Hewlett Packard Enterprise Company
It sells the servers, storage and network equipment that companies install in their own data centres. Since it bought Juniper in 2025, a quarter of it is networking.
The scores
Growth3/4 ↑
Revenue grew 33.7%, but a $13.4B acquisition and rising selling prices supplied it; the engine underneath ran flat for eight years.
Business quality2/4 ↑
In three quarters of revenue the buyer decides on price and management guides that margin toward 13%; the protected quarter is networking.
Financial position3/4 ↑
Net debt is falling, nothing is drawn on the facilities and no maturity wall sits inside the horizon; but tangible net worth is negative.
Execution record3/4 ↑
Beat guidance in seven of eight quarters and raised fiscal 2026 twice; a $1.6B goodwill writedown on the segment it reorganised away caps it.
Per share capture2/4 →
The mandatory convertible delivers 76 to 93 million shares in September 2027, and quarterly buybacks of $135M do not offset yearly stock pay.
Potential2/4 ↑
A record order backlog with nothing firm in it, and the spending decision belongs to buyers who fund their own building by selling shares.
Risk profile3/4 →
One distributor is about 10% of sales and memory supply is externally controlled; no binary event survives now the Juniper settlement is approved.
Valuation3/4 ↓
About $17 of company value for each $1 of the operating profit the 2027 framework implies - the plan working, and little beyond it.
Evidence quality3/4 →
Filed figures throughout, corroborated by four other filers; the order, backlog and 2027 numbers exist only on a call and are graded as claims.
The Verdict
We rate Hewlett Packard Enterprise neutral. It sells servers, storage and network equipment that companies install in their own data centres. Three quarters is servers and storage, where buyers decide on price and margins are falling; networking, largely bought with Juniper, is where the money is. So the business is mixed. The price is fair: about right if management's 2027 plan arrives, with little extra credit for beating it and no cushion if it slips. The view hinges on networking orders turning into revenue at networking's margin. Orders are at a record, but nothing filed yet shows them earning what the plan needs.
The long caseCompanies are moving AI work onto equipment they own instead of renting cloud capacity; HPE gains only if its networking, not its low-margin servers, carries that demand.
The evidence under this - 7 claims
- The same management guided fiscal 2026 adjusted earnings to $2.25 to $2.45 on 4 December 2025 and now guides $3.75 to $3.85, so its own one-year-ahead framework has already been revised upward by about 62% once.
- The framework excludes the AMD Helios opportunity entirely, on management's own statement, so the number is not a full forecast of the year it describes.
- Cloud & AI margin of about 13% is below the 17.0% just reported, which is management telling you the mix is dilutive.
- Every input except the enterprise value comes from a framework management has already revised upward once by about 60%, so the denominator is the least reliable number in the calculation.
- Holding amortisation flat at $1.26B assumes no further acquisition and no step-down as the Juniper intangibles age, and the fiscal 2025 10-K's own future-amortisation schedule is not flat.
- The same enterprise value against the last four reported quarters of operating profit, $2,602M, is about $37 for each $1 - more than twice as much, because the fourth quarter of fiscal 2025 earned nothing.
- Networking was $2,893M of $12,213M in the same quarter - 23.7% of revenue - and the fiscal 2027 framework grows Cloud & AI at 14% to 18% against Networking's 14% to 17%, so the mix does not move toward networking on management's own numbers.
- The quarter's revenue growth came 59% from Cloud & AI and 41% from Networking, so the company is becoming more of a server company by the only measure that had a filed figure.
The company
When a company, government or smaller cloud operator wants AI computing it owns, HPE delivers it: servers, storage and the switches between them, installed, supported for years and often leased. It buys nearly every chip inside; what it adds is the integration.
The test of what that is worth runs outside HPE. Super Micro sells the same AI racks without the service and financing, and keeps 10.8 cents of each sales dollar before running costs. HPE's server and storage business kept 17.0 cents after them last quarter.
