PUMPProPetro Holding Corp.
Industry Energy materials Updated 25 Sep 2026
The idea
A frac company becoming a power producer (PROPWR): no return in the base case, a triple if 1.1GW is earning in 2028 — a position-sizing question, not a valuation one. The bear case impairs the equity.
Events to watch
- Q3 release (late Oct): the PROPWR contract book reconciled line by line (additions vs capacity coming off contract)
- EBITDA per MW, which management has not published
How we got the target
Scorecard model table: base ~$10 at 5.5x 2028 group EBITDA. The summary line says ~$12 and bull ~$24. Time frame: 2028.
Target: from the research notes. Prices: market data get_latest_stock_price, close of 2026-09-22 (fetched 2026-09-23)
What would prove it wrong
7 warning signsContracted PROPWR megawatts fall, or fail to reach 700 MW by the Q2 2027 release, on a reconciliation that shows gross additions AND capacity that came off contract.
Details
Management's own July 2026 trajectory was 350 → 450 "very soon" → "several hundred megawatts" of data centre.
A data centre contract is not signed by the Q1 2027 release.
Details
Management says it expects "the overwhelming majority" of future capacity to go to data centres, has 1.1 GW of equipment ordered, and as of 2026-09-22 has one 60 MW deployment and no announced data centre contract. Two more quarters of oil-and-gas and industrial awards means the thesis has changed.
Another equity raise below $12.
Details
The January 2026 offering was at $10.00 and the stock is at $10.03. A third round of dilution to finance uncontracted equipment is the bear case arriving.
Completions adjusted EBITDA falls below $30m in a quarter.
Details
It was $44.8m in Q2 2026 group-wide including a −$3m power drag, on 12 fleets going to 13. The completions business is what funds the gap until 2028; if the Permian cycle takes it away, the Caterpillar order book becomes the problem.
Capital destroyed again: a second impairment, or the PROPWR programme still below a 5% return on its invested capital at the FY2027 10-K (added 2026-09-25 by the S2 capital-allocation check, which came out
destroys). The five-year record is an average ROIC of about −3% with $212.2m written off in 2024; PROPWR has absorbed $198.4m of capital expenditure incurred in 2025 and $118.4m in H1 2026 for −$11.6m and −$6.0m of adjusted EBITDA. Two more years of that with the book still not earning is the completions cycle funding an empty programme.A third equity raise, or any raise that takes the diluted count above 135m without contracted megawatts above 900 (added 2026-09-25).
Details
The company bought back ~11.5m shares at ~$9.6 or below and reissued 15.0m at $10.00 eleven months later; the count is already +18% in six months.
A second finance-leadership departure, or an ICFR exception on the power-generation accounting.
Details
RSM's sole CAM is that accounting; the CAO resigns on 2026-10-30 with the CFO doubling as interim principal accounting officer; and the proxy shows a third person held the CFO/PFO role during 2025.
How far along
5 of 10 steps done · 1 need your decision · 4 not started
Next step
on the Q3 2026 release (late October), reconcile the PROPWR contract book line by line: contracted megawatts at the start of the quarter, gross additions, capacity that came off contract, and the closing total. The 2026-09-22 release headlines a 230 MW addition and reports a 510 MW total against a 350 MW starting point — arithmetic that implies about 70 MW came off contract and is disclosed only as a subordinate clause with no number attached. Second: whether the 1.1 GW of equipment already ordered or delivered has closed any of the gap to the ~510 MW contracted.
Needs your decision
S6 product test / expert call, S7 conviction.
Failed checks
none outright. S0/S4/S6/S7 are not closed awaiting cluster-energy-materials-industrial and were not improvised, so this name has no Stage 0 lens, no industry map and no scuttlebutt in this repo. S2 passes its kill test with the loudest warning in the file, and its capital-allocation verdict is destroys (added 2026-09-25) — a ⚠️, not a kill, per §25. S3 is closed (gate-minimum) with 4 earnings-quality flags, the loudest being a headline free-cash-flow measure redefined twice in three years; its one-transcript deviation is closed. S8 is closed reliable, so the S7 valuation base case is not haircut for management's guide.
Who runs it
The Front OfficeStep by step
- S0How it does as AI gets cheapNot started
awaiting cluster pass
cluster-energy-materials-industrialnot improvised here. Note for that pass: PUMP is two businesses (Permian completions + PROPWR behind-the-meter power) and the SKILL.md §4 instrument choice is not obvious.
