1) Scoreboard vs. the original setup
The original preview framed HPE’s Q4 FY25 as a modestly bullish, AI-driven beat:
- Expected a beat on both EPS and revenue.
- Bias toward an upward gap of roughly +6% from the low-$22s.
- Directional odds around 60% in favor of an upside move.
- Guidance expected to be broadly “okay” rather than a fresh problem.
What actually printed:
- Non-GAAP EPS came in at $0.62 vs. about $0.58 expected, a solid mid-single-digit beat.
- Revenue landed at $9.7B vs. roughly $9.9B expected, a small but clear miss.
- Near-term revenue guidance for the following quarter came in below Street, even as FY26 EPS guidance was raised.
- The stock closed at $22.90 into earnings, opened at $20.90 the next morning (about -8.7% overnight), then closed at $23.33, about +1.9% vs. the pre-print close.
On the scoreboard:
- The gap move was clearly down, not up — the core directional call on the overnight reaction was wrong.
- The full-session move ended only modestly positive versus the pre-earnings close, closer to “flat-to-slightly-up” than the +6% rally that was envisioned.
- The crowd, which was 100% “Up” on this event, was also wrong on both the gap and the next-day outcome.
2) What HPE actually reported
Fundamentally, the quarter was a classic “EPS beat, revenue miss” story, with strong profitability but a softer-than-hoped top line:
- Revenue: $9.7B, up mid-teens year-on-year but below the ~$9.9B Street expectation.
- Non-GAAP EPS: $0.62 vs. ~$0.58 consensus, a ~7% beat and above the company’s own range.
- Margins: Non-GAAP gross margin climbed into the mid-30s, and non-GAAP operating margin pushed above 12%, both materially better than a year ago.
- Cash flow: Free cash flow was strong, comfortably ahead of last year.
Segment color reinforced the idea that the growth engine is heavily concentrated:
- Servers declined mid-single digits year-on-year with some margin pressure.
- Networking (now including Juniper) more than doubled vs. last year, with healthy margins but a slight step down from prior peaks.
- Hybrid cloud shrank low-teens year-on-year with weaker profitability.
- Financial services was essentially flat on revenue but better on returns.
The original thesis leaned on AI and networking to pull both EPS and revenue modestly above consensus. EPS did its job; revenue did not — and that mix matters a lot for a name that had already gone through a guidance reset earlier in the year.
3) Guidance: raised FY26 EPS, soft near-term revenue
Guidance was the fulcrum of the reaction and explains why the gap went the “wrong” way relative to the preview:
-
For the following quarter:
- Revenue was guided to roughly $9.0–$9.4B, below consensus near $9.9B.
- Non-GAAP EPS guidance of $0.57–$0.61 was slightly above the mid-$0.50s Street number.
-
For FY26:
- Management reaffirmed high-teens revenue growth targets.
- They raised their non-GAAP EPS framework to roughly $2.25–$2.45, up from prior guidance, implying notable operating leverage.
From a multi-quarter perspective, this is constructive: margins and EPS are tracking ahead of plan. But the market is currently far more sensitive to near-term revenue visibility in AI infrastructure than to out-year earnings math:
- The raised FY26 EPS target wasn’t enough to offset disappointment in the immediate revenue outlook.
- After a prior guidance reset, investors are quick to punish anything that smells like “still not quite on track.”
That’s why the quarter is best characterized here as an EPS beat with weak near-term guidance, rather than simply “inline” or “strong.” The guidance axis — which the original preview treated as manageable — turned out to be the main source of pain.
4) Price action: big downside gap, then a sharp rescue
The tape around the print followed the typical “EPS beat, revenue/guide miss” template:
- Into earnings: The stock drifted higher into the event, closing at $22.90 on December 4.
- At the open: The next morning, shares opened at $20.90, an ~8.7% downside gap, slightly larger than the options-implied move — and pointed the opposite direction of the original call.
- Intraday: Buyers stepped in aggressively as the call progressed and the raised FY26 EPS story came through more clearly. The stock traded up into the mid-$23s.
- At the close: It finished at $23.33, about 1.9% above the pre-earnings close.
So we got:
- A decisive downside gap that would have looked scary to anyone leaning long into the open.
- A strong intraday recovery that erased the gap and left only a modest net gain vs. the pre-print close.
In the scoring framework, that combination yields:
- A down outcome for the formal gap metric.
- A small positive but effectively “flat-ish” full-session move — not large enough to qualify as a meaningful reversal by the strict rules, even though the intraday path was dramatic.
5) Model vs. crowd: where the thesis broke
The preview’s core stance:
- Outcome: Beat.