The evidence under this - 3 claims
How it makes money
It is paid per system, plus multi-year support contracts and, for about a seventh of the business, lease payments to its own finance arm. Servers are tendered again on price every time. Networking, built largely on the $13.4B Juniper purchase, is where customers stay, and it earned 22.0 cents of operating profit per dollar last quarter against 17.0 for servers and storage.
The evidence under this - 2 claims
Where the money comes from, stream by stream
| Stream | Share of sales (nine months to July 2026) | How it earns it |
|---|---|---|
| Server | 51.1% | General-purpose and AI servers, including rack-scale AI systems, sold to companies, governments and cloud providers, directly and through distributors. |
| Networking: Campus & Branch | 12.4% | Aruba and Juniper Mist wireless access points, switches and management software for offices, campuses and branches. |
| Storage | 11.0% | Alletra storage arrays, data protection and GreenLake private-cloud software, sold as hardware or as a subscription. |
| Financial Services | 8.3% | Leases, loans and pay-as-you-use contracts financing customers' HPE and other equipment, earning interest and rent. |
| Networking: Routing | 7.3% | Juniper routers for the wide-area and backbone networks of telecom carriers, cloud providers and large enterprises. |
| Networking: Data Center and Security | 6.1% | Data-centre switches, including fabrics for AI clusters, and firewalls and security software. |
| Corporate Investments and Other | 2.5% | Smaller lines moved out of Networking, Telco and Instant On, both sold or being divested, and incubation businesses. |
| Cloud & AI: other | 1.4% | Remaining Cloud & AI revenue the filer does not break out further. |
Who runs it
Antonio Neri has run HPE since 2018, after joining HP in 1996; Marie Myers has run its finances since 2024, and Rami Rahim, who ran Juniper for a decade, now runs networking. Neri bought Juniper and sold the H3C stake, but sales barely grew for most of his time in charge. Insiders own well under 1% of the company. Bonuses ride on sales, profit and free cash; the long-term shares ride on profit growth and the stock, none on returns on capital.
The evidence under this - 6 claims
- Half of the long-term award vests on service alone, so it pays whatever the business does.
- No metric pays on return on capital, and HPE earns the lowest return on invested capital of its named rivals (C23).
- The committee raised the bonus and PARSU goals after the Juniper close, so the deal could not inflate payouts; the fiscal 2023 and 2024 PARSUs earned 77% and 85% of target on below-target net income growth.
- The cost reduction programme predates the agreement: $148M of it ran through the nine months to 31 July 2025, before Elliott's seat and in the same year the agreement was signed.
- HPE has divested regularly for a decade, from the HP Inc. separation onward, so portfolio simplification is the company's normal behaviour rather than evidence of an outside hand.
- No filing states a causal link, and the amendment that is filed changes only the board's maximum size.
Why it wins, and who could beat it
Why customers come back
The advantage covers about a quarter of the company. A campus network is run day to day through HPE's management software, so replacing the boxes means replacing how the team works. Its routers run chips HPE designs itself, which a telecom operator qualifies once and then keeps.
The servers have no such hold, and management says so: it expects their margin to fall to about 13% next year, and networking's to rise.
The evidence under this - 2 claims
- The same management guided fiscal 2026 adjusted earnings to $2.25 to $2.45 on 4 December 2025 and now guides $3.75 to $3.85, so its own one-year-ahead framework has already been revised upward by about 62% once.
- The framework excludes the AMD Helios opportunity entirely, on management's own statement, so the number is not a full forecast of the year it describes.
- Cloud & AI margin of about 13% is below the 17.0% just reported, which is management telling you the mix is dilutive.
The threat
In servers, Dell and Super Micro sell the same AI racks, and both earn less as that business grows: Dell's infrastructure margin fell from 12.8% to 11.7% in the year its AI server revenue rose 166%. In data-centre switching, HPE's new switch runs on the same Broadcom chip as Arista's, so there is no hardware edge there either.
The evidence under this - 3 claims
The rivals: DELL, SMCI, ANET, NTAP, CSCO, EXTR.