Full notes
The completions half is a cyclical service business; the power half is closer to the concession variant (§4.4) than to the AI-surplus four, since its revenue comes from multi-year contracted capacity. Liquidity for whichever runs it: ADV $40.0m/day (21 sessions to 2026-09-22, market data NYSE:PUMP), market capitalisation ~$1.23bn — the thinnest name in this repo so far, and any position size must respect it.
- S1Where the idea came fromDone
watchlist (owner, Apple Stocks, 2026-09-20), promoted by Tier 0 on 2026-09-22 at priority 2 for the 8-K filed 2026-09-21. That 8-K is the CAO resignation; a second 8-K the following day carries the Targa award.
Full notes
Promotion handled and removed from
queue.json. - S2Quick deal-breaker checkFlagged
**the loudest ⚠️ in this repo
read it before anything else. H1 2026 revenue $576.5m against $685.6m (−16%); hydraulic fracturing $386.6m against $515.1m (−25%) H1 adjusted EBITDA $81.2m against $122.3m (−34%) Net loss in both quarters of 2026 ($3.6m Q1, $8.1m Q2) Free cash flow −$30.4m over the half against +$40.4m a year earlier Against that, FY2026 capital expenditure guidance is $525-595m, of which $400-450m is PROPWR — roughly five times H1 EBITDA annualised It does not die, because it is pre-funded: $784m of cash at 2026-06-30 and $905m of total liquidity, after a $690m 0%-coupon convertible in May 2026 and a 15,000,000-share offering at $10.00 in January 2026 Equity $957.4m, total debt ~$797m (from ~$122m at 2025-12-31), so net debt is roughly nil today Five-year share count 114.5m (2022) → 104.3m (2025) → 122.8m issued at 2026-06-30: +18% in six months, and the converts add more above $29.49 Analyst coverage: not counted this run Verdict: survives the next twelve months on the balance sheet it has already raised; the question is what it owns at the end of them. · capital allocation (added 2026-09-25): DESTROYS. 2021-2025 operating cash flow $1,313.8m; uses $1,353.4m — capex $1,160.7m, acquisitions $81.9m, buybacks $110.8m, no dividend ever ROIC by year −7.6% / −0.2% / +10.2% / −16.3% / +0.6%, five-year average ≈ −2.7%, invested capital $714.4m → $844.1m $212.2m written off in 2024 alone ($188.6m Tier II fleet + $23.6m wireline goodwill), and the $81.9m of acquisitions also gave back earnout reversals of −$2.6m and −$4.9m Buybacks bought ~11.5m shares at ≤~$9.6 blended (the only disclosed average is Q4 2024, 423,476 shares at $7.98, FY2024 10-K Item 5) — then 15,000,000 new shares were sold at $10.00 in January 2026 Diluted count 102.7m (2021) → 105.4m (2025) up 2.7% despite $110.8m of buybacks**, then 122.8m issued at 2026-06-30 Not a kill on its own (§25) — but the team now spending $525-595m a year has a five-year record of allocating capital at a negative return.
- S3Reading the company's reportsDone
(gate-minimum)
FY2025 10-K Risk Factor Summary (three disclosed risks named below, in management's own ordering) and RSM's report. Q2 2026 10-Q (balance sheet, Note 10 related parties, segment tables); the 2026-07-29 Q2 results release.
Full notes
the 2026-09-21 and 2026-09-22 8-Ks; the 2026-01-27 offering release; the 2026-04-08 DEF 14A (ownership and bonus metrics); one earnings-call transcript; and a full efts.sec.gov red-flag sweep. RSM US LLP, PCAOB 49, and the sole critical audit matter is the accounting for the power generation arrangements — embedded-lease control and standalone selling prices for non-lease components. ICFR opined effective at 2025-12-31.
