Everpure P

0. Ticker identity — read this first

This dive was preceded by a mandatory identity check, because single-letter tickers are reassigned between issuers and screener metadata goes stale. The result of that check is material enough to lead with.

What the vendor profile says the company is: Everpure, Inc., listed on the New York Stock Exchange under the symbol P, CIK 0001474432, CUSIP 74624M102, ISIN US74624M1027. Headquartered at 2555 Augustine Drive, Santa Clara, California. Chief Executive Charles H. Giancarlo. Founded in October 2009 by John M. Hayes and John Colgrove. 6,400 employees. Website everpuredata.com. Business description: "offers cutting-edge data storage solutions and services, empowering clients to extract maximum value from their information ... committed to revolutionizing data storage and management."

Three internal consistency checks, all of which pass: (1) the insider filings in this same file name John Colgrove as "director, officer: Chief Visionary Officer" and Charles H. Giancarlo as "director, officer: CEO" — matching the profile's founder and chief executive; (2) the peer set the vendor assigns is entirely enterprise-technology and storage — Astera Labs, Credo, Ericsson, HPE, HP, Keysight, NetApp, Super Micro, Sandisk — consistent with a data-storage company and inconsistent with anything else; (3) the financial profile (70.4% gross margin, 26.3% of revenue in R&D, a subscription-and-hardware Product/Service split, a fiscal year ending in early February) is unmistakably that of an enterprise storage vendor.

Two metadata fields in the profile are stale or wrong, and we flag them rather than paper over them:

• sector/industry are tagged "Industrials / Industrial - Machinery." This directly contradicts the company's own description, its peer set, its 70.4% gross margin and its 26.3% R&D intensity. We treat the sector tag as wrong and classify this dive as Technology / Enterprise Data Storage. Any screener that filters this name into an industrials bucket is misclassifying it.
• ipoDate is 2011-06-15, which cannot be reconciled with a company founded in October 2009 whose financial history in this file begins with fiscal 2021. This is the classic symptom of a recycled single-letter symbol carrying a prior issuer's listing date. Ignore the IPO date.

Non-equity tripwire: NEGATIVE — this is a common equity, not a bond, preferred or trust security. Price $80.14 (nowhere near a $20 or $25 par), beta 1.447 (not sub-0.55), zero dividend (not a fat fixed coupon), a wide 52-week band of $54.37-$100.59 (not narrow), full estimate coverage out to FY31 with 12-15 analysts on the near years, and 45 rating actions on file. Nothing in the profile name contains NT/NTS/JR/SUB/DB/COLLATERAL/TR MT/PFD/Series. This is unambiguously an operating-company equity and an equity dive is appropriate.

One item for the editor. The two knowledge-base claims that describe this business are tagged to the entity symbol PSTG, not to P. The claims describe the same company — an enterprise storage vendor inside the data-movement bottleneck and AI-factory infrastructure themes — and the vendor profile for P describes that same business. We have counted them (kb_claim_count: 2) and flagged the symbol mismatch here explicitly. The editor should confirm the entity-to-symbol mapping in the knowledge base before this dive is cross-linked to any portfolio or scorecard record. We have not silently merged two identifiers.

Reference table

1. What they actually sell

The product disclosure is simple and, unusually for this batch, complete:

The near-even split between hardware (Product) and recurring Service is the most important structural fact here. Service has grown from $738.5M in FY22 to $1.691B in FY26 — a 129% increase over four years, versus 37% for Product. The company is becoming a recurring-revenue business one year at a time, which is what makes the 70.4% gross margin sustainable and what makes deferred revenue of $1.214B meaningful.

Geography (FY26):

International is growing at more than twice the domestic rate and has risen from 27.7% to 32.7% of the mix in two years. That is a genuine second growth vector and it is under-discussed.

What is missing: there is no split between the company's product families, no customer-concentration disclosure, no hyperscaler-versus-enterprise breakdown, and no unit or capacity metrics. For a company whose thesis rests on AI-driven demand, the absence of any disclosure isolating AI-related revenue is a real gap — we are inferring the driver from the growth rate and the timing, not reading it off a line item. Stated plainly.

2. The inflection — five quarters, and one that changes the picture

Two things to read from this table.