Who else does this, rival by rival
| Ticker | Rival | How far along | Priced against it |
|---|---|---|---|
| DELL | Dell TechnologiesPriced against it: Yes | On sale | Yes |
| SMCI | Super Micro ComputerPriced against it: Yes | On sale | Yes |
| ANET | Arista NetworksPriced against it: Yes | On sale | Yes |
| NTAP | NetAppPriced against it: Yes | On sale | Yes |
| CSCO | Cisco SystemsPriced against it: Yes | On sale | Yes |
| EXTR | Extreme NetworksPriced against it: Yes | On sale | Yes |
How big it could get
How big this could get
The market is AI servers and the networks between them. Two suppliers alone booked over $60B from it last year, Dell $24.68B and Super Micro $39.06B, at roughly 11 cents of profit per dollar. Buyers are still accelerating: Oracle, an HPE customer, spent $28.5B on capital projects in one quarter, against $8.5B a year earlier. The size is not in doubt; what it pays is.
The evidence under this - 3 claims
The results
The numbers
Sales grew 27% over the twelve months to Jul 2026, faster than the 14% in the year to Oct 2025.
Operating margin is 6.2%, about where it was (7.2%) in the year to Oct 2023.
Yearly sales growth has ranged from +2.2% to +14% over the last 5 years.
Each dollar put into the business earns 5.5% a year after tax, about where it was (3.3%) in the year to Oct 2021.
It turned 9.9% of sales into free cash, $4.16B over the last twelve months.
It owes $14.0B more than the cash it holds, about 2.1 years of the cash its operations bring in.
The share count rose 8.9% in 4.8 years, and pay in shares costs 1.9% of sales.
All the numbers
Filed figures from the company's 10-K and 10-Q, read 2026-10-04. Fiscal years end in October.
| Oct 2021 | Oct 2022 | Oct 2023 | Oct 2024 | Oct 2025 | 12 mo to Jul 2026 | |
|---|---|---|---|---|---|---|
| Size and growth | ||||||
| Sales | $27.8B | $28.5B | $29.1B | $30.1B | $34.3B | $41.9B |
| Sales growth | +3.0% | +2.6% | +2.2% | +3.4% | +14% | +27% |
| Profit | ||||||
| Operating margin | 4.1% | 2.7% | 7.2% | 7.3% | −1.3% | 6.2% |
| Net margin | 12% | 3.0% | 7.0% | 8.6% | 0.2% | 6.7% |
| What growth costs, per sales dollar | ||||||
| Research and development | 7.1% | 7.2% | 8.1% | 7.5% | 7.3% | 8.8% |
| Selling, general and admin | 18% | 17% | 18% | 16% | 17% | 17% |
| Pay in shares | 1.4% | 1.4% | 1.5% | 1.4% | 1.9% | 1.9% |
| Capital spending | 9.0% | 11% | 9.7% | 7.9% | 6.7% | 6.1% |
| Cash | ||||||
| Cash from operations | $5.87B | $4.59B | $4.43B | $4.34B | $2.92B | $6.69B |
| Free cash flow | $3.37B | $1.47B | $1.60B | $1.97B | $627M | $4.16B |
| Balance sheet | ||||||
| Cash and short-term investments | $4.00B | $4.16B | $4.27B | $14.8B | $5.77B | $6.22B |
| Debt | $13.4B | $12.5B | $12.4B | $18.2B | $22.4B | $20.2B |
| Net cash | −$9.45B | −$8.30B | −$8.09B | −$3.40B | −$16.6B | −$14.0B |
| Returns and shares | ||||||
| Return on capital | 3.3% | 2.4% | 5.8% | 5.0% | - | 5.5% |
| Shares (diluted) | 1.33B | 1.32B | 1.32B | 1.34B | 1.32B | 1.45B |
What growth costs
Growth has not made HPE leaner. Sales are up by half in five years, but research, selling and office costs still take about the same share of each sales dollar, and research took more once Juniper arrived. What decides the margin is the mix: AI servers earn thin margins and networking fat ones, so profit depends on which grows faster.
