Barclays says equities are attractive despite full valuations

Barclays remains constructive on equities, lifting its 2026 S&P 500 EPS forecast to $337 from $321 and its price target to 7,800, while projecting 2027 EPS of $389. The bank favors AI-linked sectors such as technology, media and telecoms, industrials and utilities, and sees Japan as the best risk-adjusted AI exposure outside the U.S. It is more cautious on bonds, forecasting 10-year Treasury yields at 4.65% over the next year and recommending underweight long-duration government debt.

Analysis

The market is being asked to price a rare combination: a durable capex cycle and a softening policy backstop at the same time. That is broadly supportive for the AI supply chain and the industrial enablers of data-center buildout, but it also means leadership should narrow to firms with immediate revenue conversion rather than “AI optionality” stories. The biggest second-order effect is that earnings breadth can improve even if multiples compress, because nominal growth plus higher utilization can offset the discount-rate headwind.

The more interesting asymmetry is in rates. If long-end yields grind higher while the Fed becomes less predictable, duration-sensitive equities will likely de-rate even if the macro stays constructive. That creates a hidden tax on software, REIT-like utilities, and any AI beneficiary whose valuation depends on cash flows far out on the curve; meanwhile, banks may not deserve a reflexive overweight if private credit deterioration and AI automation pressure fee pools faster than higher rates help NII.

Outside the U.S., Japan looks like the cleaner way to express the AI trade because you get operating leverage to capex plus a shareholder-return rerating and less single-name concentration risk than Korea/Taiwan. The contrarian point is that “expensive but not wrong” usually becomes wrong only when earnings revisions roll over — so the real tell over the next 1-2 quarters is whether capex plans translate into margin expansion or just capex inflation. If the latter, this becomes a stock-picker’s market fast, with utilities and industrials likely outperforming pure megacap beta.

Key Decisions for Investors

• Go long SMH vs. short IWM for the next 3-6 months: semis have the clearest earnings torque from AI capex while small caps are more exposed to higher rates and weaker pricing power; target a 2:1 reward/risk if the trade expresses as continued earnings revision divergence.
• Add long Japan equity exposure via EWJ or a basket of exporter/automation names for a 6-12 month hold; use this as a lower-volatility AI proxy with a favorable rerating catalyst from capital returns and governance reform.
• Underweight TLT / short ZROZ into the next 1-2 FOMC meetings: if the market starts pricing a higher term premium and more inflation-sensitive policy, long duration should remain the cleanest macro short with asymmetric downside versus a modest carry cost.
• Pair long XLI vs. short XLF over 3-6 months: industrials should capture more direct AI infrastructure spend, while financials face dual pressure from private credit skepticism and AI-driven commoditization of fee income.
• Sell upside volatility in high-duration AI software names on strength, or use put spreads, if implieds remain elevated: the risk/reward favors downside hedges because the main reversal trigger is not earnings collapse but any disappointment in forward capex-to-revenue conversion.