Vertiv Holdings VRT

Reference table

1. What the business is

Vertiv Holdings Co is a Delaware corporation headquartered in Westerville, Ohio, NYSE-listed since 2018, with approximately 34,000 employees, led by Chief Executive Giordano Albertazzi. It designs, manufactures and services the physical infrastructure of data centres and communications networks: power distribution, uninterruptible power supplies, thermal management (air and increasingly liquid), racks and integrated modular systems, plus a service and spares business.

The revenue split, from seg_prod — and it reconciles exactly:

The service line is the annuity underneath the equipment, and it grew $148.7 million in the June quarter and $282.1 million in the half. The archive does not disclose its margin, which is the most useful undisclosed number in this file — service revenue in an installed-base business is normally materially higher-margin than equipment, and 18.0% of revenue at a premium margin would be a meaningful share of profit.

What management says is changing, quoted because the specifics matter:

> "We continue to see very robust growth in demand for data centers supporting artificial intelligence ('AI') and high-performance compute applications and have strategically invested in expanding our global capacity in response to current and anticipated customer demand across key infrastructure segments. For the quarter, our capital investments were significantly higher than the spend in the same quarter of 2025."

> "Customer requirements are evolving toward higher-density, more complex infrastructure environments, including, but not limited to, hybrid air and liquid cooling architectures, converged physical infrastructure systems, and high voltage direct current power architectures."

> "Customers are increasingly prioritizing speed of deployment, scalability and execution certainty. We continue to invest in prefabricated, modular and factory-integrated solutions designed to reduce on-site complexity and accelerate time-to-deployment."

Half-weighted as management framing — but the three architecture changes named in the middle quotation are the substance of the exponential case and are worth separating out. Hybrid air-and-liquid cooling, converged physical infrastructure, and high-voltage direct-current power are each a change of specification rather than a change of quantity. Each obsoletes part of an installed base, each raises the equipment content per megawatt, and each favours a supplier that can sell the whole stack and then service it. None of them is quantified anywhere in this archive.

2. The June quarter — an earnings beat and a revenue miss

All figures from the 10-Q filed 2026-07-29, which is prose-only (see Section 7) but carries the MD&A numbers in full ($M):

The organic disclosure is the best thing in this filing and it deserves to be stated as a headline. Vertiv separates acquisition revenue and currency in plain language: $636.2 million of growth, of which $129.7 million was acquired and $35.9 million was currency, leaves $470.6 million on a base of $2,638.1 million — 17.8% to 18.2% depending on the base convention. Very few companies in this programme give the reader that split without being asked, and PWR, covered in this same batch, does not.

Four observations.

First, the margin expansion is real and its cause is disclosed. Gross margin rose 370 basis points to 37.7% in the quarter, attributed to "the mix of product and service sales", and over six months to "the mix of product and service sales in addition to operational leverage and improved project execution." Meanwhile SG&A grew only 10 basis points as a share of sales. That combination — gross margin up 370 basis points with operating expense flat as a percentage — is what produces a 48.6% increase in Americas operating profit on a 29.2% revenue increase.

Second, the "other operating expense" line grew $44.0 million and its composition is unflattering. The increase was "primarily due to a $26.8 increase in amortization of intangibles and a $21.4 increase in other operating expense (income), primarily due to the contingent consideration related to our recent acquisitions." In the corporate line, this is explicit: total corporate costs of $79.6 million in the quarter, up $20.9 million, "primarily due to the $28.8 loss on the change in fair value of contingent consideration associated with the PurgeRite acquisition", and $173.0 million in the half including $62.0 million of the same. A contingent-consideration charge means the acquired business is beating its earn-out targets — good news arriving as a cost — but it is a real charge and it is growing.

Third, interest expense collapsed and the reason is the recapitalisation. Net interest was $13.0 million in the first half against $46.6 million, driven by the interest-rate swap settlement and $13.7 million of incremental interest income on a much larger cash balance. Vertiv now earns nearly as much on its cash as it pays on its debt.

Fourth — and this is the finding that governs the verdict — the quarter missed on revenue.

Management's explanation, verbatim: the sales increase was driven by volume, acquisitions and currency, "which were slightly offset by temporary supply chain congestion and multi-phased project execution." The same phrase appears in the Americas segment discussion.

Why this matters more here than it would elsewhere. Vertiv is valued at 40.4 times current-year earnings on the strength of a volume ramp. An earnings beat driven by margin, alongside a revenue shortfall driven by delivery, is precisely the wrong composition for that valuation — it says the demand may be there but the company could not ship it, and for a business whose consensus requires a 36% second-half sequential revenue step, shipping is the whole question. The stock's 18.4% three-month decline and its 28.3% distance from the 52-week high are the market saying so.

The segments, and one is shrinking

EMEA is the problem and it should not be smoothed over. First-half net sales fell 8.4%, with the filing stating the decline was "due to softer market demands from prior periods, specifically in the first quarter of 2026", and product sales down $81.0 million over six months. The quarter's 1.7% increase was entirely currency: foreign exchange contributed approximately $13.7 million against a total increase of $8.0 million — so EMEA net sales in constant currency FELL in the June quarter too. EMEA is 13.6% of first-half net sales excluding intercompany, so the dollar impact is contained; the signal is that the AI infrastructure buildout the whole thesis rests on is, at Vertiv, an Americas and Asia-Pacific phenomenon and not yet a European one.

3. The recapitalisation, the balance sheet, and the cash

On 3 March 2026 Vertiv changed what kind of credit it is, and the payload has none of it.

From the 10-Q: "On March 3, 2026, the Company repaid in full all outstanding indebtedness under its Term Loan Credit Agreement and refinanced and replaced the ABL Revolving Credit Facility. Upon such repayment, all commitments under the Term Loan Credit Agreement and ABL Revolving Credit Facility were terminated and all guarantees and liens securing obligations under the Term Loan Credit Agreement and the ABL Revolving Credit Facility were released." The company recognised a loss on extinguishment of debt for the six months (the amount falls in a stripped numeric field — see Section 7), and "in connection with the repayment of the Term Loan Credit Agreement, the hedged forecasted transactions became probable of not occurring, and therefore the Company settled all outstanding interest rate swaps."