⚠️ the index carries an NT 10-K at 2024-03-01 (the FY2023 10-K then filed 2024-03-13) and NT filings in 2019-2020; no 10-K/A or 10-Q/A in the index and no "changes in accountants" hit. Deviation (first pass): one transcript, not two, and only turns 9-20 of 80 — CLOSED 2026-09-25, all eight archived calls read. · EQ (added 2026-09-25): 4 flags. (1) The headline free-cash-flow measure has been redefined twice in three years, each time to exclude the cash outflow of the moment — "Free Cash Flow adjusted for Acquisition Consideration" in 2023-24, "Free Cash Flow for Completions Business" in 2025, which adds PROPWR's operating AND investing cash flows back: $190.2m against group FCF of $81.8m. (2) "Retention bonus and severance expense" added back in 3 years of 3 ($2.3m / $2.3m / $2.6m), as is "other G&A — nonrecurring professional fees". (3) Write-offs of prematurely failed power ends reclassified into depreciation, restating FY2024 disposal from a $7.5m loss to a $4.9m gain ($12.4m moved). (4) Guided capex is "incurred", not cash: FY2025 $281.2m incurred vs $186.3m paid. Clean: SBC 1.3% of revenue, FCF after SBC 79% of reported; accruals −18.4% of assets; OCF/adjusted EBITDA 111%. Watch: DSO 49.5 → 57.7 days in 2025 while revenue fell 12% (one period, not two).
- S4Rivals and suppliersNot started
awaiting cluster pass
cluster-energy-materials-industrialnot improvised. Note for it: PROPWR's supply chain is a single supplier, Caterpillar, under a framework agreement for ~2.1 GW.
- S5Who owns the stockDone
the register turned over completely in 2026. ExxonMobil, holder of 16,600,000 shares (13.54% on the 2026 proxy) through the Pioneer merger, sold the entire position in May 2026 (13D/A filed 2026-05-20) and is no longer a related party.
Full notes
With January's 15,000,000-share offering that is ~31.6m shares of supply in five months against 122.8m outstanding. Who is on the other side now: index money (BlackRock 9.1m, State Street 3.5m), Sourcerock 7.7m (5.58% on the proxy), and a wall of event-driven and multi-strategy funds ADDING in Q2 — D. E. Shaw +1.29m, Encompass NEW 4.57m, Cooper Creek +1.34m, VR Advisory +1.63m, Citadel +1.68m, Millennium +1.53m, Two Sigma +1.15m, Wellington +2.06m, Assenagon NEW 2.61m. Against them, short interest of 16,221,951 shares at the 2026-08-31 settlement = 13.2% of shares outstanding, 5.4 days to cover — by far the most contested name in this repo (ALL 2.4%, AIG 1.7%). Insiders: 13 directors and executive officers hold 1,894,695 shares, ~1.5%. Shelf: live and used — the January offering came off a Form S-3, and the May convertible off the same structure.
⚠️ the Q2-26 13F quarter is incompletely ingested (179 filers vs 253).
⚠️
⚠️ the Sixth Street row is NOT a holding — seventh issuer, see the log.
- S6What workers and customers sayNot started
awaiting cluster pass
cluster-energy-materials-industrialno scuttlebutt of any kind exists in this repo for this name, and none was gathered this run (budget went to ALL). Product test / expert call open OPEN (user)
- S7The case for it, and pricesNot started
awaiting cluster pass
cluster-energy-materials-industrialnot improvised. Kill criteria below are written from this pass's evidence because a promoted name needs them; the thesis is the cluster's.
Full notes
Conviction open OPEN (user) · valuation (bear/base/bull; moved from S8 2026-09-23): see the valuation table below. $10.03 (2026-09-22), level with the January offering price of $10.00; P/B 1.29 on $7.80 of book. Base ~$12, bear ~$5, bull ~$24 — the widest spread of any name in this repo, and deliberately so. ADV $40m/day.
- S8Do the bosses keep their word?Done
reliable
8 rows, written 2026-09-25 from all eight archived calls plus the 8-K Ex. 99.1 outlooks.
Full notes
3 beat, 3 met, 1 missed, 1 withdrawn. Current team made 4 of 8 — Schorlemer (CFO, separated 2025 Q1) made the other four, and every one of his is a beat or a meet. The record splits by business: every beat and meet is completions or balance sheet (FY2024 capex $133.4m against a $200-250m first guide. FY2025 capex $281.2m against $300-400m; 14 fleets guided and run in Q4 2024; 14-15 in Q1 2025; the $350m Stonebriar facility guided as an LOI in October and signed in December). Both broken rows are PROPWR: the fifth FORCE fleet promised "active in early 2025" and quietly restated to "in 2025" (withdrawn — the slip is never called a delay), and "at least 2 fleet buyouts during 2026" cut to one with the second pushed to early 2027 (missed, FY2026-Q2 call).