First, the acceleration is real and it is measured against a rising base: 12.7% → 16.1% → 20.4% → 35.2%. Four consecutive quarters of accelerating year-over-year growth. That is not a comparison artefact; it is a trend, and the fifth data point is the one that changes the investment case.

Second, the quarterly seasonality is severe and it matters for how you read Q1. FY27 Q1 revenue of $1,052.9M was essentially flat against FY26 Q4's $1,058.9M, and operating margin fell from 8.2% to 1.9%. The same pattern appears a year earlier (FY25 Q4 $879.8M → FY26 Q1 $778.5M, with operating income going from +$42.5M to -$31.2M). This company's fiscal first quarter is structurally its weakest. The correct comparison is year over year, not sequential, and on that basis Q1 was excellent. But anyone reading the sequential operating-margin collapse as deterioration would be reading it wrong, and we flag it so the reader is not misled by their own chart.

Full-year record:

Note the FY24 air pocket: revenue grew just 2.8% and operating income fell. This company has had a flat year within the last three, which is a useful corrective to any assumption that 35% is now the base rate.

3. The GAAP problem — stated plainly

Free cash flow ($615.7M) is 3.3 times GAAP net income ($188.2M). The bridge is almost entirely stock compensation: $481.7M in FY26, equal to 13.2% of revenue and more than four times the $114.8M of GAAP operating income.

The consequences, stated without spin:

• Trailing GAAP P/E is 118x. The vendor reports 117.85x and our computation agrees.
• Every forward multiple in this dive — 32.9x FY27E, 26.7x FY28E — is a NON-GAAP multiple, computed on consensus EPS of $2.433 and $3.003 that exclude the stock-compensation charge. We cross-checked: consensus netIncomeAvg of $842.0M for FY27 divided by consensus EPS of $2.433 implies 346M shares, consistent with the actual diluted count, so the estimate series is internally coherent — but $842M of "net income" against $188.2M of GAAP net income in the year just closed tells you exactly how large the adjustment is.
• The mitigant is that the share count is not exploding. Diluted shares went 285.9M (FY22) → 339.2M (FY23) → 332.6M (FY24) → 342.7M (FY25) → 346.1M (FY26). That is +21% over four years against +68% revenue growth, and FY23-FY26 dilution is only +2.0% total. The buyback record is mixed (a $374.0M repurchase in FY25, but a positive $56.0M net stock issuance in FY26), so the dilution restraint is coming from grant discipline rather than from buying stock back.
• GAAP operating margin is nonetheless improving: -4.5% (FY22) → 3.0% (FY23) → 1.9% (FY24) → 2.7% (FY25) → 3.1% (FY26), and the FY26 quarterly path ran -4.0% → 0.6% → 5.6% → 8.2%. The direction is right. It is starting from a very low base.

Our position: the free cash flow is real and it is the number that matters, but the 27x FY28E headline is not comparable to a 27x multiple on a company that expenses its compensation in cash. We use the non-GAAP series because it is what the market prices, and we label it every time.

4. Balance sheet and cash flow

Balance sheet (FY26, 2026-02-01):

• Cash $854.9M + short-term investments $692.4M = $1.547B
• Total debt $216.1M (zero long-term debt, zero short-term debt — the $216.1M appears to be lease and other obligations)
• Net cash: $1.331B (my computation). The vendor's netDebt field reports -$638.7M because it nets only cash and ignores the $692.4M of short-term investments. I use the fuller figure and label it as my computation.
• Deferred revenue $1.214B (+23.2%) — this is the single most important balance-sheet line, because it is contracted revenue not yet recognised, growing faster than revenue itself
• Shareholders' equity $1.446B; accumulated deficit -$1.181B (narrowing from -$1.369B)
• Current ratio 1.62; interest coverage 49.2x

Enterprise value: market cap $26.639B minus $1.331B net cash = $25.31B (my computation). The vendor reports $26.032B, which uses the cash-only net-debt figure; the difference is the $692M of short-term investments. All EV multiples here use $25.31B.

Cash flow (four-year record):

Free cash flow has grown at a compound rate of only 0.3% since FY23 — it is lumpy, not compounding — because capex has risen 67% over the same period. The FY26 result is a new high, but this is not a smoothly growing cash-flow stream and we will not present it as one. Free-cash-flow yield is 2.3% on market cap; price-to-free-cash-flow is 51.6x.