The new capital structure, at 2026-06-30:

Against, at 2026-06-30: cash and equivalents $2,810.6M plus short-term investments $300.0M = $3,110.6M.

Vertiv is approximately NET CASH — $3,110.6 million of liquid assets against $2,950.0 million of principal debt — with $2,483.6 million of undrawn unsecured revolver behind it. A company that eighteen months ago funded itself with a secured term loan and an asset-based revolver now issues forty-year unsecured paper at 5.950%. That is a credit-quality transition, and it is the single most under-appreciated fact in this file.

A note on the 2066 notes specifically. A forty-year unsecured bond is an instrument sold to insurance companies and pension funds, and it is not available to a business the market considers cyclical. Vertiv sold $500 million of it. (For contrast in this same batch: Quanta Services, also an AI-infrastructure name, priced ten-year money at 5.550% on 3 August.)

The cash statement, from the 10-Q prose ($M):

Operating cash flow nearly tripled — and 36% of it was working capital. The filing: "Trade working capital provided $678.8 in the first six months of 2026 compared to $95.2 utilized in the first six months of 2025 primarily driven by deferred revenue." Deferred revenue is customer prepayment. On the way up it is an excellent leading indicator — customers are paying in advance to secure capacity. It is also non-repeating: cash earnings ex-working-capital were roughly $1,187.8 million in the half, not $1,866.6 million, and a normalisation of deferred revenue would reverse the flattery. Both readings are true and both are stated.

Capital expenditure is being tripled. $288.5 million in the first half against a full-year guide of $550.0–570.0 million, "in order to support capacity expansion across the business" — against $220.0 million spent in all of FY2025. That is 2.5x, it is discretionary, and it is the clearest signal management gives about demand it expects but has not yet booked. It is also the reason free cash flow will convert less well in the second half than in the first.

One unquantified item worth recording. From the MD&A: "in February 2026, the Supreme Court issued a decision invalidating tariffs imposed under IEEPA. U.S. Customs and Border Patrol has established a phased administrative process for submitting refund claims for certain IEEPA tariffs. We have initiated a process of submitting refund claims, however, the amount, timing and likelihood of any refund recovery remain uncertain." Unquantified upside; no number is assumed here.

4. Valuation — priced in or room?

At $269.93 (market cap $103.92B, ~385.0M implied shares):

Estimate coverage is strong on the near years and thin beyond. 14 analysts on FY2026 EPS, 16 on FY2027 — the anchor — 10 on FY2028, and only 4 each on FY2029 and FY2030, which are excluded from every conclusion. The FY2027 EPS range is $8.477 to $9.788, a 15.5% spread.

The core arithmetic, stated plainly: 29.7 times FY2027 consensus for 36.1% expected FY2027 earnings growth is a PEG of 0.82. That is the most attractive growth-adjusted multiple in this batch by a wide margin, and it is why this dive spends as long as it does on the reasons not to act on it.

est.ebitdaAvg and est.ebitAvg are REJECTED on two grounds — fixed ratios and arithmetic impossibility. ebitdaAvg is exactly 14.24% of revenueAvg and ebitAvg exactly 10.19% in every forward year FY2026 through FY2030. And netIncomeAvg EXCEEDS ebitAvg in every single forward year — FY2026 $2,638.2M against $1,425.1M, FY2027 $3,561.0M against $1,843.6M, FY2030 $6,222.3M against $2,872.1M — which is impossible. The implied 10.19% EBIT margin also contradicts the company's own reported figures: June-quarter operating profit across the three segments less corporate was $711.6 million on $3,274.3 million of sales, 21.7%. All forward valuation uses epsAvg.

Peer context is unusable. The vendor peer set — CTAS, EMR, ITW, JCI, NOC, PWR, TDG, TRI, UPS and WM — contains a uniform-services company, a waste company, a parcel carrier, a defence prime, a news-and-data business and four industrials. Emerson Electric and Johnson Controls are arguably adjacent; nothing here is a data-centre power and thermal specialist. The payload supplies only market capitalisations and no multiples, so no peer-multiple comparison is drawn. (PWR is separately covered in this same batch.)

4a. What today's price assumes (the inversion)

At $269.93 — 40.4x FY2026 consensus, 29.7x FY2027, 21.8x book — the price embeds:

• Revenue reaches $13,991 million in 2026 and $18,101 million in 2027. (Consensus; 19 and 21 analysts.) First-half revenue was $5,923.8 million, so FY2026 requires approximately $8,068 million in the second half — a 36.2% sequential step. This is the most fragile assumption in the price, and it is fragile for a specific, dated reason rather than a general one: the June quarter already missed by 3.1% on "temporary supply chain congestion and multi-phased project execution", and the third-quarter consensus of $3,751 million itself requires a 14.6% sequential increase.
• Adjusted EPS reaches $6.680 in 2026 and $9.091 in 2027. (Consensus; 14 and 16 analysts.) The earnings side has been the reliable half — five of the last five quarters beat, the two most recent by 6.3% and 25.9% — because margin has been the driver. It is the revenue side that has cracked.
• Gross margin holds near 37.7%. (Filing-sourced; our derivation of the requirement.) The 370-basis-point expansion came from "the mix of product and service sales" plus operational leverage. Acquisition mix, the capacity-expansion depreciation from a $550-570 million capital programme, and any pricing response to competition all work against it.
• The market keeps paying roughly 27-34x forward earnings. (Our number.) At 24x FY2027E the stock is $218; at 36x it is $327. On a 2.029 beta with a 209% twelve-month range, the realised band will be wider than that.
• EMEA stops shrinking. (Filing-sourced.) First-half net sales down 8.4% with product sales down $81.0 million. Consensus revenue growth of 29.4% for FY2027 is not consistent with a seventh of the business contracting.

4b. The return bridge (why the multiple moves)

Expected return over the next twelve months decomposes as: adjusted EPS growth (+36.1%, from FY2026E $6.680 to FY2027E $9.091) + multiple drift (COMPRESSION of roughly 16%, from today's 40.4x on the current year toward about 34x on the then-current year) + shareholder yield (+0.08%) ≈ +15%.

