Just 3 stocks are behind 75% of the S&P 500's bullish earnings revisions since the Iran war

Goldman says 75% of the S&P 500's recent positive earnings revisions came from just three stocks: Micron, Exxon, and Chevron, with Micron alone contributing 51%. The revisions were concentrated in Energy and Technology, while 8 of 11 sectors saw no increase in EPS estimates and the median S&P 500 company was unchanged for 2026. The upgrade wave has helped lift the index back to record territory amid Iran-war-driven oil price moves and AI-related strength in tech hardware.

Analysis

The market is rewarding revision breadth, but the underlying breadth is weak: this is a narrow-duration rally being carried by a few high-beta earnings levers rather than a broad upgrade to the index’s earnings power. That matters because index-level multiple expansion can persist for weeks on headline consensus upgrades even when the median company is flat; the fragility shows up when those same few names stop compounding revisions and the rest of the index has no offset. In other words, the support is real, but it is concentrated enough that a single reversal in memory pricing or crude can unwind a large share of the perceived earnings tailwind.

Micron is the cleanest expression of the current tape because the upgrade cycle is less about end-demand and more about the market re-rating the duration of AI-driven memory tightness. If that regime holds, the follow-on beneficiaries are not just adjacent chip names but equipment, packaging, and datacenter power infrastructure; if it softens, the reversal tends to be fast because consensus is still extrapolating improvement into next year. Broadcom’s contribution is smaller but important as a tell: investors are paying for AI monetization quality, not just unit growth, so any disappointment there would hit the factor basket harder than the stock alone.

Energy is more precarious. The revisions are backward-looking to a geopolitical spike, while equity prices are already sniffing out lower crude near term, implying the easy upside has likely been captured. That creates a short runway for positive estimate momentum: if oil retraces, analysts will cut numbers just as positioning is becoming crowded, which is the classic setup for a 4-8 week giveback. The second-order loser is not just upstream; refiners and service names face a mismatched setup where earnings may still print well after the stock has begun discounting normalization.

The contrarian read is that this rally is less a broad earnings improvement than a temporary transfer of earnings power from consumers and cyclicals into two capped pockets: memory and oil. That is bullish for the index in the very near term, but not necessarily for leadership durability. The market is likely underpricing the speed at which consensus can de-revise once crude stabilizes and the AI trade shifts from estimate revisions to execution risk.

Key Decisions for Investors

• Stay long MU into the next 2-4 weeks, but size it as a trade rather than a core hold; the risk/reward is strongest while revisions are still climbing, with downside if memory ASP expectations flatten and the multiple de-rates quickly.
• Use CVX/XOM strength to fade the move via a short basket of CVX, COP, EOG, and OXY over the next 1-2 months; the setup favors mean reversion if crude rolls over and analysts stop lifting forward EPS.
• Pair long MU / short a high-quality consumer or industrial proxy for 1-3 months; the thesis is that AI-led revision breadth is real while broader earnings remain stagnant, creating a relative winner from estimate dispersion.
• Buy a short-dated put spread on VLO or the broader energy complex for 4-8 weeks; refiners are vulnerable if war-related oil spikes fade faster than product crack support, compressing the upside from already elevated revisions.
• If owning AVGO, use it as a lower-beta AI expression and fund it by trimming weaker semiconductor names; AVGO offers better durability if the market shifts from pure revision momentum to monetization quality.