Between reporting periods, analysts cut estimates. That's what they do. The 5-year average first-month revision is -0.9%. The 10-year average is -1.4%. In April, Q2 2026 estimates rose +2.1% — the largest first-month increase in five years.
But that's the short version. The long version is stranger.
S&P 500 Q2 2026 earnings growth estimate — drifting UP, not down
That's 6.4 percentage points of upward drift in four months. In a normal cycle, estimates would have fallen by 3-5pp over the same period. The gap between what's happening and what normally happens is roughly 10 percentage points.
The Engine: Q1 Didn't Just Beat — It Tripled the Beat
With 63% of the S&P 500 reported, Q1 is producing numbers that break the comparison framework:
| Metric | Q1 2026 | 5-Year Avg | Multiple |
|---|---|---|---|
| EPS beat rate | 84% | 78% | — |
| Avg EPS beat magnitude | 20.7% | 7.3% | 2.8x |
| Revenue beat rate | 81% | 70% | — |
| Blended EPS growth | 27.1% | — | Highest since Q4 2021 |
| Blended revenue growth | 11.1% | — | — |
The average company isn't beating by a penny. It's beating by 20.7%. The last time the beat rate hit 84% was Q2 2021, the post-COVID reopening quarter. This is that magnitude of surprise — without the reopening excuse.
The Concentration Problem
Three companies — Alphabet, Amazon, and Meta — account for 71% of the dollar-level earnings increase this season. Strip them out, and the S&P 500 is still growing, but the blended rate drops from 27.1% closer to mid-teens.
This matters for the drift. When analysts revise Q2 estimates upward, they're largely revising the same three names (plus Microsoft, AMD, ARM). The AI capex chain is running estimates in one direction. Everything else is drifting laterally.
The Divergence: Two Sectors in Decline
| Sector | Q1 EPS Growth | Q2 Est. Revision (Apr) | Net Margin vs 5yr Avg |
|---|---|---|---|
| Comm Services | +53.2% | Up | Above avg |
| Info Technology | +50.0% | Up | 29.1% vs 25.4% |
| Consumer Disc | +39.0% | Up | — |
| Energy | Decline | +45.1% | 7.4% vs 9.6% |
| Health Care | Decline | Mixed | — |
| Industrials | Growth | -2.9% | — |
The strangest line in this table: Energy had the largest upward Q2 revision of any sector (+45.1%) — while reporting year-over-year declining earnings in Q1. Analysts saw $110+ oil in April and extrapolated Q2 derivative unwinds. XOM and CVX both showed billions in timing mismatches that were expected to reverse.
Then oil dropped from $113 to $100 in two days on Iran deal hopes. Today, US and Iranian forces exchanged fire in the Strait of Hormuz — the most serious confrontation since the ceasefire began. Brent bounced to $103. Trump says the ceasefire still holds.
That +45.1% energy revision is now sitting on top of a binary outcome.
What the Drift Is Actually Telling You
Decompose the 6.4pp upward drift into its components:
1. AI capex chain flowing through (~3-4pp). MSFT Azure +40%, AMD data center +57%, ARM licensing +29%, GCP +63%. These Q1 beats are being mechanically extrapolated into Q2. PLTR US commercial +133% confirms enterprise AI spend is accelerating. The chain hasn't broken yet — 8 of 8 links beat or raised.
2. Energy timing unwinds (~1.5-2pp). Q1 derivative losses at XOM/CVX were timing, not permanent. Analysts expect reversal in Q2. But this assumes physical deliveries resume — which assumes Hormuz reopens.
3. Margin expansion in tech (~1pp). IT net profit margin hit 29.1% vs 25.4% a year ago. AI revenue is scaling faster than AI costs. MSFT's capex efficiency improved from ~$10 per $1 AI revenue to ~$3.40. More revenue per dollar spent = higher forward estimates.
What Breaks It
The S&P 500 closed at 7,399 today — another record — trading at 20.9x forward earnings. That multiple is pricing in the 21.3% growth. If the drift reverses:
Hormuz collapse scenario. If today's fire exchange escalates and the ceasefire breaks, oil returns to $110+. But that doesn't help energy estimates — it means more derivative timing delays, not more realized revenue. Meanwhile, airlines re-slash guidance, consumer discretionary weakens (MCD already showed comps halving from 7% to 3.9%), and Industrials estimates (-2.9% already) drop further. Net effect: drift stalls or reverses 2-3pp.
AI demand question. OpenAI missed internal revenue targets in Q1. Anthropic passed them at $30B vs $24B ARR. MSFT's RPO is 45% OpenAI ($281B) — and OpenAI can now serve on AWS/GCP. If the largest single customer of the largest AI infrastructure buildout is struggling to monetize, the extrapolation math changes. This wouldn't show up until Q3 estimates, but it's the crack in the supply-demand logic.
Concentration risk. Three companies = 71% of the beat. If even one disappoints in Q2 (GOOGL, META, and AMZN all report late July), the mechanical extrapolation reverses faster than it arrived.
Verdict
The 6.4pp upward drift is real, earned, and concentrated. It's not analyst optimism — it's analysts catching up to AI revenue that materialized faster than any prior technology cycle. The beat magnitude (20.7%, nearly 3x normal) is forcing revisions upward because the gap between estimates and actuals was historically wide.
But the drift has three fragilities: geographic (Hormuz controls the energy revision), customer (OpenAI's health controls the AI extrapolation), and structural (three names control 71% of the beat). A record-high index at 20.9x forward on 21.3% growth is not expensive if the growth arrives. The market is making a bet that none of these three fragilities materialize simultaneously.
Today, US and Iranian forces traded fire in the Strait of Hormuz while the S&P 500 hit a new all-time high. The market has decided that AI earnings growth outweighs geopolitical risk. The drift says they're right — so far.
S&P 500 Q1 2026 (63% reported): Blended EPS growth 27.1%. Beat rate 84%. Avg beat magnitude 20.7% (5yr avg 7.3%). Revenue growth 11.1%. Q2 EPS growth estimate 21.3% (up from 14.9% in January). Bottom-up Q2 EPS estimate $80.47 (up 2.1% in April — largest first-month increase since Q2 2021). Forward P/E 20.9x. S&P 500 close: 7,399 (record). Brent crude ~$103. Sources: FactSet Earnings Insight, S&P Global.