The financial assets of the world have never been more concentrated. That is the unsettling thesis at the center of Goldman Sachs Global Strategy Paper No. 77, published July 10, 2026, in which Christian Mueller-Glissmann, CFA, and colleagues at Goldman Sachs International lay out a careful, historically grounded argument: the "World Portfolio" has drifted so far toward US equities and the Technology sector that it now carries structural risks investors can no longer afford to ignore.
The diagnosis is precise. Three years of strong recovery following the 2022 "Balanced Bear" drawdown have left global multi-asset benchmarks backward-looking and momentum-dependent. Mueller-Glissmann states that "the composition of the 'World Portfolio' has materially drifted," and that "the current AI capex boom increases the risk that falling profitability for mega-cap Tech stocks materially drags on equity returns before benefits from AI adoption show up."
That sentence carries real weight. The hyperscalers, those platform companies whose capital-light models drove a decade of outsized earnings, are now allocating virtually all of their operating cash flows to AI infrastructure spending. Return on equity for mega-cap Tech is expected to decline by an average of 7 percentage points beginning in 2027, according to consensus estimates cited in the paper. The S&P 500 Shiller P/E has climbed back toward Tech Bubble territory. The Tech sector's weight in US equities has already surpassed its late-1990s peak.
And yet the paper is not a bear case dressed up in academic clothing. It is something more honest. Mueller-Glissmann acknowledges plainly that "equities deliver some of their strongest returns in the final years of a bull market, often led by the sector that outperformed in the preceding years." Being underinvested is a real cost. Timing the market always competes with time in the market.
The structural framing is what sets this paper apart. Innovation and inflation, the authors argue, have historically been the dominant drivers of multi-asset structural cycles. Balanced portfolios performed best during periods of low inflation and rising productivity, specifically the Golden 20s and 50s, the 1990s tech boom, and the post-GFC era. The risk today is that the World Portfolio has been shaped by the last era's winners, not calibrated for the next.
Higher inflation volatility compounds that risk. Mueller-Glissmann points to three converging structural headwinds: deglobalisation, decarbonisation, and demographics. Each of these creates upward pressure on inflation over time. With bond yields near their long-run average rather than below it, bonds offer less of a buffer than they did in the pre-2022 period. The paper notes that "in all but an optimistic 'Goldilocks + AI boom' scenario, expected long-term equity returns are below the long-run average."
The prescription is neither to sell equities nor to stay passive. It is to rebalance with discipline across five identified strategies. Selective real asset exposure, including commodities, infrastructure, and real estate stocks, adds inflation protection and has historically cushioned Tech-led drawdowns. Style diversification within equities, specifically toward low-volatility and high-dividend-yield names, helped investors through every major TMT correction since 1950. Regional diversification, away from the current 63% US weight in the MSCI AC World, reflects the finding that non-US equities have tended to outperform during Tech-led US sell-offs. Long-dated call options offer upside convexity without the full drawdown risk of direct equity exposure. Finally, selective alternatives including commodity carry strategies, CTAs, and equity long/short funds can provide meaningful diversification when traditional assets are moving together.
The regime-neutral portfolio analysis is particularly compelling. Regularly rebalanced portfolios built around historical average weights have outperformed the World Portfolio on a risk-adjusted basis since 1950, a finding that holds across multiple construction methodologies including equal-weighting and risk parity. They currently point to meaningfully lower equity weights, lower Tech exposure, and higher real asset allocations than any market-cap benchmark carries today.
Five Key Takeaways for Advisors and Investors
- The World Portfolio is not a neutral starting point. It is a product of the last decade's winners and is currently overweight US equities, overweight Tech, and underweight inflation protection.
- The AI capex cycle introduces a J-curve risk. Productivity benefits from AI may take years to materialise, while profitability among the hyperscalers is already showing early signs of pressure.
- Inflation volatility is a structural condition, not a temporary one. Deglobalisation, decarbonisation, and demographics all point toward a higher-volatility inflation environment that bonds alone cannot offset.
- Leaning against momentum is expensive in the short run but rewarding over the medium term. Regime-neutral portfolios have outperformed over long horizons, but underperformed during strong Tech-led bull markets.
- The five strategies, real assets, style diversification, regional diversification, options, and alternatives, are not defensive postures. They are tools for staying invested while building resilience.
Footnote:
1 Mueller-Glissmann, Christian, Andrea Ferrario, Alessandro Giglio, Giovanni Ferrannini, Elena Porfidia, and Peter Oppenheimer. "Balancing Innovation and Inflation in Portfolios." Goldman Sachs Global Strategy Paper, no. 77, Goldman Sachs International, 10 July 2026.