Our base assumes meaningful multiple compression and the reason is the growth path itself. Consensus has revenue growth decelerating from 36.8% to 29.4% to 22.0% and EPS growth from 61.7% to 36.1% to 27.1%. A 40x multiple is a 60%-growth multiple; a 30x multiple is a 30%-growth multiple; and the transition from one to the other happens whether or not the company executes. The base case therefore lets a very large earnings increase carry the return and gives back a substantial part of the multiple — which is why a +36% earnings year produces a +15% price outcome.

There is no shareholder yield here to speak of: the dividend is $0.225 a share, 0.083%, and the payload shows no buyback in FY2025. Every dollar of return has to come from earnings and the multiple.

The sensitivity: at 24x FY2027E the price is $218 (−19.2%); at 36x, $327 (+21.1%); at 40x — today's multiple on the current year — $364 (+34.7%), which is approximately the street's target.

4c. Variant perception (where we differ, what would surprise)

• We DO have a variant perception and it is NEGATIVE relative to a near-unanimous street. Our base of $310 is 13.3% BELOW the consensus target of $357.47 and 12.7% below the median of $355, against 19 buy ratings, 1 hold and 0 sells with the lowest published target ($277) ABOVE the market price. When twenty analysts agree and none of them will publish a target below spot on a stock that has just fallen 28% from its high and missed on revenue, the disagreement is worth stating explicitly: we think the second-half revenue ramp is a genuine open question and the published targets do not reflect that it is.
• We think the market — and the sell side — is treating the revenue miss as noise and it may not be. Watchable number: third-quarter revenue against the $3,751 million consensus, which requires a 14.6% sequential increase from a quarter that already fell short. A second consecutive revenue miss would break the FY2026 consensus and the multiple simultaneously; a clean beat would resolve the single largest uncertainty in this dive and is the named upgrade trigger in Section 8.
• We think EMEA is being ignored. Down 8.4% in the half, with the quarter's 1.7% gain entirely currency. Watchable number: EMEA constant-currency net sales in the October print. If the AI infrastructure cycle is genuinely global, a seventh of Vertiv's business should not be contracting in the second year of it.
• We are more constructive than the multiple implies on the balance sheet, and it is under-discussed. A company that in eighteen months went from a secured term loan and an asset-based revolver to approximately net cash, a $2.5 billion unsecured revolver and forty-year unsecured notes has changed credit category. That is worth something to the equity — it removes financing risk from a cyclical capital-goods business at exactly the moment it is tripling capital expenditure — and neither the multiple nor the analyst commentary in this file reflects it.
• Positive surprise that would force a re-rate: third-quarter revenue at or above $3,751 million with the supply-chain congestion described as resolved; EMEA returning to constant-currency growth; or a disclosed liquid-cooling revenue line, which the company has bought into repeatedly (PurgeRite and "solution providers in liquid cooling technologies, heat-exchange and heat rejection technologies") and has never quantified.
• Negative surprise that would break the thesis: a second consecutive revenue miss; gross margin below 36%; EMEA declining further; a large negative revision to the FY2026 revenue consensus, which currently requires a 36% second-half step; or evidence that the deferred-revenue inflow of $678.8 million is reversing, which would take operating cash flow down sharply while the capital programme is running at $550-570 million.

Synthos fair values

All three anchors are multiples of the FY2027 consensus adjusted EPS distribution (mean $9.091, low $8.477, high $9.788, 16 analysts), cross-checked against FY2028E.

• Bear ~$180 — 21.2x the FY2027 consensus LOW of $8.477, and 15.6x FY2028E. Cross-check: 47.8% above the 52-week low of $121.82; 27.2% below the 200-day average of $247.20; 35.0% below the lowest street target. The scenario: the revenue shortfall proves structural rather than temporary, EMEA keeps shrinking, the second-half ramp does not arrive, deferred revenue normalises, and a decelerating grower de-rates toward a normal capital-goods multiple. −33.3%.
• Base ~$310 — 34.1x the FY2027 consensus MEAN of $9.091, and 26.8x FY2028E. Cross-check: 13.3% BELOW the street's $357.47; 17.6% below the 52-week high; 2.1% above the 50-day average of $303.62. Sensitivity, stated openly: 30x gives $273 — essentially spot — and 38x gives $345. The whole answer lives in a 30-38x band on one estimate row. The scenario: the congestion was temporary, the second half ramps close to consensus, margin holds near 37.7%, and the multiple compresses modestly as growth decelerates from 62% to 36%. +14.8%.
• Bull ~$420 — 42.9x the FY2027 consensus HIGH of $9.788, and 36.3x FY2028E. Cross-check: 11.6% above the 52-week high of $376.23; 16.0% below the highest street target of $500. The scenario: the second half beats, EMEA recovers, liquid cooling is disclosed and is large, and the market re-rates the pure-play supplier of the two things a data centre cannot do without. +55.6%.

Base is 14.8% above spot; asymmetry roughly 1.67:1 to the upside (33.3% down, 55.6% up), with essentially no shareholder yield. That is the best raw arithmetic in this batch and it still does not clear the Buy bar, for the reasons in Section 8 — a 3.1% revenue miss on a volume valuation, a contracting European segment, a 2.029 beta, prose-only filings, and a knowledge base with zero name-level coverage.

5. Knowledge base — zero on the name, and a lane that fails its own test

Raw hits: 8. Case-sensitive entity hits on VRT, Vertiv or Vertiv Holdings: ZERO. Name-level claims: 0. Discarded: 8.

Not one of 51,928 distilled claims names Vertiv. For the largest independent supplier of data-centre power and thermal-management equipment, in a knowledge base saturated with artificial-intelligence-infrastructure argument, that is a striking absence and it is reported as the finding rather than papered over. The credible-thinker universe in this store discusses the AI buildout constantly and has never once named the company that supplies its cooling.

The free-text sweep on "liquid cooling", "Vertiv", "data center cooling", "thermal management" and "power distribution unit" returned 8 hits across 4 channels. The relevant ones, verbatim:

> 2025-07-20 · bullish · conviction 82 · horizon: thesis · channel: jordi_visser · skill 2.0

> "AI compute demand projected 10,000x higher than supply for the foreseeable future; funding continues as a race for obsolescence and military supremacy — own hardware, data center semis, HBM, liquid cooling, power, not software."

