U.S. Major Banks Kick Off Q3 Earnings Season: AI Stock Concentration and Inflation Take Center Stage
nashnova research
JPMorgan, Goldman Sachs and peers report Q3 results tomorrow. S&P 500 earnings growth is expected at 29.6%, roughly half last quarter's pace, with just ten stocks driving nearly 70% of the gain — AI concentration and consumer inflation are the season's two proving grounds.
How much weaker is Q3 versus last quarter?
S&P 500 Q3 consensus: revenue growth 12.4%, earnings growth 29.6%, margin 15.0%.
Q2 comparison: revenue growth 15.6%, earnings growth 52.3%, margin 17%. This means → earnings growth has roughly halved and margins slipped two points; the market's faith in sustained high growth is being repriced.
Full-year view: FY2026 EPS growth forecast at 32.7%, FY2027 at 15.8% — the deceleration is already baked into consensus.
Who is carrying almost all the growth?
Goldman Sachs data: the top ten contributors are expected to account for 68% of total S&P 500 earnings growth, up from 47% in Q1 and 48% in Q2 — concentration is accelerating.
The breakdown: Micron alone contributes 19% of index-level growth, Nvidia 15%, Meta 7%, Alphabet 6%, Broadcom 5%, followed by ExxonMobil, Chevron, Marathon Petroleum and Boeing.
In plain terms = the entire S&P 500 "growth story" is really fewer than ten companies talking; the other 490-plus stocks contribute roughly 30% combined.
How weak is the median stock, really?
The median S&P 500 stock is expected to grow Q3 EPS by just 9%, down from 14% in Q2. Every sector except IT and communication services is decelerating quarter-over-quarter.
Analysts have cut the median stock's Q3 margin estimate by 11 basis points; IT is the only sector not revised down.
This reflects a gap the index headline hides — the S&P 500 sits roughly 2% from its all-time high, yet the median stock is down 17% from its peak. Market breadth has fallen to its narrowest since the dot-com bubble.
What are the four things to watch this season?
Market breadth: Bloomberg data shows 57% of S&P 500 constituents are in a technical bear market (down over 20% from highs). This means → any pullback in leading tech names could trigger a broader correction with no breadth cushion underneath.
The beat-and-punish asymmetry: in four of the past six earnings seasons, stocks that beat estimates posted smaller one-day gains than the historical median, while misses were punished more severely than average — good news discounted, bad news amplified.
Capex signals: the market sees the trigger for ending the AI rally as one or more hyperscalers cutting capex. JPMorgan considers this extremely unlikely in 2026 but will watch for indirect signals such as reduced shareholder returns.
Why are inflation and consumer stocks called out separately?
JPMorgan frames inflation through the consumer sector: that sector is currently trading closer to a "consumer recession" scenario, with consumer confidence at a historic low.
In plain terms = consumer stock prices are already saying "recession is here," but the Fed's policy path is not following a recession script — there is a clear disconnect between the two.
The season's core test: whether AI infrastructure spending can keep delivering upside surprises, and whether actual consumer data can correct the market's over-pricing of a recession scenario.
市场有风险，内容仅供研究参考，不构成投资建议。