AI Drives 80% of S&P 500 Gains, Valuation Remains Attractive

Overview

This report focuses on structural opportunities in the US stock market under the AI wave. The report points out that although Q1 saw record-breaking earnings growth (86% of companies reported earnings growth), market gains were highly concentrated in AI-related areas. The AI sector contributed over 80% of the S&P 500 index's gain this year, accounting for 38.7% of index weight. The core conclusion is: AI theme fundamentals are strong, valuation remains attractive, momentum comes from earnings rather than multiple re-rating, which is sustainable. The report suggests continuing to focus on high ROIC enterprises and profit-momentum GARP (Growth at a Reasonable Price) strategies.

Core views

AI returns fully driven by earnings. AI-related stocks account for 38.7% of S&P 500 weight, contributing the vast majority of YTD gains, with AI servers, optics, and storage sub-sectors performing best, while supercomputing centers and chip designers lagged relatively. Decomposing returns into "E" (earnings) and "P" (valuation multiples), the AI basket's rise was mainly driven by earnings, not multiple re-rating, indicating a solid and sustainable trend foundation. Without AI, S&P 500 was up only 2% this year, sufficiently illustrating extreme market differentiation. AI valuations remain the cheapest after pullback. Although the AI basket rose 30% this year, it still trades at 25x forward PE, below +1 standard deviation, PE/G 0.6x — lowest among S&P 500 sectors. By comparison, Non-AI S&P 500 PE/G as high as 1.6x. From sub-sector view, storage and computing stocks most attractive on PEG basis, while chip design and semiconductor capex relatively expensive. AI earnings upgrades strong and reliable. AI sector 2026 expected EPS higher than mid-2025 by over 30%, raised 13% in last three months, consensus expects 2026-27 EPS CAGR reach 38.5%, far surpassing non-AI department's 11.9%. Analyst sentiment follows upward — Q1 2026 nearly 58% companies received positive analyst ratings (up from 48% in 4Q25). Management attitude also stable optimistic, 95% enterprises kept positive remarks in earnings calls. Non-AI and commodities sector upgrades stalled. S&P 500 2026 EPS rose 6% in past three months, but excluding AI and commodities only 0.3%, highlighting huge difference between trending themes and other markets. Similarly, Non-AI and Commodities excluded S&P 500 growth rate about 11%, while overall index growth rate 19%. Other sectors except energy, materials had limited rises, consumer discretionary, industrial and consumer staples even saw downward revisions. Q1 beat expectations dismissed and punished. Although 86% of companies reported earnings growth (historic high), 82% reported sales growth (also historic high), stock price performance disappointing — except AI and few sectors, earnings growth companies' stocks did not rise due to this, instead companies missing expectations faced heavy losses. However positive signal: earnings growth accompanied rise, misses accompanied fall, implying analyst rating adjustments normal and market expectations clear, earnings risk in controlled position. Iran conflict impact limited but needs watch. Based on AlphaSense AI analysis of approx 330 companies Q1 earnings calls, 44% companies believe Iran war has negative impact, 16% beneficial, 40% limited impact. Main negative factors include specific industry shocks, supply chain disruption, increased market volatility and consumer sentiment decline. But overall, this factor caused no substantive disturbance to the market.

Analysis framework

Report adopts layered quantitative research method. First, verify source of gain through return decomposition (E vs P), confirming earnings-driven rather than valuation bubble. Second, based on PE/G valuation multiples for horizontal comparison, measure relative attractiveness of each sector. Third, judge sustainability of trend through analyst expectation raises and revision speed (by time span statistics). Finally, construct investment portfolio framework on two dimensions of ROIC and GARP: ROIC embodies "moat" and earnings stability (select LTM ROIC >20%, above WACC and with improvement momentum companies), GARP combines growth and valuation efficiency (PE/G <1.5x, 2026-27 EPS CAGR >10% and recent raise positive). This method captures mainstream theme core drivers while reducing risk concentration through quality screening.

Methodology notes

• Valuation MethodPEG Valuation

  PE/G (Price-to-Earnings / Growth) is indicator measuring match between growth and valuation, lower value indicates cheaper valuation given growth rate

  Report uses PE/G to judge AI sector valuation attractiveness. AI PE/G 0.6x, meaning although PE 25x seems not low, considering annual 38.5% EPS growth rate, its cost-performance far exceeds other parts of S&P 500 (PE/G 1.6x). This reminds investors, should not look only at absolute PE, but compare relatively with growth rate

• Company Fundamentals and Financial FrameworkROIC–WACC spread

  Companies where ROIC (Return on Invested Capital) exceeds WACC (Weighted Average Cost of Capital) are true value creators, and if ROIC continuously improves prospects are better

  ROIC stars screening criteria require LTM ROIC >20% and above WACC, these companies have durable competitive advantages (moats). Report emphasizes ROIC improvement momentum (wROIC momentum), identifying those with both moats and expanding quality enterprises, such enterprises have largest long-term return potential

• Industry/Industrial Analysis FrameworkSupply and Demand Framework

  In AI supply chain, different links (servers, chips, storage, optics, etc.) have different supply-demand patterns, leading to sub-sector performance differences

