Key Takeaways

• In 2026, both legs of the 60/40 have taken turns disappointing: bonds through September, then stocks in early November.
• Through September 30, a moderate-risk portfolio’s real assets returned about 4% while its bonds returned about -2%.
• Real assets are a diversified sleeve, not a hedge against everything.

The 60/40 portfolio rests on a simple idea: stocks drive growth, and bonds cushion the fall when stocks stumble. Through September 30, bonds did not hold up their end. The ICE BofA US Broad Market Index was down 2.69% for the year⁽ⁱ⁾. In a moderate-risk multi-asset portfolio, fixed income averaged about 30% of assets and returned roughly -2%, while the equity sleeve, about 58% of assets, returned 16%. Real assets, about 10% of the portfolio, returned about 4%.

By early November, there was a risk reversal, where the positions that had worked best were suddenly not, even as rates pulled back from their highs, with the 10-year Treasury yield easing from about 5.3% to below 5%.⁽ⁱⁱ⁾

The lesson is not that either leg is broken. With yields reset, bonds pay you to wait again. The lesson is that a portfolio with two return engines is exposed whenever either one stumbles. A third engine, tied to different economic forces, makes the whole portfolio more resilient. That is where we think real assets belong, and the case is structural rather than a one-year story.

The supply side: restraint in oil

Energy producers have been prioritizing balance sheets and shareholder returns over growth, buying back shares and paying dividends while spending less on expansion. Global oil and gas capex is projected to fall well below its prior peak by the end of the decade When supply is slow to respond, demand surprises tend to show up in prices.

Source: Goldman Sachs. Data as of December 2025. For illustrative purposes only. Not a projection of future results. Not intended as an offer to buy or sell any securities mentioned herein or as any call to action. Past performance is not indicative of future results.

The demand side: the physical cost of AI

Artificial intelligence is often discussed as software, but it runs on power, grid capacity, materials, and buildings. VanEck analysis estimates AI compute could reach about 62 gigawatts by 2030, roughly five times today’s fleet (see Chart below). Capex from the largest AI spenders now dwarfs that of the global oil and gas industry. Someone has to supply the electricity, the copper, and the construction. Those are real assets.

Source: VanEck analysis; DC Market Insights (12.4 GW AI capacity 2025, 62.4 GW 2030 base case, interpolated); Yield Theory (big-four capex). Today’s ~17 GW split 30% inference / 70% training; training and other inference held flat. Knowledge-worker inference = seats plus agents on the indicative path. Capex at $45bn per GW new build, $30bn re-chipped, spent a year ahead. Indicative only. For illustrative purposes only. Not a projection of future results. Past performance is not indicative of future results.

Gold plays a different role

Gold is driven less by the buildout and more by confidence in currencies and debt. Central banks are a key buyer. Third-quarter central bank buying was the largest since 2024. Net monthly central bank buying has stayed near record levels, with 89% of surveyed central bankers expecting global reserves to rise over the next 12 months.⁽ⁱⁱⁱ⁾ Gold hit about $5,600 intraday in January and ended June near $4,000, roughly a 25% pullback from its highs. That volatility is real. But the buyers setting the long-term price appear to be treating gold as a strategic holding, not a trade.

What real assets are not

They are not a hedge against everything. The AI-linked parts of the complex can sell off with sentiment. Nuclear energy equities were down almost 30% for the year as of early November. Gold and commodities pull back too. That is why we treat real assets as a diversified sleeve with several different drivers, rather than a single bet.

The Takeaway

Ask whether your portfolio depends on two engines. If it does, consider a meaningful but bounded allocation to assets tied to physical scarcity, diversified across gold, energy, and infrastructure, and sized so no single holding can sink the year. Rebalance with discipline, trimming what has run and adding to what has lagged.