiscal stress or a term-premium shock drew 25%. The report closes with a selective cross-asset message. Equity dispersion had reached 25 points and was normalizing as AI winners emerge and higher yields compete for capital; recurring momentum crashes remain possible. The dollar was viewed as broadly range-bound despite uncertainty around its relationship with higher US yields. Emerging-market inflows have returned, including an estimated $75bn into EM equities, but would be vulnerable to a faster US rates backup or stronger dollar. No strong conviction view was offered on cryptocurrency.

Analysis framework

The report synthesizes views from 15 macro and market speakers at JPMorgan’s September 10 Global Macro Conference and a live survey of roughly 110 participants. It connects macro variables—growth, inflation, central-bank policy, fiscal issuance and term premia—to equities, bonds, currencies, commodities and AI infrastructure, using historical comparisons, survey results, financing estimates and scenario analysis.

Methodology notes

• MacroeconomicsTaylor rule

  Balanced Taylor-rule assessment of developed-market policy rates

  The report uses a Taylor-rule reference point to argue that developed-market policy rates could be roughly 100bp higher, helping frame the expected hiking cycle.

• Fixed Income and CreditSpread analysis

  Term-premium and long-end yield analysis

  The report explains higher long-end yields through fiscal deficits, debt issuance and a rising term premium, rather than only expected policy rates.

• Industry AnalysisUpstream-Midstream-Downstream Transmission

  AI infrastructure supply-chain and financing analysis

  The report traces AI investment from hyperscaler spending to chips, data centers, power, transmission, construction inputs and multiple funding channels.

Key data

• S&P 500 year-end target8,000JPMorgan equity strategists’ cited year-end level.
• Survey: strongest total return through year-endS&P 500 43%; gold 24%; oil 22%Based on roughly 110 Global Macro Conference survey responses.
• Survey: 10-year Treasury yield expectation~60% above 4.75%; 19% below 4.25%Year-end 2026 expectations.
• Estimated term premium~125bpUp from negative territory a decade ago, according to the report’s cited dealer-survey cross-check.
• Five largest US hyperscalers’ 2026 capex guidance>$750bnExpected to exceed $1.1tn in 2027.
• Total AI capex through 2030$5.5tnReport estimate.
• Data-center new-build concentration98% single-tenant by megawattsHighlights tenant-credit and execution risk.
• Brent 2027 scenario$87 versus $64 baselineJPMorgan commodity-strategy estimate for a prolonged conflict versus peace entering 2027.

Impact & implications

The report’s cross-asset implication is that higher long-end yields need not immediately end the equity advance, provided earnings, AI monetization and the rates adjustment remain orderly. It favors attention to attractive real-income opportunities while urging selectivity in equities, AI infrastructure and emerging markets because funding, execution, policy and geopolitical risks can quickly alter the outlook.

Risks

• A rapid rather than orderly increase in long-end yields could pressure risk assets and force AI-capex reassessments.
• AI monetization may fail to justify elevated investment and valuations, particularly if open-source or non-US competition weakens pricing power.
• Permitting, transmission, power, labor, chip availability, local opposition and poor construction execution can delay or impair AI infrastructure projects.
• Cybersecurity and fraud costs may rise as AI strengthens offensive capabilities, with much of the required spending defensive rather than revenue-generating.
• Persistent geopolitical conflict, especially around the Strait of Hormuz, could keep oil prices, inflation and risk premia elevated.
• Post-midterm fiscal, tariff, energy and geopolitical actions could increase Treasury-market and policy uncertainty.
• A stronger dollar or faster US rates backup could threaten returning emerging-market flows and carry trades.

What to watch

• The pace of the next move higher in long-end yields, Treasury issuance and forthcoming liquidity-rule proposals.
• 3Q26 earnings, AI capex guidance and evidence that AI revenue monetization is closing the valuation gap.
• Data-center permitting, transmission capacity, power availability, project delays, termination notices and commodity-cost overruns.
• Labor-market breadth, wage pressure, productivity data and the upcoming BEA core-PCE calculation revision.
• Central-bank communications and the scale and pace of the expected developed-market hiking cycle.
• Iran and Strait of Hormuz developments, oil flows and signs of sustained energy-price pressure.
• US midterm and post-midterm policy developments, including tariffs, fiscal packages and AI regulation.
• The US dollar’s response to Fed policy, September Japanese flow data and emerging-market capital flows.