HPE at ~13x FY27 Earnings: A Selloff Built on the Wrong Variable

HPE at ~13x FY27 Earnings: A Selloff Built on the Wrong Variable

In an ironic move, Hewlett Packard Enterprise is trading down roughly 9% over the last five sessions and sitting about 17% below its high — sold off for the exact reason its most recent quarter disproved. The trigger is the usual August scare: a broad technology selloff trained on AI infrastructure names, plus a fresh U.S. International Trade Commission investigation into DDR5 memory modules opened by licensor Netlist. The ITC — the U.S. trade tribunal with import-blocking power — is a real overhang, but it is a small number against this company's scale. After more than doubling this year to 2.7x its 52-week low of $19.84, HPE is a momentum name, and momentum names bleed hardest when the AI-hardware trade rotates.

But the market is pricing the wrong engine. Wall Street spent years valuing HPE on hyperscaler server-box volume — a thin-margin, lumpy business. The company's own numbers say the earnings power has already migrated: into the high-margin networking and enterprise-AI mix that the Juniper Networks acquisition locked in. That is the divergence this pullback quietly re-prices, and it leaves the forward math looking better than the panic.

The wrong engine

The quarter that ended April 30 was the evidence. Revenue rose 40% year over year to $10.68 billion, adjusted EPS of $0.79 nearly doubled, and non-GAAP gross margin expanded about 750 basis points to 36.9%. The margin is the story, because it proves the mix shift is real, not aspirational.

Networking — the second engine Juniper created, after the $14 billion deal cleared the Justice Department's divestitures and licensing commitments and closed in mid-2025 — is where it shows up. Networking revenue grew 148% year over year to $2.7 billion at roughly a 22% operating margin. Routing, Juniper's backbone business, went from $1 million a year ago to $775 million. Security grew 155%, data-center networking 233%. Even the old server core grew 33%, because enterprises are now buying for agentic AI — software agents running as core workloads — rather than the commodity boxes of the hyperscaler build-out.

The point is not any single line. It is that the profitable segments are compounding fastest. A company whose highest-margin franchise compounds at triple-digit rates — in fiscal Q4 2025, networking ran 150% higher than the year before — is a different asset than the hardware assembler the selloff is repricing.

The forward math

At about $53.45, the arithmetic runs like this:

1. The multiple. HPE sits at roughly 15.7x the midpoint of its raised fiscal 2026 adjusted-EPS guidance ($3.35–$3.45) and only about 13.2x fiscal 2027 consensus EPS of about $4.04. Against management's own FY27 framework — 8–12% revenue growth, 12–16% adjusted-EPS growth — that is a PEG (price-to-earnings-to-growth) near one, the classic GARP boundary. On trailing EV/EBITDA, earnings before interest, taxes, depreciation and amortization, HPE runs about 17x against Dell's roughly 31x — half the peer's claim on future cash flow.

   

2. The floor just moved up. Three months ago consensus put FY27 EPS at $2.75; it now sits near $4.04, a ~47% upward revision in a single quarter. That happened because management pulled its 2028 financial targets forward by two years, raising full-year 2026 revenue guidance to 29–33% growth and adjusted EPS to $3.35–$3.45 from a prior $2.30–$2.50. The stock rose a lot; the earnings floor rose faster. That is a gap closing, not a bubble inflating.

3. The honest counter. Reported GAAP EPS of about $2.47 puts the multiple near 21x, not 13x — the gap is mostly amortization of Juniper's acquired intangibles plus stock comp, real for years but non-cash. The cash anchor is cleaner: management guides to at least $4.5 billion of free cash flow in FY27 against a market cap around $70.8 billion, roughly a 6% cash yield. The 13x is the operating-business frame; on cash it is cheaper still.

4. Orders outrank headlines. HPE booked $1.8 billion in AI systems orders in a single quarter, and hyperscalers have budgeted more than $700 billion of AI infrastructure spending this year. The bear case has to rest on AI capex cracking, networking growth decelerating from +148% to single digits, or gross margin rolling back over. Nothing in the twice-raised guidance or the order book supports that. Orders are data; "sustainability of AI spending" is a headline.

5. The catalysts. HPE finished exiting the H3C Chinese joint venture in late May, bringing total pretax consideration to roughly $3.5 billion and clearing a geopolitical overhang. Management has earmarked that cash to cut net leverage toward roughly 2x by the end of fiscal 2027 — every dollar lowers interest expense and funds buybacks, mechanically lifting the very EPS line the selloff just marked down. Even the usually reactive sell-side consensus carries a median target near $69, about 29% above the current price. AInvest's aggregate signal labels the stock Buy; I use that only to cross-check — the math, not the label, does the work.

Bottoming condition

The stock may still need to find a bottom — momentum names don't bottom on demand, and the Netlist ITC case is a live overhang even if the sums are small. That is a timing gate, not a thesis breaker. On about 13x fiscal 2027 EPS against a raised floor and a contracted order base, the forward math already beats the panic narrative. The break conditions are named: AI spending rolls over hard, networking growth stalls to single digits, or margins unfurl as mix shifts back to boxes. Until one of those shows up, this is a temporary headwind against a structural engine — and the market is pricing the wrong variable.