5. Data integrity — one clear corruption, flagged and worked around

REJECTED (partially) — the most recent quarter's share-count and interest-expense fields. inc_q[0] (FY27 Q1, ending 2026-05-03) reports:

• weightedAverageShsOut = 326,539,000 and weightedAverageShsOutDil = 326,539,000 — identical to each other, and exactly identical to the year-ago quarter (2025-05-04). Every other quarter in the file has diluted shares meaningfully above basic (Q4 FY26: 330.5M basic versus 346.1M diluted, a 4.7% gap).
• interestExpense = $1,807,000 — exactly identical to the year-ago quarter, while the intervening quarters ran $1.051M, $0.266M and $0.251M.

Two fields duplicating the year-ago quarter to the dollar is the vendor-corruption signature this house screens for. Conclusion: the FY27 Q1 share-count and interest-expense fields appear to be carried forward from the prior-year quarter and are not reliable. The reported GAAP EPS of $0.07 is arithmetically consistent with that share count ($24.078M ÷ 326.539M = $0.0737), which means the EPS figure inherits the same doubt.

What we did instead: revenue ($1,052.9M) and net income ($24.1M) from that quarter are used, because they reconcile independently — revenue matches the earnings-calendar actual to the dollar, and the non-GAAP EPS actual of $0.47 against a $0.398 consensus is a separate, corroborating series. No per-share arithmetic in this dive is built on the FY27 Q1 share count. Share-count-dependent work uses either the FY26 diluted count of 346.1M or the market-implied 332.4M (market cap ÷ price).

Also flagged (not rejected):

• The consensus EBITDA series is internally implausible. FY27 estimated EBITDA of $930.1M against estimated net income of $842.0M implies that depreciation, amortisation, interest and tax together consume only $88M — impossible for a company with $264M of annual capex. No EBITDA-based forward multiple is computed in this dive. This is the same failure mode observed across multiple files in this batch; it appears to be a systematic vendor issue rather than a company-specific one.
• Two different 52-week ranges. The profile reports $54.37-$100.59; the technical block reports $54.51-$98.70. Small, but they do not agree, and the "percent from high" figure depends on which you use (-20.3% versus -18.8%).
• ratios_ttm.returnOnEquityTTM is null while km_ttm reports 16.13%. Use the latter.
• Sector and IPO-date metadata are wrong — see section 0.

Checked and found clean: the FY26 annual statements (revenue, gross profit, operating income, net income and EPS all reconcile); the earnings-calendar series (five quarters of actuals, all matching the income statement revenue to the dollar); the balance sheet (assets, liabilities and equity foot); the estimate EPS series (FY26 consensus $1.946 against a non-GAAP actual sum of $0.29+$0.43+$0.58+$0.69 = $1.99, a 2% error).

6. Knowledge base — thin, positive, and from a high-skill source

Two tagged claims, both from the same tracked source, at the highest skill weighting in the panel:

• jordi_visser (skill 2.0), 2026-01-18 (bullish, conviction 58, entity: PSTG): "Pure Storage is a great name that has done well but corrected over the last 18 months; fits the data-movement/storage bottleneck theme."
• jordi_visser (skill 2.0), 2026-06-17 (bullish, conviction 65, entities: DELL, HPE, PSTG): "AI-factory infrastructure names — Dell, HPE, Pure Storage — are part of the buildout; Dell already reached the 'insane numbers' earnings mode."

One claim EXCLUDED: a third claim on this business (2026-06-04, bullish, conviction 58 — "ready to start its move as a big player") is tagged to the quarantine_misattributed source and is therefore excluded from the count and from the conviction rating, per house policy. We note its existence rather than pretending the search returned two hits; excluding it is the correct handling and it is the reason kb_claim_count is 2 and not 3.

One theme-adjacent claim from a second, independent source:

• gavin_baker, 2024-02-01 (bullish, conviction 72): "All hard drives will be replaced by flash storage; storage is the layer below something like Snowflake, and the old EMC-style 'how to invest in data storage' opportunity is coming again."

That claim does not name the company, but it is the clearest statement in the knowledge base of the secular substitution this business is the direct beneficiary of, and it predates the tagged claims by two years.