> 2026-05-11 · bullish · conviction 80 · horizon: thesis · speaker: Jordi Visser · channel: jordi_visser · skill 2.0

> "The physical world was underinvested for this moment; agentic inference buildout needs power, liquid cooling, chemicals and optical fiber — all early-inning physical plays."

> 2026-05-24 · neutral · conviction 72 · horizon: thesis · speaker_role: independent · channel: jordi_visser · skill 2.0

> "AI is a physical-world CapEx cycle (HBM, racks, liquid cooling, copper, fiber, substations), only ~12-18% spent; bottlenecks make the buildout lumpier, more inflationary and more volatile than the market models."

> TEST (registered with the claim): "AI buildout produces cost inflation, delays and revenue-recognition problems over the next year."

That third claim is the most useful thing the knowledge base contributes to this dive, and it is worth being precise about why. It was made on 24 May 2026 with an explicit, falsifiable test attached: that the buildout would produce "delays and revenue-recognition problems over the next year." Ten weeks later, on 29 July 2026, Vertiv reported net sales 3.1% below consensus and attributed the shortfall to "temporary supply chain congestion and multi-phased project execution." That is a delay and a project-phasing problem, in the revenue line, at a company squarely inside the category the claim names. The test appears to have come true, inside the security being valued, and it is the reason the short-horizon confidence in this dive is Low. The claim names no company and carries no conviction weight; it is the most important discarded claim in this batch.

> 2024-06-21 · neutral · conviction 68 · horizon: fact · speaker: Brian Venturo · speaker_role: MANAGEMENT · channel: odd_lots · skill 1.0

> "Liquid cooling cuts electricity utilization ~60-70%, not the 30-40% people assume, because dense AI servers no longer need power-hungry internal fans; new builds combine liquid and air cooling."

Half-weighted as a management voice. It is a technical claim about a product category rather than about the speaker's own financials, and it quantifies why the liquid-cooling transition is not optional for dense AI racks — which is the demand case for a large part of Vertiv's portfolio.

> 2024-12-12 · bullish · conviction 85 · horizon: thesis · channel: no_priors · skill 1.0

> "Synthetic-diamond cooling attacks heat at the chip/material level and layers on top of any air/liquid cooling, opening extra performance and thermal margin."

This one runs the other way and is reported as such. A thermal solution "at the chip/material level" moves value toward the semiconductor and away from the facility. It is described as layering on top of rather than replacing liquid cooling, so it is a partial rather than total substitution risk — but it is the only technology-displacement claim in the lane and it is bullish at conviction 85 on the displacing technology.

CONCENTRATION SENSITIVITY TEST — REQUIRED AND RUN.

jordi_visser supplies 4 of 8 hits, 50% of the lane — far above the one-third threshold that mandates the test. Stripping that channel entirely leaves four claims:

The result of the test, stated honestly: stripping the dominant channel leaves two management voices for other companies, one half-weighted technical fact, and one substitution threat. The bullish reading of this lane depends almost entirely on one thinker — and that thinker's most recent and most specific claim is a warning about delays and revenue-recognition problems, not an endorsement. This is the cleanest illustration in this batch of why the concentration test exists.

Attribution note. speaker_role fired on two claims (odd_lots "management", jordi_visser "independent") and failed on the jensen_huang claims, which are management-channel content with a null role — consistent with the store's confirmed unreliability of that flag; management voice was judged by content. One jensen_huang claim carries the speaker slug jensen_huang_ai, which is a synthetic-speaker identifier and is noted for the record. The remaining claims carry speaker: null and are channel-attributed.

Conclusion. Breadth 0, claim count 0, net conviction none. The Synthos knowledge base has no view on Vertiv Holdings, and the adjacent thematic lane fails its concentration test. What it does supply is one dated, falsifiable, independent warning about exactly the operational failure this company reported six days before this dive — which is worth more than a dozen supportive claims would have been.

6. Technicals and insiders

• Price $269.93. −28.3% from the 52-week high of $376.23; +121.6% above the 52-week low of $121.82. The 52-week range spans 209% of the low — by far the widest in this batch. Position within the annual range: 58th percentile.
• Below the short-run average and only modestly above the long. 11.1% BELOW a 50-day moving average of $303.62; +9.2% above a 200-day average of $247.20. The 50-day has rolled over.
• RSI 39.4 — weak, approaching but not at oversold. MACD −15.97, the most negative reading in this batch by a wide margin.
• Maximum drawdown from peak over the trailing year: −28.25%, identical to pct_from_hi — the ADP-class signature — but CHECKED and consistent: ret_12m of +90.6% implies approximately $141.6 twelve months ago against lo52 of $121.82, so the low preceded the high and the current price is the low point since the peak. Verified, not flagged.
• Relative performance is a story of two regimes. 3-month −18.4% against SPY +7.6% and QQQ +7.7% — a 26-point underperformance; 6-month +42.1% against SPY +11.1%; 12-month +90.6% against SPY +24.3% and QQQ +30.8%. The stock roughly tripled off its low, peaked, and has given back nearly a third. The three-month leg coincides with the revenue miss.
• Sentiment: 19 buy, 1 hold, 0 sell across 20 analysts — the most one-sided coverage in this batch; consensus target $357.47 (+32.4%), median $355, high $500 (+85.2%), low $277 — 2.6% ABOVE spot. Not one covering analyst has a target below the market price on a stock that has fallen 28% from its high and missed on revenue. We treat that unanimity as a caution rather than a comfort, and our base sits 13.3% below it.

Today's move. VRT closed 2026-08-04 at $269.93, up 2.62% or $6.88 from $263.05. It opened at $275.21, traded $268.12 to $276.52, and closed 2.4% below the day's high, in the bottom quarter of the range, on 5.63 million shares — roughly $1.52 billion of turnover, the heaviest in this batch. An up day that closes near the low on heavy volume is a distribution pattern, and it is recorded without being over-read.

Insiders — eight transactions, all dividend-equivalent accruals

The reading: nothing whatsoever, and the fractional share counts prove it. Eight A-Award transactions on the same date for fractional amounts between 0.38 and 9.53 shares at a zero price. These are dividend-equivalent accruals on unvested awards — the mechanical crediting of a $0.0375 quarterly dividend onto restricted stock — and they carry no information of any kind. The chief executive received 9.53 shares.