⚠️ the one PROPWR beat — ≥220 MW contracted by end-2025, delivered at ~240 MW — is scored on committed megawatts, a bonus-plan metric the company reports GROSS (see the finding paragraph); it is a beat against the number management gave, not necessarily against the book. In flight, not scored: FY2026 capex, first guided $390-435m and raised to $540-610m one quarter later (+24% to +40%) — with rows 1 and 4 that makes three years of capex guidance with no predictive value in either direction; and "PROPWR contributing meaningful earnings in 2H 2026", against −$6.0m of H1 2026 power adjusted EBITDA, first scoreable on the late-October Q3 release. Three finance leaders in nineteen months (Schorlemer separated 2025 Q1; Davila interim PFO then CAO, resigning 2026-10-30; Weatherl CFO from 2025-07-14 and interim PAO from 2026-10-30).
- S9What to watch nextDone
**re-checked 2026-09-23 by slot
names-ai-capex-4; unchanged, and the promotion that prompted the re-check is stalesee the note below. trigger written; next earnings expected late October 2026 (DERIVED from filing cadence: Q3 results 8-Ks landed 2025-10-29 and the Q2 2026 one 2026-07-29) Tier 0 EDGAR sweep covers the feed (cik 1680247), and on this name 8-K Item 7.01 is the form to watch** — every PROPWR contract number arrives that way and none of them are in a periodic filing.
Sources and notes
Sources 11 documents
| Short name | Document | Where it came from | Read on |
|---|---|---|---|
| 10K-2025 | ProPetro 10-K for FY2025, filed 2026-02-19 | https://www.sec.gov/Archives/edgar/data/1680247/000168024726000028/pump-20251231.htm | 2026-09-23 |
| 10Q-Q2-26 | ProPetro 10-Q for the quarter ended 2026-06-30, filed 2026-07-30 | https://www.sec.gov/Archives/edgar/data/1680247/000168024726000097/pump-20260630.htm | 2026-09-23 |
| 8K-TARGA | ProPetro 8-K filed 2026-09-22, Exhibit 99.1 — PROPWR commits ~230 MW to a Targa Resources subsidiary | https://www.sec.gov/Archives/edgar/data/1680247/000168024726000112/propwrtargaannouncement.htm | 2026-09-23 |
| 8K-CAO | ProPetro 8-K filed 2026-09-21 (event 2026-09-16), Item 5.02 — Chief Accounting Officer resigns | https://www.sec.gov/Archives/edgar/data/1680247/000168024726000110/pump-20260916.htm | 2026-09-23 |
| 8K-Q2-REL | ProPetro 8-K filed 2026-07-29, Exhibit 99.1 — second quarter 2026 results release | https://www.sec.gov/Archives/edgar/data/1680247/000168024726000095/exhibit991-earningsrelea.htm | 2026-09-23 |
| 8K-EQUITY-JAN26 | ProPetro 8-K filed 2026-01-28, Exhibit 99.1 — pricing of an upsized public offering of common stock | https://www.sec.gov/Archives/edgar/data/1680247/000110465926007199/tm264193d1_ex99-1.htm | 2026-09-23 |
| DEF14A-26 | ProPetro DEF 14A filed 2026-04-08 (ownership table and 2025 Annual Bonus Plan metrics) | https://www.sec.gov/Archives/edgar/data/1680247/000110465926040891/tm261395-1_def14a.htm | 2026-09-23 |
| 13F-Q2-26 | Institutional holders by manager, 13F report date 2026-06-30 (ingestion incomplete — see the edge note) | python ledger/research/tools/threespread.py holders PUMP | 2026-09-23 |
| 13DA-XOM | Schedule 13D/A on ProPetro, filed 2026-05-20 for event date 2026-05-18 | https://www.sec.gov/Archives/edgar/data/1680247/000119312526232878/0001193125-26-232878-index.htm | 2026-09-23 |
| CALL-Q2-26 | ProPetro Q2 FY2026 earnings call, 2026-07-29, speaker-labelled transcript | Equibles GetEarningsCallTranscript(ticker=PUMP, fiscalYear=2026, fiscalQuarter=2) — https://equibles.com/stocks/pump/earnings-calls | 2026-09-23 |
| QUEUE | Tier 0 promotion record — ledger/research/queue.json, promoted 2026-09-22, priority 2 | ledger/research/queue.json | 2026-09-23 |