Read. This is a Moderate conviction lane: thin in count (2), but both claims come from the panel's highest-skill-weighted source, both are recent (January and June 2026), both are bullish, and both place the company inside a named structural theme rather than making a price call. The adjacent Baker claim supplies the multi-year substitution logic. It is genuine support and it is not enough on its own to carry a full-size position — which is one reason the verdict is Stage-In. The symbol-mapping caveat in section 0 applies.

7. Valuation — priced in or room?

At $80.14 (market cap $26.639B, net cash $1.331B, EV $25.31B):

Trailing GAAP P/E: 118x. Price-to-free-cash-flow: 51.6x. Free-cash-flow yield: 2.3%. The FY29 EPS estimate rests on a single analyst and should be ignored.

7a. What today's price assumes (the inversion)

At $80.14 — 32.9x FY27E and 26.7x FY28E non-GAAP earnings, 5.6x FY27E sales — the price embeds the following falsifiable claims:

• Revenue reaches ~$4.49B in FY27 (+22.5%) and ~$5.16B in FY28 (+15.0%) (consensus, 15 analysts each). Note this is below the +35.2% just delivered — the price does not require the Q1 growth rate to persist; it requires roughly two-thirds of it. That is the most encouraging thing in this section.
• Non-GAAP EPS reaches ~$3.00 by FY28 (consensus), i.e. 23% compound growth from the $1.99 FY26 base, which requires operating leverage on a cost base where R&D alone is 26.3% of revenue.
• Gross margin holds above 70% (FY26: 70.4%; four-year range 67.5%-71.4%) — no price war with the larger storage vendors in the peer set (HPE $66.5B, NetApp $35.9B, Sandisk).
• Stock compensation grows more slowly than revenue, so that the gap between $842M of consensus "net income" and GAAP profitability narrows rather than widens. If stock compensation stays at 13.2% of revenue, FY28 would carry a $680M charge — and the non-GAAP multiple would remain a fiction rather than a forecast.
• The market keeps paying ~27-33x forward non-GAAP earnings — no de-rating, despite a trailing GAAP P/E of 118x and a 2.3% free-cash-flow yield.

The single most fragile assumption: that one quarter is a trend. The entire re-rating case rests on FY27 Q1's +35.2%, a single observation following a full year at +15.6% and, two years before that, a year at +2.8%. Consensus for the 2026-08-26 quarter implies +28.0% ($1.102B against $861.0M) — i.e. the street itself is modelling deceleration from Q1. If the actual comes in near +20%, the inflection narrative loses its evidence base and 33x forward has nothing holding it up.

7b. The return bridge (why the multiple moves)

Expected return ≈ EPS growth + multiple drift + shareholder yield.

• EPS growth: consensus +22.3% (FY27E, $1.99 → $2.433) and +23.4% (FY28E, → $3.003). This is the dominant and most reliable term.
• Multiple drift: our base case assumes modest COMPRESSION — we apply 30x to FY28E against the 32.9x the market pays on FY27E today. The rationale: the growth curve itself decelerates (from +22.5% to +15.0% to +12.5% on consensus), and multiples follow growth rates down. We explicitly do not count the mechanical roll from 32.9x FY27E to 26.7x FY28E at a constant price as compression — that is arithmetic, not re-rating. What we are assuming is that the market pays 30x at the horizon versus 33x today, a ~9% de-rating, offset by ~23% earnings growth.
• Shareholder yield: approximately zero, and possibly negative. No dividend. FY26 showed a net stock issuance of +$56.0M rather than a buyback, after a $374.0M repurchase in FY25. There is no cash-return cushion here.

The bridge reads: ~23% earnings growth, less ~9% multiple compression, plus ~0% yield ≈ +12-14% over the horizon. That matches our base fair value of $90, and it is honest about where the return comes from — all of it is earnings growth. If earnings growth disappoints, there is nothing else in the bridge.

Scenario bands (stated arithmetic, no DCF):

• Bear ~$54 — 22x FY27E non-GAAP EPS of $2.433. The Q1 acceleration proves to be a one-quarter comparison effect, growth reverts to the mid-teens, and the market applies a mature-hardware multiple. Roughly the 52-week low. -33% from spot.
• Base ~$90 — 30x FY28E non-GAAP EPS of $3.003. Growth lands between the Q1 rate and the consensus curve, gross margin holds at 70%, the multiple compresses modestly with the growth rate. +12% from spot.
• Bull ~$127 — 35x FY29E non-GAAP EPS of $3.619. The AI-infrastructure storage pull proves durable, revenue growth holds above 20% through FY29, and the market re-rates the name toward its AI-adjacent peers. +58% from spot. Note the FY29 estimate rests on one analyst; this scenario is the least well-supported of the three and we say so.