What the file does not contain: a single open-market purchase, a single open-market sale, and any transaction by the chief financial officer. On a stock that has fallen 28% from its high, the absence of insider buying is the only thing here worth a sentence, and it is a weak signal. (Separately, the 10-Q discloses a Rule 10b5-1 trading arrangement in the name of a "Mr. Monser" — a director — whose terms fall outside the extracted text.)

7. Data integrity — what we rejected and why

Six findings. Vertiv's file has one corrupt record, one systematic estimate fabrication, and — the most consequential issue — an entire filing archive with NO statement tables, which lowers confidence in this dive and is disclosed as such.

1. EVERY FILING IN THIS ARCHIVE IS PROSE-ONLY — no preserved [TABLE] statement data anywhere. All six documents (10-K filed 2026-02-13, 10-Qs filed 2026-04-22 and 2026-07-29, and 8-Ks of 2026-06-12, 2026-06-18 and 2026-07-29) carry tables: false in the manifest, and a direct check of the June 10-Q returns zero [TABLE] markers. There is no balance sheet, no income statement, no cash-flow statement and no segment table readable from the filings. Compounding this, the extraction has STRIPPED numeric values in several narrative passages — for example the acquisition note reads "net of cash acquired of $" with the figure absent, and the loss on extinguishment of debt is "$" followed by nothing. What survives, and is used throughout, is the MD&A prose, which fortunately carries the income-statement, segment, liquidity and cash-flow figures in full sentences. Confidence is explicitly lowered on any figure not present in that prose, and the three 8-Ks are stubs of 1.2–1.8 KB whose exhibits are not in the archive at all — so the second-quarter earnings release itself is unavailable and no guidance figure of any kind can be sourced.

2. earn_cal contains a CORRUPT RECORD: the Q1 2026 row has been overwritten with Q2 2026 data. The 2026-04-22 entry reports epsActual: 1.52, revenueActual: 3,274,300,000 and revenueEstimated: 3,379,044,000 — identical in all three fields to the 2026-07-29 entry — with lastUpdated: 2026-07-29. inc_q reports Q1 2026 revenue of $2,649.5M and the 10-Q confirms first-half net sales of $5,923.8M against a June quarter of $3,274.3M, so Q1 was $2,649.5M. The Q1 2026 adjusted EPS actual is therefore NOT RECOVERABLE from this file, and no trailing-twelve-month adjusted EPS is computed anywhere in this dive. All trailing multiples shown are on GAAP. This is the "corrupt individual record" defect class.

3. est.netIncomeAvg EXCEEDS est.ebitAvg in every forward year — arithmetically impossible — and both EBITDA and EBIT carry the fixed-ratio fabrication signature. REJECTED. ebitdaAvg is exactly 14.24% of revenueAvg and ebitAvg exactly 10.19% in FY2026 through FY2030. Meanwhile netIncomeAvg is $2,638.2M against ebitAvg of $1,425.1M in FY2026, $3,561.0M against $1,843.6M in FY2027, and $6,222.3M against $2,872.1M in FY2030. Net income cannot exceed EBIT. The implied margin is also wrong against the filing: June-quarter segment operating profit less corporate was $711.6M on $3,274.3M of sales, 21.7%, against the estimate block's 10.19%. All forward valuation uses epsAvg, which is internally consistent with netIncomeAvg and the share count.

4. enterpriseValueTTM implies $445.6M of net debt against bal_a's $1,675.0M — but here the EV field is the CLOSER of the two to the filing. enterpriseValueTTM of $104,365.4M less market capitalisation of $103,919.8M gives $445.6M. The bal_a block at 2025-12-31 says $1,675.0M. The 10-Q at 2026-06-30 gives cash of $2,810.6M plus short-term investments of $300.0M against principal debt of $2,950.0M — approximately net cash, or roughly −$160M. So the EV field is stale by less than the balance block is, and both are out of date. This is the same inverted-resolution pattern found on GD in this batch, and we use the filing. (Note: bal_a's totalDebt of $3,403.4M includes $245.2M of capitalLeaseObligations, a 7.8% lease-inclusion overstatement of the T/TJX class.)

5. seg_geo sums 8.3% above reported revenue and drops the US split. The FY2025 entry gives Americas $6,423.9M, EMEA $2,382.2M and Asia Pacific $2,274.7M, summing to $11,080.8M against reported FY2025 revenue of $10,229.9M — an 8.3% overage, because these are segment figures including intercompany sales while the 10-Q presents them excluding. The FY2020 entry, by contrast, carried a "UNITED STATES" line; no year since does. All segment figures in this dive come from the 10-Q's ex-intercompany presentation. (By contrast seg_prod — Product $8,390.6M plus Service $1,839.3M — sums to $10,229.9M, EXACTLY the reported FY2025 revenue. Clean.)

6. inc_a FY2022 reports eps of −$0.038 against netIncome of +$76.6M. A positive net income with a negative earnings per share is possible only through preferred dividends or a noncontrolling-interest allocation, and minorityInterest is zero in every year of the block. The row is not used and is recorded as an unexplained inconsistency in the historical series. Related: FY2024 totalOtherIncomeExpensesNet of −$613.4M against FY2025's −$155.1M is a $458M swing on a company whose interest expense fell only $64M, which the filings attribute to warrant and derivative fair-value movements not visible in this block.

