Goldman Sachs: The Postmodern Cycle Has Begun—AI and Geopolitics Drive a Capital Expenditure Supercycle

Overview

This report introduces the concept of the ‘Postmodern Cycle,’ arguing that the global economy and investment landscape are undergoing a structural transformation since the pandemic. This cycle marks the end of decades-long trends—disinflation, deregulation, declining interest rates, and globalization—that have dominated markets, giving way to rising macroeconomic volatility, higher real interest rates, greater state intervention, and supply-chain regionalization. Against this backdrop, a capital-expenditure (Capex) supercycle driven by the artificial-intelligence (AI) revolution and geopolitical tensions is taking shape, profoundly reshaping equity-market return dynamics, making earnings growth rather than valuation expansion the primary driver.

Core views

Core View One: A fundamental reversal in the macro environment. Unlike the ‘Modern Cycle’ of the 1980s to 2007—characterized by low inflation, low volatility, and globalization dividends—the ‘Postmodern Cycle’ is accompanied by higher capital costs, rising government debt, and increasing trade barriers. U.S. tariff levels have climbed to their highest since the 1930s, and the policy uncertainty index remains at multi-year highs, intensifying global capital competition and pushing long-term bond yields significantly higher. Core View Two: The dual engines of the capital-expenditure supercycle. On the private side, the emergence of large language models has sparked an unprecedented infrastructure-investment boom among tech giants. In Q1 2026, S&P 500 constituents’ capital expenditures grew 38% year over year, far outpacing buyback activity; the consensus 2026 capital-spending forecast for the top five hyperscalers has been raised by roughly $80 billion to $755 billion, up 80% from a year earlier. On the public side, driven by energy security and geopolitical considerations, governments worldwide are sharply increasing defense and critical-infrastructure spending, as evidenced by surging defense orders in countries like Germany. Core View Three: Severe intra-sector differentiation within the tech space and a reevaluation of the ‘old economy.’ Light-asset software companies, once favored during the zero-interest-rate era, are seeing their valuations contract sharply as agentic AI threatens to disrupt their business models and downward revisions to terminal-value assumptions weigh on their multiples; the global P/E ratio for the software sector has fallen by about five points in the short term. By contrast, semiconductor hardware, benefiting from compute demand, is performing strongly. Meanwhile, for the first time, the fates of the virtual and physical worlds are closely intertwined, with spillover effects from tech giants’ capital spending reaching data centers, power suppliers, and industrial equipment—traditional heavy-asset industries—unlocking structural growth opportunities. Core View Four: Broadening sources of market returns. The highly polarized market landscape dominated by U.S. equities, tech stocks, and growth names over the past fifteen years is changing. Market dispersion is rising, median stock returns are moderating, but companies aligned with the capital-expenditure theme—including industrials, commodities, tech hardware, and utilities—are posting robust earnings momentum, with consensus EPS estimates revised up by about 25% year over year, offering investors more avenues to capture alpha through diversification.

Analysis framework

The firm employed a long-cycle historical-comparison methodology, contrasting the pre-1980 ‘Traditional Cycle,’ the 1980–2007 ‘Modern Cycle,’ and the post-pandemic ‘Postmodern Cycle’ to identify the current cyclical phase by analyzing shifts in macro variables (inflation, interest rates, volatility), policy orientations (globalization vs. regionalization), and industrial structures (light assets vs. heavy assets). Specifically, the report deployed ‘value-volume decomposition’ and ‘supply-chain transmission’ logic: first identifying AI technology and geopolitics as the two primary drivers of capital spending; second, tracing the flow of capital expenditures along the value chain—from upstream semiconductors and hardware, through midstream data-center construction, to downstream power supply and industrial equipment; finally, by comparing the relative performance of the ‘capital-spending beneficiary’ portfolio against global indices and analyzing the convergence of valuation premiums between software and hardware segments, validating the shift in market style from ‘long-duration growth’ to ‘real assets and earnings certainty.’

Methodology notes

• Cycle and Business-Cycle FrameworkKondratiev Long Wave

  Long-wave cycles driven by technological revolutions

  The report implicitly draws on long-wave theory, positing that the AI revolution, like major historical innovations (e.g., the internet, electrification), triggers decades-long infrastructure rebuilding and capital-spending booms, thereby shaping long-term asset-return patterns.

• Industry/Industrial Analysis FrameworkSupply-demand framework

  Capital intensity and supply constraints

  The report examines the supply-side shift from ‘light assets’ to ‘heavy assets.’ With constrained capacity in electricity, computing power, and defense, capital-intensive firms gain pricing power and valuation premiums thanks to their scarce physical assets—a classic supply-demand mismatch analysis.

• Valuation MethodologyDCF Discounted Cash Flow

  Rising discount rates suppressing long-duration assets

  The report notes that as real interest rates and capital costs rise, the present value of future cash flows declines, putting valuation pressure on ‘long-duration’ assets reliant on distant growth prospects (such as software stocks) and explaining the market’s pivot toward assets with near-term earnings visibility.