Base sits 12% above spot and 5% below the $94.38 street consensus. We are marginally more cautious than a sell-side that is 33-to-1 positive.

7c. Variant perception (where we differ, what would surprise)

• We differ from the street on the insider tape, and we appear to be alone in weighting it. Thirty-four analysts rate this a buy or strong buy against a single sell, with a $94.38 target — and in the four weeks before this dive the chief executive and the founder-Chief Visionary Officer sold a combined ~$13.1M of stock at $79-82. That is the sharpest divergence between the published view and the revealed preference of the people with the most information. We do not treat it as a thesis-breaker — executives sell for many reasons, and the sales were programmatic in appearance (multiple tranches across two days at incremental prices) — but we treat it as sufficient reason to stage an entry rather than complete one. Section 8 develops this.
• We differ from the street modestly on the durability of the inflection. Consensus models +22.5% for FY27 against the +35.2% just delivered, so the street is already discounting it; our base case sits below the consensus target because we discount it slightly further. We do not claim a variant perception on the business — our revenue and margin path is consensus-derived. Our edge here, such as it is, is on position sizing and entry, not on forecasting.
• Positive surprise that would force a repricing: the 2026-08-26 print delivering revenue at or above $1.16B (+35%) — above the +28.0% consensus — with gross margin holding at 70% and deferred revenue growth accelerating above 25%. That would convert one quarter into a trend and would justify the bull-case multiple directly. Watchable, dated, binary.
• Negative surprise that would force a repricing: revenue growth on 2026-08-26 below ~+20%; or any decline in deferred revenue, which would indicate the forward book is thinning; or gross margin below 68%, which would signal price competition from the larger peers; or further large insider sales at or below current prices.

8. Insider activity — the single most important qualitative signal

The Form 4 record in this file is eight consecutive sale transactions across two days, with zero purchases:

Totals: Giancarlo 65,800 shares for ~$5.27M; Colgrove 98,053 shares for ~$7.82M. Combined ~163,853 shares, ~$13.09M, in two days, at $79.46-$82.14 — against today's price of $80.14.

How to read it, fairly and in both directions.

Arguments that this is benign: the transactions have the fingerprint of a pre-arranged trading plan — Colgrove's sales are broken into multiple tranches at incrementally rising prices across a single session, which is what algorithmic plan execution looks like rather than a discretionary decision to exit. Both men remain very large holders: Colgrove retains 2,565,000 shares (~$206M at spot) after selling 3.7% of his position, and Giancarlo retains 1,749,597 shares (~$140M) after selling 3.6%. Neither is leaving. Founders and long-tenured chief executives diversify.

Arguments that it matters: it is $13.1M in 48 hours, by the two people who best understand what the FY27 Q2 quarter looks like, executed roughly three weeks after the quarter began and roughly six weeks before it is reported, at a price essentially identical to today's. And there are no purchases anywhere in the record — not by these two, not by any other officer or director. In a file where the sell-side is 34-to-1 positive and the stock has just posted its best growth quarter in years, the complete absence of insider buying is itself information.

Our read: not a thesis-breaker, but a decisive argument for staging. We size the first tranche below target and hold capital back through the 26 August print. If the quarter confirms the inflection and the insiders stop selling, we complete the position. If the insiders keep selling into strength, the position stays small regardless of what the quarter says. This is exactly the situation position-sizing discipline exists for.

9. Moat and competitive position

The competitive position is a product-architecture and software moat rather than a scale moat. This company is materially smaller than several of its assigned peers — HPE at a $66.5B market cap, NetApp at $35.9B, Sandisk, Super Micro — and cannot win on purchasing power. What the financials say it wins on instead:

• 70.4% gross margin in a hardware-adjacent business, versus what commodity storage economics would produce. That is a software-and-architecture margin.
• 26.3% of revenue in R&D — a scaled, self-funded innovation base, and among the highest R&D intensities in this batch.
• A Service line at 46.2% of revenue and growing 129% over four years, converting a hardware sale into a recurring relationship. This is the moat being built, visibly, in the disclosure.
• $1.214B of deferred revenue growing 23.2% — customers committing cash ahead of delivery.