Where vendor and filing AGREED — recorded:

• inc_q Q2 2026 revenue of $3,274.3M matches the 10-Q's stated net sales exactly, as does Q2 2025's $2,638.1M and the derived first-half total of $5,923.8M.
• inc_q Q2 2026 gross profit of $1,234.9M matches the 10-Q exactly, as does the 37.7% margin.
• inc_q Q2 2026 interest expense of $17.4M matches the 10-Q exactly ($17.4 net), and Q1's zero is consistent with the disclosed first-half net of $13.0M after the swap settlement.
• seg_prod sums exactly to reported revenue in every year shown.
• tech.max_dd_from_peak of −28.25%, which carries the ADP-class signature, was CHECKED and found CORRECT — ret_12m of +90.6% implies ~$141.6 a year ago against lo52 of $121.82, so the low preceded the high.
• capexToDepreciationTTM of 1.111 is consistent with FY2025 capex of $220.0M against D&A of $308.6M rising through 2026, and no COP-class understatement is present; the filing's FY2026 capex guidance of $550.0–570.0M is 2.5x the FY2025 figure and is disclosed, not inferred.
• Disclosure controls were concluded EFFECTIVE at 2026-06-30 by the chief executive and chief financial officer. The forward-looking-statement risk list retains legacy language about "failure to remediate material weaknesses in our internal controls over financial reporting", and this dive checked Item 4 specifically because of it: no current material weakness is disclosed.

Vendor composite rating — accepted, and the shape is instructive. B / 3 overall, with 5/5 on both return on equity and return on assets, 2/5 on discounted cash flow and debt-to-equity, and 1/5 on both price-to-earnings and price-to-book. That is an accurate summary: an exceptionally profitable business at a demanding price. The 2/5 debt-to-equity score is now stale — it predates the March recapitalisation that left the company approximately net cash.

Non-equity tripwire — checked and passed. VRT is Class A common stock, NYSE-listed, per the Form 4 filings. Price of $269.93 is not par-like; beta is 2.029, the highest in this batch; the dividend is a nominal $0.225 trailing, 0.083%; volume was 5.63M shares (~$1.52B of turnover); the 52-week band of $121.82 to $376.23 is a 209% range. Note the four series of Senior Notes and the Senior Secured Notes due 2028, which are NOT this security. This is common equity.

8. Verdict, kill-criteria and flip conditions

Watch — and this is the closest call in this batch, so the reasoning is set out in full.

What the arithmetic says. 29.7 times FY2027 consensus for 36.1% expected earnings growth is a PEG of 0.82 — the most attractive growth-adjusted multiple in this batch. The stock is 28.3% below its 52-week high after an 18.4% three-month decline, RSI is 39.4, and the street is 19 buy, 1 hold, 0 sell with a $357.47 target and no published target below the market price. Our base of $310 is 14.8% above spot with 1.67-to-1 asymmetry. The underlying business is doing everything right: organic growth of 18.2% in the quarter and 20.9% in the half, gross margin up 370 basis points to 37.7%, Americas operating profit up 48.6% at a 27.6% margin, first-half operating cash flow of $1,866.6 million against $626.2 million, and a balance sheet rebuilt in March into approximately net cash with $2,483.6 million of undrawn unsecured revolver and forty-year unsecured paper outstanding.

Why it is not a Buy, stated as five specific reservations.

First, and decisively: the June quarter missed on revenue. $3,274.3 million against a $3,379.0 million consensus, −3.1%, while adjusted EPS beat by 6.3%. For a security valued at 40.4 times current-year earnings on a volume ramp, a shortfall in volume offset by a beat in margin is the wrong composition. The FY2026 consensus requires approximately $8,068 million of second-half revenue against $5,924 million in the first — a 36% sequential step — from a company that just told us it could not ship what it sold.

Second, the independent voice in the adjacent lane called it. jordi_visser, 24 May 2026, skill 2.0, registered test: "AI buildout produces cost inflation, delays and revenue-recognition problems over the next year." Ten weeks later Vertiv reported a revenue miss caused by "temporary supply chain congestion and multi-phased project execution." We do not have a name-level claim on this company. We have something more useful and it is a warning.

Third, EMEA is contracting. Down 8.4% in the first half, with the quarter's 1.7% gain entirely currency, on "softer market demands." A seventh of the business shrinking in the second year of the buildout is not consistent with 29.4% consensus revenue growth in FY2027.

Fourth, the security itself is extreme. Beta 2.029, a 209% twelve-month price range, no dividend to speak of, no buyback in the payload, and 21.8 times book with tangible book of $1.75 a share. Every dollar of return has to come from earnings and multiple, and the multiple has already moved 28% in one direction this quarter.

Fifth, we cannot see the statements. Every filing in this archive is prose-only. There is no balance sheet, income statement, cash-flow statement or segment table to read, and the earnings-release exhibit is absent, so no company guidance of any kind is available. Confidence is lowered accordingly and the short-horizon confidence rating in this dive is Low for that reason as much as any other.

Pre-registered UPGRADE conditions — what would take this to Buy — Tactical. The first one is the whole question:

• A third-quarter revenue print at or above the $3,751 million consensus, with the supply-chain congestion described as resolved. That single number settles whether the June miss was a phasing artefact or the start of something, and it arrives on 2026-10-28. If it clears, this becomes a Buy at almost any price under $330 and we will say so.
• A price near $235, which is 25.8x the FY2027 consensus of $9.091 and 5% below the 200-day average of $247.20. At that level the base case is +32%, the asymmetry exceeds 2.5:1, and nothing has to change at the company. This requires a 13% decline — well inside this security's normal weekly range.
• EMEA returning to constant-currency growth.
• Gross margin holding at or above 37.5% for two consecutive quarters while capital expenditure runs at the guided $550-570 million pace.
• A disclosed liquid-cooling revenue line. Vertiv has bought PurgeRite and other "solution providers in liquid cooling technologies, heat-exchange and heat rejection technologies" and has never quantified any of it. A number would let the highest-growth part of the portfolio be valued directly.

Pre-registered KILL criteria — what would take this to Avoid:

• A second consecutive revenue miss in the October print.
• A downward revision to FY2026 revenue consensus below $13.0 billion, which would mean the second-half ramp has been abandoned.
• Gross margin below 36.0%, from 37.7%.
• EMEA net sales declining more than 10% year on year in constant currency.
• A reversal in deferred revenue taking operating cash flow materially negative against a capital programme running at $550-570 million a year.
• Evidence that chip-level thermal solutions are displacing facility-level cooling content — the substitution risk the knowledge base's no_priors claim identifies at conviction 85.

