Citi Model: Nearly 90% of Tech Stocks Set for Earnings Surprises, NVIDIA and AMD Leading the Pack
nashnova research
Citi's quant model flags 66.2% of Russell 1000 names for positive earnings surprises this quarter, with tech leading at 88.1%. Yet the highest growth forecast — energy at 118.6% — carries the lowest beat probability.
What exactly is this model predicting?
Citi's quantitative model scans Russell 1000 constituents, flagging stocks likely to beat Wall Street earnings estimates and post positive price reactions. This quarter's hit rate: 66.2%, up 5.3 percentage points from last quarter — the highest since Q4 2021.
The S&P 500 consensus for Q3 year-on-year earnings growth has been revised up from 26.7% at the start of the quarter to 29.6%, with all 11 sectors expected to grow.
This means → the bar is rising, yet Citi believes most companies can still clear it. Confidence and absolute expectations are strengthening in tandem.
Which sectors does the model favor most?
Ranked by positive-surprise probability: IT at 88.1% > Healthcare at 75.9% > Industrials at 73.8% — tech leads by a wide margin.
Citi's named tech picks cluster around AI compute infrastructure: Nvidia, AMD, Intel, Texas Instruments, Applied Materials, KLA and other large-caps.
In plain terms = Citi is not calling a broad tech rally. The thesis is that capital will concentrate in the most certain AI hardware leaders, with healthcare and industrials as the next wave of rotation.
How are the Magnificent Seven splitting?
Positive model signals: Alphabet, Microsoft, Apple, Nvidia. Negative signals: Tesla, Meta, Amazon.
Meta and Amazon still carry "Buy" fundamental ratings — This means → Citi sees no medium-term problem, but their near-term quarterly prints may disappoint.
Crowding data — a gauge of how concentrated long positioning is — stands out: Meta at 0.955, Nvidia at 0.876, well above Apple at 0.671, Microsoft at 0.684. In plain terms = the more crowded the trade, the sharper the selloff if earnings miss.
Why is healthcare more likely to beat despite slower growth?
Healthcare's earnings-growth forecast has been cut from 8.7% to 5.1%, yet its positive-surprise share rose to 75.9%, up roughly 18 percentage points from last quarter.
This reflects a counter-intuitive pattern: once expectations are lowered, actual results cross the reduced bar more easily — creating "surprises" almost mechanically.
Citi's healthcare buy-rated picks include J&J, Abbott, Intuitive Surgical, Danaher, and Thermo Fisher. Several carry notably low crowding scores — Medtronic at 0.170, Intuitive Surgical at 0.266 — far below popular compute names.
Energy has the highest growth forecast — why the lowest beat odds?
Energy's earnings-growth estimate surged from 79.3% to 118.6%, the highest of any sector. Yet its positive-surprise share is just 48.1%, well below healthcare's 75.9%.
In plain terms = expectations are already so elevated that actual results struggle to clear an extreme bar. This means → high absolute growth ≠ high beat probability. The real opportunity in earnings season lies in the gap between results and prior expectations, not in the absolute growth number.
This reflects the model's core logic: beat probability = f(expectation gap), not f(absolute growth).
Does JPMorgan's view point to the same test?
JPMorgan notes that semiconductor forward-12-month EPS estimates have been revised up 30%–40% since June. It projects hyperscaler capex at roughly $950 billion in 2026, $1.4 trillion in 2027, and at least $3 trillion by 2030.
Citi and JPMorgan converge on one verification point: can AI compute leaders convert strong order books into above-consensus earnings this quarter?
This means → if they deliver, current elevated valuations gain fundamental support. If they fall short, crowded long positions face concentrated unwinds — that is the single most important question of this earnings season.
市场有风险，内容仅供研究参考，不构成投资建议。