Where the moat is thin: there is no customer-concentration disclosure, so we cannot assess dependence on a handful of hyperscale buyers; the AI demand pull that is driving the acceleration is available to every vendor in the peer set, not just this one; and the FY24 result (+2.8% revenue growth, operating income down) demonstrates that when enterprise budgets pause, this company's moat does not prevent a flat year.

Assessment: a real but narrow moat — architecture and recurring service, inside a market where much larger competitors are chasing the same AI-driven demand. It supports a premium multiple; it does not support an unlimited one.

10. Verdict, kill-criteria and flip conditions

Stage-In. The operating case is documented and dated: revenue growth accelerated through four consecutive quarters to +35.2%, gross margin is 70.4%, R&D reinvestment is 26.3% of revenue, deferred revenue is up 23.2%, free cash flow is $615.7M, the balance sheet holds $1.331B of net cash, international revenue is compounding at more than twice the domestic rate, and the panel's highest-skill-weighted source has placed the business inside two named structural themes. Our base fair value of $90 (30x FY28E) is 12% above spot, and the sell-side is 33-to-1 positive with a $94.38 target.

Three specific reasons the position is staged rather than completed: (1) the inflection rests on one quarter, and consensus itself models deceleration to +28.0% for the quarter reporting on 26 August; (2) GAAP operating margin is 3.1% and the trailing GAAP P/E is 118x — the attractive multiples are all non-GAAP and exclude a charge equal to 13.2% of revenue; (3) the CEO and the co-founder sold ~$13.1M in two days last month at today's price, with no purchases anywhere in the record.

Staged entry — build a 2-3% position in three tranches:

1. Tranche 1 — now (~1/3 of target), at ~$80. Above a rising 50-DMA and a rising 200-DMA, a documented four-quarter acceleration, net cash, and 12% to base fair value. You are buying the trend, not the print.

2. Tranche 2 — the reaction to the 2026-08-26 print. Add on revenue at or above +28% with gross margin at 70% and deferred revenue still growing above 20%. If the quarter comes in below +20%, do not add — reassess against the kill-criteria below.

3. Tranche 3 — either a pullback to the 200-DMA (~$74) that holds, or a confirmed second quarter of 30%+ growth. Whichever arrives first.

Pre-registered KILL criteria (any one converts Stage-In → Avoid, and stops further tranches):

• Revenue growth below +20% on 2026-08-26 — the inflection was a comparison artefact and the multiple has no support.
• Deferred revenue declining in any quarter — the forward book thinning is the earliest possible warning and it precedes revenue.
• Gross margin below 68% — the four-year floor; a break signals price competition from the larger peers.
• Stock compensation rising above 14% of revenue, which would mean the gap between the non-GAAP multiple everyone quotes and economic reality is widening rather than closing.
• Continued large insider selling below $80, particularly by the chief executive — selling into strength is one thing; selling into weakness is another.
• The 200-DMA (~$74) decisively lost on company-specific news rather than market beta.

Pre-registered FLIP TO HIGH-CONVICTION (upsize toward the top of the 2-3% range):

• 2026-08-26 revenue at or above $1.16B (+35%) with gross margin at 70% and deferred revenue growth above 25% — one quarter becomes a trend.
• Any insider PURCHASE, which would resolve the single largest qualitative concern in this dive at a stroke.
• A third tracked source naming this business, broadening a conviction lane that currently rests on one voice.

Where P fits in the Synthos Framework Portfolio. The AI-infrastructure sleeve, storage and data-movement layer — a distinct tier from compute, networking and optical, and complementary rather than redundant to them. Suggested weight 2-3%, built in three tranches, sized below a core position because the KB lane rests on a single source and the insider tape is unfavourable. Logged as a tracked Synthos call (Stage-In) as of 2026-08-04 at $80.14.

Single biggest risk: that the +35.2% quarter does not repeat. Everything in the bull case — the multiple, the base fair value, the sell-side's 33 buy ratings — is downstream of one observation. This company grew 2.8% in FY24, three years ago.

Most fragile assumption in the price: that the market continues to pay ~30x forward non-GAAP earnings for a business whose GAAP operating margin is 3.1%, whose stock compensation is 13.2% of revenue, and whose free-cash-flow yield is 2.3%. The valuation depends entirely on investors continuing to look past a $482M annual expense. They have so far. That is a market convention, not a law.

