How sustainable is the AI rally? Jefferies weighs in By Investing.com

Jefferies says AI has driven more than 80% of the S&P 500’s gains in 2026, with the basket’s forward EPS estimates up over 30% since mid-2025 and consensus CAGR at 38.5% for 2026-27 versus 11.9% for non-AI sectors. The basket still trades at about 25x forward earnings and a 0.6x PEG, which the firm argues supports the rally, while first-quarter earnings beats were a post-COVID record at 86% and sales beats reached 82%. The article also flags U.S.-Iran tensions as a risk, with 44% of companies citing it as a negative factor due to oil/inflation and supply-chain concerns.

Analysis

The market is paying up for a very narrow earnings compounding story, but the bigger risk is not valuation in isolation — it is factor crowding. When one theme explains most of index returns and most positive surprises are being concentrated in the same names, passive and benchmark-aware capital mechanically chases the winners, which can sustain the tape longer than fundamentals alone would suggest. That said, the breadth data imply the rest of the market is becoming a low-beta, low-upside funding source, so any AI wobble would likely trigger a fast de-grossing rather than a gentle rotation.

Within the AI complex, the second-order winner is not necessarily the highest-quality platform name; it is the parts of the stack with the best near-term earnings revision torque and the least disputed valuation. That argues for staying closer to the picks-and-shovels beneficiaries with visible order books and operating leverage, while being selective on the expensive design/exposure names where expectations are already embedded. If geopolitical friction lifts energy and inflation expectations together, that is a subtle headwind for long-duration growth multiples because it raises the discount-rate debate at the exact moment the market is most concentrated.

The contrarian read is that the market may be underestimating how long earnings revisions can stay positive even if sentiment worsens. Management teams are still guiding conservatively, so beats are more likely to keep translating into estimate raises over the next 1-2 quarters, which can preserve leadership. The real tell is whether revision momentum broadens beyond AI and commodities; if it does not, index-level returns remain fragile and easily reversible on any earnings miss or macro shock. If it does, the rally can broaden without a major multiple reset, but that is the lower-probability path given current concentration.

Key Decisions for Investors

• Stay tactically long SMCI on pullbacks over the next 2-6 weeks; use it as a high-beta beneficiary of AI capex persistence, but size modestly because the market will punish any sign of order digestion. Risk/reward is favorable if revisions keep accelerating, but stop out if the stock loses relative strength versus the AI basket.
• Hold APP as a selective long for 1-3 months, but pair it against a more expensive AI software or semiconductor design name to neutralize market beta. The trade works if the market keeps rewarding earnings surprise and cash-flow conversion rather than narrative-only growth.
• Put on a relative-value pair: long SMCI / short a higher-multiple chip-design or capex proxy over the next quarter. The rationale is that the market is still paying more for perceived quality than for actual revision momentum, and that gap typically closes when breadth is narrow and expectations are elevated.
• Add an inflation/geopolitics hedge via short-duration calls on energy or a tactical long in XLE for the next 1-2 months. If oil keeps feeding inflation fears, duration-sensitive AI multiples are the first place the market will reprice.
• Avoid adding broad index exposure at current levels; if the AI basket loses leadership, the S&P can gap lower quickly because the rest of the index lacks earnings support. Prefer raising cash from low-conviction cyclicals and redeploying only after revision breadth improves.