  Report divides AI basket into 11 sub-themes by supply chain position (servers/storage, chip design, compute, supercomputer centers, storage/packaging, new cloud, networking, optics/photonics, physical data center, power semiconductors, semiconductor capex, thermal management). Servers, optics, storage tight supply-demand leads highest gains; supercomputer centers and chip design lag due to increased competition or demand peaking. Such sub-sector analysis helps investors precisely allocate, rather than mechanically viewing whole AI sector

• Event Gambling and Behavioral FinanceExpectation Gap/Expectation Management

  Market expectation of listed company performance determined by analyst consensus, when actual results exceed consensus is "surprise", market reaction depends on pre-event expectation levels

  Report points out Q1 although record 86% earnings growth (absolute growth high), stock price performance flat (even punished). Shows investor expectations already very high, growth fast but not exceeding expectations. Conversely, performance missed companies hit hard, implies market expectation set already at critical level. This reminds investors cannot look only at absolute performance, key is whether exceeding market expectation

• Quantitative/Factor/Portfolio TheoryMulti-factor model

  Score enterprises by multiple dimensions (growth, valuation, quality, momentum, etc.), construct comprehensive investment portfolio to balance risk-return

  JEFQ USA portfolio integrates ROIC (Quality/Moat) and GARP with momentum (Growth/Valuation/Momentum) two dimensions, equivalent to 2D factor combination. More comprehensive than single factor, grabs both质地 stable enterprises and growth accelerating enterprises, and confirms trend through momentum, thereby outperforming S&P 500 by 14.6% from late 2022 to now

Key data

• AI Sector Contribution Share of S&P 500 Gains>80% (2026) vs 60% (Avg Since 2023)AI theme's driving force on market hits historic high
• Non-AI S&P 500 Year-To-Date Gain2%Market highly differentiated, non-AI segments barely rose
• AI Basket 2026 Gain and Current PE/G30% / 0.6xGain considerable but valuation still cheapest
• AI Basket Forward PE25x (below +1 std dev)Relative to 25x PE, if growth reaches 38.5% CAGR, then valuation reasonable
• AI Basket 2026-27 EPS CAGR38.5%Far exceeds non-AI department's 11.9%
• S&P 500 2026 EPS Three-Month Raise6% (Total) / 0.3% (Excluding AI and Commodities)Raise fully driven by AI and Commodities
• AI Basket 2026 EPS Raise (Since Mid-2025)30% (6 Months) / 13% (Past 3 Months)Strong raise but recent momentum moderate
• Q1 Earnings Growth / Sales Growth Record High86% / 82%vs 4Q25's 75% / 74%, post-pandemic high
• Management Positivity / Analyst Positive Sentiment95% / 58%vs 4Q25's 94% / 48%, both improved
• S&P 500 Index AI Related Weight38.7%Nearly 40% weight concentrated on AI theme
• Iran War Negative Impact Assessment44% Companies Negative / 40% Limited Impact / 16% PositiveOverall impact controllable
• JEFQ USA Portfolio Outperformance Since Last Update6.8% (Since Jan 2026) / 14.6% (Since Dec 2022)ROIC stars outperformed 10.9%, GARP with momentum outperformed 1.2%

Impact & implications

Report views carry multi-layer implications for market. On one hand, AI fundamentals strong and valuation relatively cheap support continuing overweight or holding AI related positions, especially sub-sectors with supply constraints like servers, storage, optics. On other hand, market severe differentiation (non-AI up only 2%) means other segment allocation value may be severely undervalued, but need wait for their rise trigger (such as geopolitical easing, interest rate expectation adjustment etc.). For individual investors, current appropriate to use AI high-growth GARP and high ROIC quality enterprises as core, but should not overly concentrate, retain flexibility for potential style switching. For institutional investors, report prompts significance of high differentiation and expectation management,提示 need go beyond simple "AI Long" narrative, turn to finer sub-sector screening and gain logic verification.

Risks

• AI related stock valuation although attractive on PE/G basis, still exists risk of 37-40% EPS growth expectation not met, once revised valuation drop risk large
• Market expectation for AI earnings growth set at high level, growth "surprise" space limited, subsequent mainly depends on whether reaching or slightly exceeding consensus
• US-Iran conflict currently assessed as limited impact, but geopolitical escalation may cause supply chain disruption and market volatility rise
• Non-AI S&P 500 segment gain limited (only 0.3%), if AI relative performance stalls, may trigger valuation reallocation and market style switch
• AI chip design and supercomputing center sub-sectors relatively lagging, need monitor increased competition or demand peak risks

What to watch

• Subsequent AI basket EPS raise dynamics, especially after Q2 raise momentum whether continuous, to judge consensus whether exceeded or stalled
• Supply/Demand situation evolution of AI servers, storage, optics sub-sectors, whether face price pressure due to capacity release
• Trigger point for non-AI segment gain, such as interest rate expectations, geopolitical easing catalysts, to judge timing of style switch
• Analyst expectation dynamics, especially direction of adjustment for non-AI segment raises, reflect market expectation changes for future growth
• Follow-up development of Iran conflict and its actual shock degree on supply chain and consumer sentiment