Provenance & disclosures

• TICKER IDENTITY — see section 0. The vendor profile identifies the issuer behind symbol P as Everpure, Inc. (NYSE, CIK 0001474432, CEO Charles H. Giancarlo, Santa Clara CA, founded October 2009 by John M. Hayes and John Colgrove), an enterprise data-storage company. Two profile metadata fields are stale or wrong and are not used: the sector/industry tag ("Industrials / Industrial - Machinery", contradicted by the description, peer set, 70.4% gross margin and 26.3% R&D intensity) and the ipoDate (2011-06-15, irreconcilable with a 2009 founding and a financial history beginning in FY2021 — the signature of a recycled single-letter symbol). The non-equity tripwire test is NEGATIVE: this is a common equity, not a bond, preferred or trust security. For the editor: the two KB claims used here are tagged to the entity symbol PSTG, not P; the symbol-to-entity mapping should be confirmed before this dive is cross-linked to any portfolio or scorecard record.
• Traceability: 2 tagged KB claims (breadth 2, net conviction positive-moderate), both from jordi_visser (skill 2.0) — 2026-01-18 (conviction 58, the data-movement/storage bottleneck theme) and 2026-06-17 (conviction 65, AI-factory infrastructure alongside Dell and HPE). One further claim (2026-06-04, conviction 58) is EXCLUDED because it is tagged to the quarantine_misattributed source, per house policy. One theme-adjacent claim is cited and labelled as adjacent: gavin_baker, 2024-02-01 (conviction 72) on flash displacing all hard drives.
• Data as-of: fundamentals 2026-05-03 (FY27 Q1; fiscal year ends late January/early February) · estimates 2026-08-04 · prices 2026-08-04 ($80.14, 50-DMA $75.51, 200-DMA $74.22, RSI 54.8) · KB claims 2026-08-04. Sole market-data source: the Synthos vendor file for this ticker.
• Data REJECTED: (1) the FY27 Q1 share-count and interest-expense fields — weightedAverageShsOut, weightedAverageShsOutDil (both 326,539,000) and interestExpense ($1,807,000) are exactly identical to the year-ago quarter, and basic equals diluted where every other quarter shows a 3-5% gap. No per-share arithmetic in this dive uses that row; revenue and net income from it are used because they reconcile independently against the earnings calendar. (2) The consensus EBITDA series — FY27 estimated EBITDA of $930.1M against estimated net income of $842.0M implies only $88M of combined D&A, interest and tax for a company with $264M of annual capex. No forward EV/EBITDA is computed. This EBITDA failure appears across several files in this batch and looks systematic rather than company-specific.
• Data gaps explicitly noted: no product-family, customer-concentration, or AI-specific revenue disclosure — the AI demand driver is inferred from growth rate and timing, not read off a line item; the FY29+ estimate rows rest on a single EPS analyst; the profile and technical blocks report different 52-week ranges ($54.37-$100.59 versus $54.51-$98.70); ratios_ttm.returnOnEquityTTM is null (the key-metrics figure of 16.13% is used).
• Vendor figures recomputed by hand and labelled: net cash of $1.331B (cash $854.9M + short-term investments $692.4M − total debt $216.1M); the vendor's netDebt of -$638.7M nets cash only. Enterprise value of $25.31B; the vendor reports $26.032B on the same cash-only basis. All EV multiples here use $25.31B.
• Non-GAAP caveat: every forward P/E in this dive (32.9x FY27E, 26.7x FY28E) uses consensus non-GAAP EPS, which excludes stock compensation of $481.7M — 13.2% of revenue and 4.2 times GAAP operating income. Trailing GAAP P/E is 118x. This is disclosed prominently rather than buried, because it is the largest single difference between the headline valuation and the economic one.
• Fair-value caveat: the $54 / $90 / $127 anchors are scenario multiples applied to consensus non-GAAP EPS, not a discounted cash-flow model. The base case assumes modest multiple compression (30x at the horizon versus 32.9x FY27E today) in line with a decelerating consensus growth curve. The bull case rests on an FY29 estimate carried by a single analyst and is the weakest-supported of the three.
• Not investment advice. Independent research, educational and informational only, never personalised.
• Version: 2026-08-04-full.