Where VRT fits in the Synthos Framework Portfolio. No position today, and a dated review. It goes on the watch list with two triggers, either of which qualifies: a price near $235, or a clean third-quarter revenue print on 2026-10-28 — for a 1.5% initial sleeve in the AI-infrastructure bucket, scaling to 3% on the second confirming quarter. Sizing note: this is a 2.029-beta security with a 209% twelve-month range, and the correct discipline is a small position taken on evidence rather than a full position taken on a multiple. The 28% drawdown is not, by itself, the evidence — the revenue miss that caused it is unresolved. Logged as a tracked Synthos call (Watch) as of 2026-08-04 at $269.93, with the fair-value anchors, the kill criteria and both upgrade triggers gradeable.

Single biggest risk: that the revenue miss was not temporary. Vertiv's entire valuation is a volume valuation. Consensus needs revenue to compound from $10,230 million in FY2025 to $22,081 million in FY2028, and the near-term path requires roughly $8,068 million in the second half of 2026 against $5,924 million in the first. The June quarter missed by 3.1% because of "temporary supply chain congestion and multi-phased project execution", EMEA is shrinking, and the one independent voice in our knowledge base with a registered, dated, falsifiable test on this exact question predicted precisely this failure mode ten weeks before it happened. The business is excellent, the balance sheet is now a fortress, the margin trajectory is outstanding, and the price is 28% off its high — and none of that answers the question the October print will answer. We would rather pay $290 in November knowing, than $270 in August guessing.