Links to other companies 21 links
| From | Link | To | Date | Where it says so |
|---|---|---|---|---|
| Targa Resources Corp. (NYSE: TRGP) | customer of | PUMP | 2026-09-22 | 8K-TARGA — Exhibit 99.1, headline and first two paragraphs |
| PUMP | context | PUMP | 2026-09-22 | 8K-TARGA — Exhibit 99.1, first paragraph, sentence beginning 'With the addition of these contracts' |
| PUMP | context | PUMP | 2026-07-29 | 8K-Q2-REL — Exhibit 99.1, 'Recent PROPWR Highlights', first bullet |
| PUMP | governance structure | PUMP | 2026-04-08 | DEF14A-26 — Compensation Discussion and Analysis, 2025 Annual Bonus Plan metrics paragraph |
| PUMP | context | PUMP | 2026-07-29 | CALL-Q2-26 — Q2 FY2026 call, CFO answer to Saurabh Pant (Bank of America) on liquidity and capex |
| Exxon Mobil Corporation (including Pioneer and XTO Energy) | customer of | PUMP | 2026-06-30 | 10Q-Q2-26 — Note 10 - Related-Party Transactions, ExxonMobil and Pioneer |
| Exxon Mobil Corporation (including Pioneer and XTO Energy) | insider transaction | PUMP | 2026-05-18 | 13DA-XOM — Schedule 13D/A filed 2026-05-20, event date 2026-05-18 (filer identity not opened this run — see the dossier's access limits) |
| Exxon Mobil Corporation (including Pioneer and XTO Energy) | context | PUMP | 2026-06-30 | 10Q-Q2-26 — Note 10 - Related-Party Transactions, ExxonMobil and Pioneer, third sentence |
| Celina A. Davila (Chief Accounting Officer until 2026-10-30) | key person risk | PUMP | 2026-09-16 | 8K-CAO — Item 5.02, first paragraph |
| Caleb Weatherl (CFO; interim principal accounting officer from 2026-10-30) | officer of | PUMP | 2026-09-16 | 8K-CAO — Item 5.02, first and second paragraphs |
| Samuel D. Sledge (CEO) | officer of | PUMP | 2026-09-22 | 8K-TARGA — Exhibit 99.1, management quotation |
| RSM US LLP (PCAOB ID 49) | auditor of | PUMP | 2026-02-19 | 10K-2025 — Reports of Independent Registered Public Accounting Firm, Critical Audit Matters |
| Caterpillar Inc. (PROPWR equipment supplier, strategic framework agreement) | supplier dependency | PUMP | 2026-07-29 | 8K-Q2-REL — Exhibit 99.1, PROPWR Update, CEO commentary |
| Caterpillar Financial Services Corporation | financing partner | PUMP | 2026-06-30 | 8K-Q2-REL — Exhibit 99.1, Liquidity and Capital Spending |
| Goldman Sachs & Co. LLC (lead book-running manager, January 2026 equity offering) | event | PUMP | 2026-01-27 | 8K-EQUITY-JAN26 — Exhibit 99.1, first paragraph |
| BlackRock, Inc. | holds | PUMP | 2026-06-30 | 13F-Q2-26 — 13F-HR, accession 0002012383-26-003238 (holders table, row 1) |
| Sourcerock Group LLC | holds | PUMP | 2026-06-30 | 13F-Q2-26 — 13F-HR, accession 0001822531-26-000009 (holders table, row 2) |
| Encompass Capital Advisors LLC | holds | PUMP | 2026-06-30 | 13F-Q2-26 — 13F-HR, accession 0001172661-26-003734 (holders table, row 4) |
| Sixth Street Partners Management Company, L.P. | context | Charles Schwab Investment Management, Inc. | 2026-06-30 | 13F-Q2-26 — threespread.py holders PUMP — identical-share-count rows 16 and 17 |
| PUMP | context | PUMP | 2026-02-19 | 10K-2025 — Risk Factor Summary, 'Risks Inherent in Our Business and Industry', bullets 1-3, p.iii |
| PUMP | context | PUMP | 2026-09-22 | QUEUE — queue.json, promoted[], ticker PUMP, why '8-K filed 2026-09-21 (current report)' |
History 4 entries
- 2026-09-23 — first pass, slot