- Direction: Up on the gap, with an expected move around +6% from the reference price.
- Guide tone: Inline-to-constructive, with some caution but no fresh guidance trauma.
- Directional conviction: ~60% odds that the gap would be higher, with the rest split between a downside surprise and a near-flat result.
Where reality diverged:
-
Top line vs. expectations
- The setup assumed revenue would land at or slightly above consensus.
- Instead, revenue missed by about 2%, which was enough to validate fears that demand or timing in non-networking segments remained wobbly.
-
Guidance mix
- The thesis treated guidance risk as largely “already in the price.”
- In practice, a below-Street near-term revenue guide carried more weight than anticipated, even with raised out-year EPS. Investors are still in “show me” mode after the earlier reset, and they leaned heavily into the short-term revenue caution.
-
Investor sensitivity to near-term revenue
- The model underweighted how much the market would care about next quarter’s revenue line in an AI infrastructure name, relative to a stronger long-term EPS story.
Scorecard:
- Gap call: Wrong. Expected an upside gap; got a mid-to-high single-digit downside move.
- Full-session call: Wrong under the strict framework. The thesis envisioned a mid-single-digit rally; the actual net move was a small gain, closer to “flat.”
- Crowd: With 1 vote, the crowd was unanimously “Up,” and therefore also wrong on both the opening move and the day-after close.
Given the stated confidence (~60%), this sits exactly where a probabilistic model should sometimes fail: a modest edge that simply didn’t pay out, rather than a high-conviction stance that completely misread the regime.
6) How the example trades would have fared
The preview sketched three representative structures: a short-dated bull call spread, a put spread below support, and a short-front/long-back call diagonal.
High-level outcomes:
a) Short-dated bull call spread (directional)
- At the open: A nearly 9% gap down crushed the value of any near-the-money call spread centered in the low-$20s. Mark-to-open drawdowns would have been severe.
- By the close: The recovery into the mid-$23s rescued some value, but the net move from the pre-earnings close was still much smaller than the targeted +6–7%. Many strike choices would have ended with disappointing or negative P&L, especially for traders who stopped out near the open.
Net: The pure directional upside expression struggled, even though the stock eventually closed slightly higher than it started.
b) Short put / long lower put spread under the low-$20s
- At the open: The gap toward $20.90 pushed the short leg close to or through strike, creating ugly marks and real stress for anyone watching the tape tick-by-tick.
- Into expiration: With the stock back above $23 by the end of the reaction session, put spreads struck around the $20 area likely expired worthless, allowing sellers to keep the entire credit.
Net: Structurally, the idea of selling downside where protection demand had been heavy worked, but only for traders willing (and sized appropriately) to sit through a sharp opening drawdown.
c) Front-short / back-long call diagonal
- The short front-week call benefited from the event-vol crush and the fact that the stock did not finish wildly above pre-earnings levels.
- The longer-dated call kept exposure to the raised FY26 EPS and AI/networking story without being overly punished by the short-term shakeout.
- As long as the short strike wasn’t too tight, this structure likely performed best relative to its intent, monetizing the rich event vol while preserving upside in the cleaned-up post-print narrative.
Net: Structures that primarily targeted vol dislocation across the curve, rather than a precise directional gap outcome, came through the event in better shape.
7) Lessons for future AI-infrastructure prints
This event offers a few clear takeaways for similar setups:
-
“EPS beat, revenue miss, soft near-term guide” is usually a downside-gap template.
Even with raised out-year EPS, investors anchored on the immediate revenue disappointment and punished the stock at the open. -
Near-term revenue guidance can dominate long-term EPS improvements.
For AI infrastructure names with a history of resets, the next-quarter revenue line is often the primary risk axis — and deserves heavier weight in the directional call. -
Implied move sizing can be right while the direction is wrong.
The options market was pricing a high single-digit move; that’s roughly what happened, but on the downside gap, not the upside. Directional confidence needs to account for this symmetry more explicitly. -
Separating vol trades from directional bets matters.
The diagonal-style structure, which focused on harvesting event vol while keeping longer-term upside, fared better than pure bullish spreads that relied on a clean up-gap.
In short: the fundamental story — record revenue, improving margins, and raised FY26 EPS — was broadly in line with a constructive long-term view. But the near-term revenue guide and the market’s sensitivity to any whiff of “still not fixed yet” around guidance made this a land mine for an up-gap call. Both the model and the crowd ended up on the wrong side of the overnight move, and this print is a useful reminder to treat near-term revenue guidance as a first-class driver in similar AI-heavy setups.