• Event Game Theory and Behavioral FinanceExpectation Gap/Expectations Management

  Market repricing of AI business models

  The report analyzes the evolution of investor expectations: from initial blind enthusiasm for all things AI to questioning the capital returns of hyperscale cloud providers and concerns about AI’s impact on software business models—an expectation gap that has triggered sharp rotation across sectors.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• META.US (Meta Platforms)

  As one of the leading hyperscale cloud providers, META is both a primary driver and beneficiary of AI-related capital spending, though it also faces market scrutiny over the sustainability of its capital-return profile.

  Strengths

  Pioneering investments in AI infrastructure, with robust data and application ecosystems.

  Weaknesses

  Substantial capital expenditures may erode free cash flow, and there are lingering concerns about the long-term viability of its investment returns.

  Comparison

  More hardware-focused than pure-software companies; positioned as the payer rather than the direct recipient of capital-spending flows compared to semiconductor manufacturers.

  Risks

  AI monetization falling short of expectations, leading to valuation pullbacks; regulatory risks.

• Semiconductor Industry (Semiconductors)

  Directly benefiting from surging demand for AI compute power, serving as a core upstream supplier in the capital-spending chain.

  Strengths

  Seemingly limitless demand, highly predictable earnings growth, and relative outperformance compared to hyperscale cloud providers.

  Weaknesses

  Valuations already reflect elevated growth expectations, with cyclical volatility risks.

  Comparison

  Within the tech sector, significantly outperforming software and services segments.

  Risks

  Overcapacity risks; supply-chain disruptions stemming from geopolitical tensions.

• Software and IT Services Sector (Software & IT Services)

  The disadvantaged or lagging segment, confronting both the threat of AI-driven business-model disruption and a restructuring of its valuation framework.

  Strengths

  Some firms may leverage AI integration to enhance efficiency.

  Weaknesses

  High valuations are hard to sustain; lowered terminal-value assumptions leave the sector vulnerable to a ‘Kodak-style’ disruption.

  Comparison

  Significantly underperforming hardware and semiconductor segments.

  Risks

  Agentic AI replacing traditional software-licensing models; customer-spend cuts.

• Traditional Heavy-Asset Industries (Industrials, Utilities, Energy)

  Indirect beneficiaries, absorbing spillover effects from tech giants’ capital spending and benefiting from increased defense outlays.

  Strengths

  Relatively low valuations, improving earnings growth amid tight supply-and-demand conditions, backed by tangible assets.

  Weaknesses

  Sensitive to macroeconomic cycles, with longer capital-return horizons.

  Comparison

  Transforming from the overlooked ‘old economy’ into a sector with structural growth opportunities.

  Risks

  Energy-price volatility; execution of policies falling short of expectations.

Key data

• Top Five Cloud Providers’ 2026 Capital-Spending ForecastApproximately $755 billionUp roughly $80 billion from last year’s consensus estimate, representing an 80% year-over-year increase
• S&P 500 Companies’ Q1 2026 Capital-Expenditure Growth Rate+38%Far outpacing the 1% growth in share buybacks over the same period, signaling strong investment appetite
• Software Sector’s Global P/E Ratio ChangeDown approximately five pointsA significant valuation contraction in a short span, driven by rising risk premiums and lowered terminal-value assumptions
• Capital-Spending Beneficiary Portfolio’s Year-to-Date GainsAbout 25%Outperforming broader markets, with consensus EPS estimates revised up roughly 25% year over year
• U.S. Public Debt as a Share of GDPRising from 55% to 124%Reflecting increased government borrowing, which has pushed long-term interest rates and capital costs higher

Impact & implications

From an investment-strategy perspective, the report concludes that the era of profiting solely from index beta is over, making stock-picking ability (alpha) paramount. Investors should overweight sectors directly benefiting from accelerating capital spending, including AI hardware infrastructure, electricity and energy security, defense industries, and high-asset-density traditional sectors. For tech stocks, caution is advised regarding ongoing valuation downgrades in the software segment, with a shift toward hardware components such as semiconductors. Moreover, as global trade fragments and regionalizes, emerging markets, Japan, and certain European countries with heavy-asset economies may offer diversification benefits.

Risks

• Escalating geopolitical conflicts could further disrupt supply chains or cause cost overruns
• Slow commercialization of AI technologies might prevent massive capital expenditures from translating into expected profits
• Unsustainable government debt could trigger sovereign-credit crises or hyperinflation
• Rapid AI-driven disruption of software business models could precipitate systemic valuation collapses

What to watch

• Hyperscale cloud providers’ capital-spending guidance and actual implementation
• Relative valuation-premium changes between semiconductors and software
• Execution progress of defense budgets and order-release rhythms in major global economies
• Trends in long-term real interest rates and their impact on growth-stock valuations