names-insurance-2, Tier 0 promotion (priority 2, 8-K of 2026-09-21), model-drafted unattended by the cloud research routine. S1, S2, S3 (gate-minimum), S5 and S9 closed; S8 modelled and left open. S0, S4, S6 and S7 are ⬜ awaitingcluster-energy-materials-industrialand were not improvised — this name is in a cluster that has not been passed, exactly as AX was on 2026-09-22. Run finding: the 2026-09-22 PROPWR release headlines a 230 MW Targa award and reports a 510 MW total against a 350 MW starting point, implying roughly 70 MW came off contract, undisclosed as a number — while "PROPWR committed megawatts" is one of four quantitative metrics in the annual bonus plan. Second finding: 1.1 GW of equipment ordered or delivered against ~510 MW contracted, roughly $825-885m of uncontracted capital at the company's own cost per megawatt. Third: ExxonMobil, the largest customer, sold its entire 13.54% stake in May 2026, the XTO fleet agreement expires in late 2026 and the company says it does not expect renewal, and ExxonMobil revenue is already −45% year on year in the quarter. Fourth: the Chief Accounting Officer resigns effective 2026-10-30 with the CFO taking the role on an interim basis, at a company whose auditor's sole critical audit matter is the accounting for the power arrangements and whose 2026 proxy names a third person as having held the CFO/PFO role during 2025. Tooling: the Sixth Street / Charles Schwab IM 13F defect appears for the seventh issuer (2,086,479 shares); row discarded. Evidence and access limits indossier-2026-09-23.md. Nothing here is human-verified. - 2026-09-23 — S9 re-check only, slot
names-ai-capex-4, as the fourth member under the promotion rule. No new filing since 2026-09-22; no stage advanced; no figure changed. The promotion was a duplicate of onenames-insurance-2had already handled the same day, and PUMP is removed fromqueue.jsonpromoted. See the S9 re-check section above. - 2026-09-23 — stage redefinition, owner decision (WATCHLIST_PLAN.md §25): S8 is now Management credibility and is not yet run for this name; the old S8 valuation figures moved to the
valuation:line under S7, and any proposed buy price or position size was removed (the pipeline no longer drafts either). No research was redone. S2 capital allocation and S3 earnings quality are also pending, in the same backfill slot. - 2026-09-25 — backfill slot
backfill-2026-09-23-8, the three checks added on 2026-09-23 (WATCHLIST_PLAN.md §25). No other stage touched. S8 written: reliable — 8 rows, 3 beat, 3 met, 1 missed, 1 withdrawn, current team 4 of 8. S2 gained its capital-allocation table: verdictdestroys— five-year average ROIC ≈ −3%, $212.2m impaired in 2024, buybacks at ≤~$9.6 followed by an issue at $10.00, diluted count up 2.7% over five years despite $110.8m of repurchases; two kill criteria added. S3 gained its earnings-quality check: 4 flags, the headline one being a free-cash-flow measure redefined twice in three years (completions-only FCF $190.2m against group FCF $81.8m). Dodge log opened, three entries, all FY2026-Q2, two of them the same topic. All eight archived calls were read, which closes the first pass's one-transcript deviation. Sources were EDGAR XBRL, the 8-K Ex. 99.1 releases of 2024-02-21, 2025-02-19 and 2026-02-18, the FY2024 and FY2025 10-Ks, four Item 5.02 8-Ks and_transcripts/PUMP/; no Equibles call was spent. Evidence and locators indossier-2026-09-25.md. Nothing here is human-verified.
Full notes, 25 Sep 2026 17 min read
PUMP — dossier 2026-09-25 (backfill slot backfill-2026-09-23-8)
Model-drafted, unattended, by the cloud research routine. Nothing here is human-verified. Scope: only the three checks added on 2026-09-23 (WATCHLIST_PLAN.md §25) — S8 management credibility, S2 capital allocation, S3 earnings quality. Every other stage stands as written in dossier-2026-09-23.md; this file does not revisit S0/S1/S4/S5/S6/S7/S9.