Provenance & disclosures

• Traceability: 0 knowledge-base claims name Vertiv Holdings out of 51,928 distilled claims (raw hits 8, case-sensitive entity hits ZERO, name-level 0, discarded 8; breadth 0, net conviction none). For the largest independent supplier of data-centre power and thermal-management equipment, in a store saturated with AI-infrastructure argument, the total absence is itself the finding. The free-text lane on "liquid cooling", "Vertiv", "data center cooling", "thermal management" and "power distribution unit" returned 8 hits across 4 channels, of which the most consequential is jordi_visser, 2026-05-24, neutral, conviction 72, skill 2.0, speaker_role: independent — "AI is a physical-world CapEx cycle (HBM, racks, liquid cooling, copper, fiber, substations), only ~12-18% spent; bottlenecks make the buildout lumpier, more inflationary and more volatile than the market models" — carrying the registered TEST "AI buildout produces cost inflation, delays and revenue-recognition problems over the next year", which Vertiv's 3.1% revenue miss on "temporary supply chain congestion and multi-phased project execution" appears to have confirmed ten weeks later. Also: jordi_visser 2025-07-20 (bullish 82, "own hardware, data center semis, HBM, liquid cooling, power, not software"), jordi_visser 2026-05-11 (bullish 80, named speaker), odd_lots 2024-06-21 (Brian Venturo, speaker_role: management, half-weighted, "liquid cooling cuts electricity utilization ~60-70%"), and no_priors 2024-12-12 (bullish 85) — a technology-SUBSTITUTION risk: "Synthetic-diamond cooling attacks heat at the chip/material level and layers on top of any air/liquid cooling." CONCENTRATION SENSITIVITY TEST — REQUIRED AND RUN: jordi_visser supplies 4 of 8 hits (50%), far above the one-third threshold. Stripping that channel leaves two jensen_huang claims (a MANAGEMENT channel for a different registrant, one of them about a competitor), one half-weighted management technical fact, and one substitution threat — so the bullish reading of this lane depends almost entirely on one thinker, whose most recent claim is a warning. speaker_role fired on two claims and failed on the management-channel jensen_huang claims; management voice was judged by content, not by the flag. One jensen_huang claim carries the synthetic speaker slug jensen_huang_ai, noted for the record. All quotes are verbatim from the stored claim text.
• Data as-of: all operating figures — net sales, organic growth split, gross margin, SG&A, segment results, interest, tax, liquidity, debt and cash flow — through 2026-06-30, from the 10-Q filed 2026-07-29, read as narrative prose because the filing carries no statement tables · annual and quarterly financial data from the vendor payload · estimates 2026-08-04 · prices 2026-08-04, quote timestamp 1785873603 = 2026-08-04T20:00:03Z ($269.93, +2.62%; 50-DMA $303.62; 200-DMA $247.20; RSI 39.4; MACD −15.97) · knowledge-base claims 2026-08-04. Vertiv's fiscal year is the calendar year. All figures come from the Synthos vendor data file for VRT or from the SEC filings in the VRT archive; no figure comes from memory, recall or external retrieval.
• Filing archive contents, and a MATERIAL LIMITATION disclosed up front: 10-K filed 2026-02-13 (fiscal 2025); 10-Q filed 2026-04-22 (March 2026 quarter); 10-Q filed 2026-07-29 (June 2026 quarter — the source of every operating figure in this dive); 8-K filed 2026-06-12 (Item 7.01 Regulation FD, exhibit not extracted); 8-K filed 2026-06-18 (Item 5.07 annual-meeting results); 8-K filed 2026-07-29 (Item 2.02 results and Item 7.01, exhibit NOT extracted). NONE of the six carries preserved [TABLE] statement data — the manifest records tables: false for every document and a direct check of the June 10-Q returns zero table markers. There is therefore no readable balance sheet, income statement, cash-flow statement or segment table in this archive, and several narrative numeric values have been stripped in extraction (the acquisition note's "net of cash acquired of $" and the loss on debt extinguishment both terminate without a figure). What survives is the MD&A prose, which fortunately carries the income-statement, segment, liquidity and cash-flow figures in complete sentences, and which is the source of every filing-derived number here. The three 8-Ks are 1.2–1.8 KB stubs, so the second-quarter earnings release is NOT available and NO company guidance of any kind — revenue, EPS, EBITDA or free cash flow — can be sourced. This dive does not estimate around it: all forward figures are consensus, labelled as such. Confidence is explicitly lowered on this name relative to the table-carrying names in this batch.
• Where the filings contradicted or corrected the vendor (detailed in Section 7): earn_cal contains a CORRUPT RECORD — the 2026-04-22 (Q1 2026) row carries epsActual: 1.52, revenueActual: 3,274,300,000 and revenueEstimated: 3,379,044,000, identical in all three fields to the 2026-07-29 (Q2 2026) row, with lastUpdated: 2026-07-29, while inc_q and the 10-Q both establish Q1 2026 revenue as $2,649.5M — so the Q1 2026 adjusted EPS actual is not recoverable and no trailing adjusted EPS is computed anywhere in this dive; est.netIncomeAvg EXCEEDS est.ebitAvg in every forward year (FY2026 $2,638.2M against $1,425.1M; FY2030 $6,222.3M against $2,872.1M), which is impossible, alongside a fixed-ratio signature (exactly 14.24% and 10.19% of revenue in every forward year) and an implied 10.19% EBIT margin against a filing-derived 21.7%; enterpriseValueTTM implying $445.6M of net debt against bal_a's $1,675.0M — a $1,229M internal disagreement in which, as on GD in this same batch, the EV field is the CLOSER of the two to the filing's approximately net-cash position at 2026-06-30 ($2,810.6M cash plus $300.0M short-term investments against $2,950.0M principal debt); bal_a totalDebt including $245.2M of lease obligations, a 7.8% overstatement of the T/TJX class; seg_geo summing 8.3% above reported revenue because it presents segment figures including intercompany sales while the 10-Q presents them excluding, and dropping the United States split present in FY2020; and inc_a FY2022 reporting eps of −$0.038 against netIncome of +$76.6M with zero minority interest, an unexplained inconsistency, alongside a $458M swing in totalOtherIncomeExpensesNet between FY2024 and FY2025 that the block does not explain. Where vendor and filing AGREED — recorded: inc_q Q2 2026 net sales of $3,274.3M, gross profit of $1,234.9M and net interest of $17.4M all match the 10-Q's prose exactly, as does Q2 2025's $2,638.1M; seg_prod sums to reported revenue exactly in every year (FY2025: Product $8,390.6M + Service $1,839.3M = $10,229.9M); tech.max_dd_from_peak of −28.25%, carrying the ADP-class signature of equalling pct_from_hi, was CHECKED against the return series and found CORRECT (ret_12m of +90.6% implies ~$141.6 a year ago against lo52 of $121.82, so the low preceded the high); capexToDepreciationTTM of 1.111 shows no COP-class understatement, and the filing independently guides FY2026 capital expenditure to $550.0–570.0M against FY2025's $220.0M; and Item 4 was checked specifically because the risk-factor list retains legacy "material weakness" language — the chief executive and chief financial officer concluded disclosure controls were EFFECTIVE at 2026-06-30 and no current material weakness is disclosed.
• Basis note: analyst estimates are on adjusted diluted EPS. No GAAP-to-adjusted reconciliation is available in this archive because the earnings-release exhibit was not extracted; the June-quarter gap can be bounded from the payload — GAAP net income of $497.8M over 392.7M diluted shares is approximately $1.27 against an adjusted $1.52, a 20% adjustment — and the principal components visible in the 10-Q are intangible amortisation (up $26.8M year on year in the quarter) and the $28.8M PurgeRite contingent-consideration charge ($62.0M in the half). All forward multiples are struck on the adjusted basis to match consensus; trailing multiples shown are GAAP only, because the corrupt earn_cal record makes a trailing adjusted figure unrecoverable.
• Estimate coverage: 14 analysts on FY2026 EPS, 16 on FY2027 — the anchor for all three fair values — with an FY2027 range of $8.477 to $9.788, a 15.5% spread; 10 on FY2028; 4 each on FY2029 and FY2030, which are excluded from every conclusion. Revenue coverage: 19, 21, 21, 10 and 10 analysts respectively.
• Peer note: the vendor peer set — CTAS, EMR, ITW, JCI, NOC, PWR, TDG, TRI, UPS and WM — contains a uniform-services company, a waste company, a parcel carrier, a defence prime and a news-and-data business alongside four industrials. Emerson Electric and Johnson Controls are arguably adjacent; nothing here is a data-centre power and thermal specialist. The payload supplies only market capitalisations and no multiples, so no peer-multiple comparison is drawn. PWR is separately covered in this same batch.
• Fair-value caveat: the $180 / $310 / $420 anchors are multiples of the FY2027 consensus adjusted EPS distribution — 21.2x the low of $8.477, 34.1x the mean of $9.091, and 42.9x the high of $9.788 — each cross-checked against FY2028E (15.6x, 26.8x, 36.3x). Stated arithmetic, not a discounted cash flow. The base is sensitivity-disclosed: 30x gives $273, essentially spot; 38x gives $345. The base case assumes meaningful forward-multiple COMPRESSION, from 40.4x on the current year toward roughly 34x, because consensus itself has revenue growth decelerating from 36.8% to 22.0% and EPS growth from 61.7% to 27.1% across three years — the transition from a 60%-growth multiple to a 30%-growth multiple happens whether or not the company executes. Our base of $310 sits 13.3% BELOW the street consensus of $357.47, and that negative divergence against a 19-buy, 1-hold, 0-sell coverage set with no target below spot is stated as an explicit variant perception.
• Half-weight management framing: the descriptions of demand as "very robust", of the supply-chain congestion as "temporary", of customer requirements "evolving toward higher-density, more complex infrastructure environments", and the capacity-expansion rationale are all company statements and are half-weighted under house rules, identified as such at every use. The net sales figures, the organic/acquisition/currency split, the segment results, the gross-margin percentages, the debt schedule, the cash balances and the cash-flow line items are quantified filing disclosures and are treated as facts. One knowledge-base claim used in this dive carries speaker_role: management (Brian Venturo, odd_lots) and is half-weighted accordingly.
• Timing: second-quarter 2026 results were released 2026-07-29, six days before this dive: adjusted EPS of $1.52 beat the $1.43 consensus by 6.3%, while revenue of $3,274.3M MISSED the $3,379.0M consensus by 3.1%. The next print is 2026-10-28, 85 days away (consensus adjusted EPS $1.81 and revenue $3,751M, the latter requiring a 14.6% sequential increase). The most recent insider filings are dated 2026-06-26 and are dividend-equivalent accruals. 2026-08-04 carried no company-specific filing.
• Accessibility note: no information in this dive is conveyed by colour. All emphasis is carried by bold text, table structure and explicit labelling.
• Not investment advice. Independent research, educational and informational only, never personalised. No recommendation to buy, sell or hold any security is made to any person.
• Version: 2026-08-04-full.