The evidence under this - 3 claims
Juniper joined in July 2025, so the later years carry a networking business with higher research spending; the rise in R&D is partly perimeter, not a worsening of the old business.
- The same management has raised fiscal 2026 guidance twice this year, so its expectations about its own margin have been conservative rather than accurate.
- The moderation is guided rather than observed: gross margin has risen in each of the last four reported quarters, and nothing in a filing yet shows it turning.
- Part of the mix effect runs the other way - networking, which carries the higher margin, is guided to grow 14% to 17% in fiscal 2027 against the company's 13% to 17%.
- The fiscal 2025 fourth-quarter release gives GAAP gross margin of 33.5% against non-GAAP of 36.4% in the same quarter, so the two definitions differ by roughly three points and this calculation is closer to the second.
- The comparative quarter contains only one month of Juniper, whose networking margins are far above the server business's, so part of the 10.9-point move is the acquisition arriving rather than pricing.
The last report (2 Sep)
Quarter ended July 2026, filed figures.
In the third quarter of fiscal 2026 revenue rose 33.7% to $12.21B, and operating profit reached $1.39B against $247M a year earlier. Orders grew faster still, 42% on management's measure, and networking orders 36%, about 3.5 times networking revenue growth.
The detail is less comfortable. Server revenue grew 35% because prices rose while short supplies of parts capped how many shipped. And inventory nearly doubled to $11.82B in nine months, paid for by owing suppliers more.
What it did to the thesisIt confirmed strong demand, and the order growth is what the market prizes. It also supported our caution: servers grew on price while parts were short, inventory piled up, and networking sales have not yet matched networking orders. The price leaves little room, so the view stays Neutral.
The quarter in detail
Sales growth from the previous quarter, filed figures.
Earnings per share $1.11 against $0.93 expected; $1.06 as filed, the gap being what the company adds back.
The evidence under this - 6 claims
- Networking revenue grew 10% on the same normalised basis in the quarter, which an analyst on the call called light against what peers were reporting, so the order figure has not yet shown up as revenue anywhere.
- No filing carries an order or backlog figure for any segment, so the whole series exists only in management's telling and cannot be checked against a statement.
- Data centre networking revenue actually fell 6% in the quarter on shipment timing, which is the part of the order book the AI story rests on.
- Neri told Mark Newman on the same call that units would strengthen in Q4, so the company reads the constraint as temporary rather than structural.
- Dell's own filing attributes its traditional-server price rise to 'richer configurations' rather than to scarcity, which would make part of the same move a genuine content gain rather than pure price.
- Super Micro's filing reports a 34% average selling price rise alongside higher shipment volume, so ASP and units are not always the trade-off this claim assumes.
What management did with the money
It paid about $13.4B in cash for Juniper and is taking cost out, aiming for $800M a year by fiscal 2028, up from $600M. Since then it has sold its last stake in H3C and a telecom unit, promised to return at least 75% of free cash, and seated a partner of Elliott, an activist investor, on its board.
The evidence under this - 4 claims
- Neither figure appears in any filing. The 10-Q carries no synergy number and no leverage ratio, and the ratio's definition - which debt, which profit - is stated nowhere.
- Net leverage of 1.8 times almost certainly excludes the debt funding HPE Financial Services, which is most of a $20.24B total, so it is not comparable to a gross figure a reader would compute.
- The $600M target runs to the end of fiscal 2028, two years out, and management gives the run rate and the end date but not the phasing - so nothing in it can be checked against a print.
- The raised $800M target is a further management projection in an investor-day release; no filing carries a synergy figure, so it cannot be checked against a statement.
- 2026-10-01editedNetworking Investor Day release of 2026-09-30 raised the Juniper savings target from $600M to $800M a year
- The cost reduction programme predates the agreement: $148M of it ran through the nine months to 31 July 2025, before Elliott's seat and in the same year the agreement was signed.
- HPE has divested regularly for a decade, from the HP Inc. separation onward, so portfolio simplification is the company's normal behaviour rather than evidence of an outside hand.
- No filing states a causal link, and the amendment that is filed changes only the board's maximum size.
The price, and where we disagree
In short. Fair, with little to spare. The price already pays for management's plan, and treats HPE as the networking company its chief executive says it is becoming.
The shares rose 154% over the past year, and 12.4% on 11 September alone on Oracle's spending figures rather than any HPE news. In nine months management raised its fiscal 2026 earnings guidance from $2.25 to $2.45 a share to $3.75 to $3.85.
The full-year results on 3 December are the first to show whether the networking orders arrive as revenue at networking's margin.
The evidence under this - 3 claims
- Networking revenue grew 10% on the same normalised basis in the quarter, which an analyst on the call called light against what peers were reporting, so the order figure has not yet shown up as revenue anywhere.
- No filing carries an order or backlog figure for any segment, so the whole series exists only in management's telling and cannot be checked against a statement.
- Data centre networking revenue actually fell 6% in the quarter on shipment timing, which is the part of the order book the AI story rests on.
How we read the multiple
Counting debt, a buyer pays about $28 for each dollar of last year's operating profit, what is left from sales after running costs. Its rivals sit around $29.
On the profit management's plan for fiscal 2027 implies, after the cost of past acquisitions, it is about $17. That pays for the plan and not much more.
The plan needs networking to grow fast enough to outweigh servers, whose margin management expects to fall.
Against its own history. Far dearer than its own normal on every measure: buyers have stopped pricing it as the server company that went eight years without growing.
The evidence under this - 6 claims
- Every input except the enterprise value comes from a framework management has already revised upward once by about 60%, so the denominator is the least reliable number in the calculation.
- Holding amortisation flat at $1.26B assumes no further acquisition and no step-down as the Juniper intangibles age, and the fiscal 2025 10-K's own future-amortisation schedule is not flat.
- The same enterprise value against the last four reported quarters of operating profit, $2,602M, is about $37 for each $1 - more than twice as much, because the fourth quarter of fiscal 2025 earned nothing.
- The same management guided fiscal 2026 adjusted earnings to $2.25 to $2.45 on 4 December 2025 and now guides $3.75 to $3.85, so its own one-year-ahead framework has already been revised upward by about 62% once.
- The framework excludes the AMD Helios opportunity entirely, on management's own statement, so the number is not a full forecast of the year it describes.
- Cloud & AI margin of about 13% is below the 17.0% just reported, which is management telling you the mix is dilutive.
- Networking was $2,893M of $12,213M in the same quarter - 23.7% of revenue - and the fiscal 2027 framework grows Cloud & AI at 14% to 18% against Networking's 14% to 17%, so the mix does not move toward networking on management's own numbers.
- The quarter's revenue growth came 59% from Cloud & AI and 41% from Networking, so the company is becoming more of a server company by the only measure that had a filed figure.
Priced against its rivals
The competitors its own annual report names, each confirmed against the rival's filing. Measured on company value for each $1 of operating profit, cheapest first:
| Ticker | Company | company value for each $1 of operating profit |
|---|---|---|
| SMCI | Super Micro Computer | 9.1× |
| NTAP | NetApp | 21.1× |
| DELL | Dell Technologies | 27.7× |
| HPE | Hewlett Packard Enterprise Company | 27.8× |
| CSCO | Cisco Systems | 30.2× |
| EXTR | Extreme Networks | 45.6× |
| ANET | Arista Networks | 54.8× |
The middle of the other 6 is 28.9. Hewlett Packard Enterprise Company is 4th cheapest of 7.
Rivals' figures come from a data provider, so read the spread of the group rather than any single pair.
Close to the middle of the rivals its annual report names, so the price reflects the whole group's AI boom, not HPE being valued above its rivals.
Also checked and set aside: International Business Machines (a substitute, not the same product); Nutanix (a substitute, not the same product); Ubiquiti (a substitute, not the same product); Ciena (a substitute, not the same product); Nokia (a substitute, not the same product).
Where we disagree with the market
An AI infrastructure winner, and the record order backlog is the reason to own it.
The earnings doubling is real but mostly bought with Juniper, and the part growing fastest earns the least. Management says so itself.
The price moves on the order number. The segment note and the rivals' own filings, all public, show what those orders earn: the AI server business is guided to about 13%, and Dell and Super Micro make less as their AI sales grow.
We would be wrong if networking grows 14% to 17% at a margin above 25% through fiscal 2027, with the server margin holding above 15%.
What analysts say, and the evidence
The evidence under this - 5 claims
- The same management guided fiscal 2026 adjusted earnings to $2.25 to $2.45 on 4 December 2025 and now guides $3.75 to $3.85, so its own one-year-ahead framework has already been revised upward by about 62% once.
- The framework excludes the AMD Helios opportunity entirely, on management's own statement, so the number is not a full forecast of the year it describes.
- Cloud & AI margin of about 13% is below the 17.0% just reported, which is management telling you the mix is dilutive.
What could go wrong, and how we would know
The risks
- Built on goodwill. Purchased goodwill and intangibles exceed shareholders' equity, leaving tangible net worth near minus $2.88B, and fiscal 2025 already carried a $1.6B write-down.
- Margins falling back. Management expects gross margin to return toward historical levels as AI systems become a bigger share of sales.
- Weakest returns in its group. HPE earns the lowest return on the money invested in it of the six rivals its annual report names.
- Insiders selling. Insiders sold 25 times in the year and bought none; the chief executive sold 150,000 shares in April.
The evidence under this - 5 claims
- The same management has raised fiscal 2026 guidance twice this year, so its expectations about its own margin have been conservative rather than accurate.
- The moderation is guided rather than observed: gross margin has risen in each of the last four reported quarters, and nothing in a filing yet shows it turning.
- Part of the mix effect runs the other way - networking, which carries the higher margin, is guided to grow 14% to 17% in fiscal 2027 against the company's 13% to 17%.
What we still do not know
- How much of next year's growth is already ordered. Management publishes no backlog figure.
- Whether customers stay with HPE's network software. Nobody in this market publishes a retention figure.
The evidence under this - 2 claims
What would change our view
- Neutral→AttractiveThe network half shows up in growth and cash
It needs all of:
- ○Networking revenue grows more than 14% for two quarters running, while servers and storage keep their margin above 13%.checked at the next results, due by 3 Dec
- ○Free cash flow for fiscal 2027 exceeds $4B.checked at the next results, due by 3 Dec
- Neutral→UnattractiveThe server line is price, or the goodwill goes again
It needs any of:
- ○Server revenue keeps growing through higher prices rather than more units, all through fiscal 2027.checked at the next results, due by 3 Dec
- ○The goodwill carried for servers and storage is written down again.watched as it happens
We would also be wrong if
- Networking growth stays under 14% for two quarters despite the record orders.
- The server margin falls below 13% before fiscal 2027 ends.
- Fiscal 2027 free cash flow comes in below $4B.
What is coming
- December 2026 Fiscal 2026 annual report (10-K): The first full year on the new segments with twelve months of Juniper, the year's goodwill test, and whether the distributor stays near 10% of revenue.
Every dated event ahead
- December 2026 Fiscal 2026 annual report (10-K): The first full year on the new segments with twelve months of Juniper, the year's goodwill test, and whether the distributor stays near 10% of revenue.
- 3 December 2026 Q4 and full-year fiscal 2026 results: Whether Networking grows at least 11% at a segment margin above 22%, Cloud & AI margin holds in the mid-teens, and the fiscal 2027 cash floor is restated.
- February 2027 2027 proxy statement: What fiscal 2026 pay earned against the free cash flow metric added this year, what insiders own after this year's sales, and the board seats under the Elliott agreement.
- September 2027 Mandatory convertible preferred converts to common: Whether 76 to 93 million new shares arrive as scheduled and whether buybacks absorb any of them, the per-share test.
Since we wrote this
- 5 OctThesis reviewThe long case now sits under the verdictAfter the five-name pilot, the owner asked for the long case on every thesis.
- 4 OctThesis reviewOwner's checklist: revenue streams, cost of growth, management, catalystsThe owner's checklist found the method silent on who runs the company, what growth costs, where the money comes from and what is dated ahead; this pass fills those four blocks from the proxy and filings.
- 1 OctNewsNetworking outlook and savings raised, $1.2B order; view heldHPE now projects high-teens to low-20s Networking growth next year and $800M of Juniper savings, plus a first $1.2B AMD rack order. These are projections; reported results on 3 December decide.
- 26 SepThesis reviewJournal completed from the written record; view heldThe opening entry lacked the view and conviction the thesis was written with; they are filled from the original file. Nothing in the argument moved.
- 12 SepThesis writtenNeutral, medium convictionWritten from the third-quarter fiscal 2026 filing, whose segment note showed the fast-growing server business earning less than the networking quarter bought with Juniper.
Also checked 4 times with nothing new to act on.
Sources 15
Sources - 15
Every claim above cites one of these, or states the reasoning that derived it. The quote is frozen at the date it was read: a link can rot, and an agent fetch returns a reading of a page rather than the page. Backs counts the whole claim ledger, so it can name a claim this page does not surface.
- S01primarycompany filing2026-09-03the income statement, the balance sheet, the cash flow, the segment note, goodwill, borrowings and the H3C gainbacksC01C02C03C04C10C11C12C16C17C18C19C36C42C45C46C47
- S02primarycompany filing2025-12-18the business description, the named competitor lists segment by segment, the Hybrid Cloud goodwill impairment, distributor concentration and the mandatory convertible's termsbacksC13C18C30C32C45C46C47
- S03primarytranscript2026-09-02orders, backlog, Networks for AI order figures, the fiscal 2027 framework, and eleven analyst questions including the one on margin normalisationbacksC05C06C14C15C25C26C27C31C34C35
- S04primarycompany filing2025-12-04the fiscal 2026 guidance as first issued, and the old five-segment margins including Server at 9.8% and Hybrid Cloud at 5%backsC14
- S05primarycompany filing2026-03-16a rival's filed AI-optimized server revenue and the segment margin it earned while that revenue tripledbacksC07
- S06primarycompany filing2026-08-31the gross margin a pure rack-scale AI systems maker earns, and its own average selling price disclosurebacksC08
- S07primarycompany filing2026-02-17a competitor's own statement that its switches run on Broadcom merchant silicon, which is the test of where HPE's networking advantage is notbacksC31
- S08primarycompany filing2026-09-11a named HPE customer's own capital spending and remaining performance obligations, and how that spending was fundedbacksC28C29
- S09secondaryvendor data2026-09-12Research pack, frozen 12 September 2026content/HPE/research-pack.jsonprice, the multi-year multiple medians, the analyst split, insider filings, short interest and the filed-competitor cohortbacksC09C22C23C24C33
- S10primarycompany filing2026-09-12the filed annual revenue and operating margin series back to fiscal 2013, restated as the company restated itbacksC09C46C47
- S11primarycompany filing2026-06-01the fiscal 2027 framework as first issued, which the Q3 release then raisedread, but no claim rests on it
- S12primarycompany filing2026-09-03the board-size cap, the Investor Director seat and the date of the original agreementbacksC38
- S13primarycompany filing2026-06-01who the Investor Director is, which committees he sits on, and Elliott's own statement of what it wantsbacksC38
- S14primarycompany filing2026-09-30Raised fiscal 2027 Networking growth outlook and margin range, segment growth rates to fiscal 2029, Juniper savings target raised to $800M, and the $1.2B Vultr AMD Helios orderbacksC40C41
- S15primarycompany filing2026-02-11beneficial ownership at 31 December 2025, the fiscal 2025 and 2026 bonus metrics and weights, the PARSU design and payouts, the CEO's long-term award mix and the Juniper goal adjustmentsbacksC43C44