Sources read this run, all free: SEC EDGAR XBRL companyfacts for cik 1680247 (accession 0001680247-26-000028 carries the FY2023–FY2025 columns); the 8-K Ex. 99.1 earnings releases of 2024-02-21 (0001680247-24-000034), 2025-02-19 (0001680247-25-000026) and 2026-02-18 (0001680247-26-000024); the FY2024 and FY2025 10-Ks; the Item 5.02 8-Ks of 2025-03-04, 2025-07-14, 2026-05-28 and 2026-09-21; and all eight earnings calls in ledger/research/_transcripts/PUMP/ (FY2024-Q3 through FY2026-Q2, market data). No Equibles call was made for this name.
Transcript guard (the check the 2026-09-24 AX pass asked for): each of the eight files names its own fiscal quarter in its opening turn. FY2024-Q4 opens with an IR turn rather than the operator's, and names "fourth quarter and full fiscal year 2024" inside the same turn. None is mislabelled.
Speaker-name caution, same class of defect as the OMAB CFO: market data renders the interim principal financial officer as "Selena Davila" in the FY2025-Q1 file and "Celina A. Davila" in FY2025-Q2. The 8-K of 2025-07-14 is the authority: Celina A. Davila. It also renders the CEO as both "Sam Sledge" and "Samuel D. Sledge", and writes PROPWR as "PROPOWR"/"PROPWER" in the 2026 files.
S8 — management credibility
Who was in the chair, from EDGAR
| Person | Role | From | To | Source |
|---|---|---|---|---|
| Samuel D. Sledge | CEO | before this window | current | throughout |
| David Schorlemer | CFO | before this window | separated 2025 Q1 | 8-K 2025-03-04, 0001104659-25-020011 (Item 5.02, separation) |
| Celina A. Davila | interim principal financial officer, then CAO | 2025 Q1 | resigns as CAO effective 2026-10-30 | 8-K 2025-07-14 0001680247-25-000112; 8-K 2026-09-21 0001680247-26-000110 |
| Caleb L. Weatherl | CFO (PFO from the Q2 2025 10-Q), interim PAO from 2026-10-30 | 2025-07-14 | current | 8-K 2025-07-14; 8-K 2026-09-21 |
Three finance leaders in nineteen months, and from 2026-10-30 the CFO is also the principal accounting officer of a company whose auditor's sole critical audit matter is the accounting for the power-generation arrangements. That was already the fourth finding of the 2026-09-23 pass; it is repeated here because it decides how much of the S8 record belongs to the people now in charge.
The say-vs-do table
Locators: the call is named by its fiscal quarter and call date (earnings-day precision, owner decision 2026-09-23); a written outlook is named by its 8-K accession. Speaker names are read off the transcript at the promise's position; roles come from the EDGAR table above as of the call date.
| # | Promise (metric, period, figure) | Made by / where | Path | Result / where | Verdict |
|---|---|---|---|---|---|
| 1 | FY2024 capital expenditures incurred $200–250m | company (written), 8-K Ex. 99.1 of 2024-02-21 (0001680247-24-000034), "Guidance and Recent Results" | cut three times during 2024; by the Q3 2024 call (2024-10-30) David Schorlemer, CFO, gave "$150 million to $175 million, down from our most recent prior guidance of $175 million to $200 million" | $133.4m incurred — segment table, 8-K Ex. 99.1 of 2026-02-18 | beat · departed |
| 2 | 14 active frac fleets in Q4 2024 | David Schorlemer, CFO — FY2024-Q3 call, 2024-10-30 | none | "In the fourth quarter of 2024, we had 14 active hydraulic fracturing fleets in line with our prior guidance" — FY2024-Q4 call, 2025-02-19 | met · departed |
| 3 | 14–15 frac fleets in Q1 2025 | David Schorlemer, CFO — FY2024-Q4 call, 2025-02-19 | none | "the 14 to 15 fleets we ran throughout the first quarter" (Sledge) — FY2025-Q1 call, 2025-04-29 | met · departed |
| 4 | FY2025 capital expenditures $300–400m (completions $150–200m, PROPWR $150–200m) | David Schorlemer, CFO — FY2024-Q4 call, 2025-02-19, and the same release | cut four times: $295–345m (Davila, Q1, 2025-04-29) → $270–310m (Davila, Q2, 2025-07-30) → $270–290m (Weatherl, Q3, 2025-10-29) | $281.2m incurred — segment table, 8-K Ex. 99.1 of 2026-02-18 | beat · departed |
| 5 | Fifth FORCE electric fleet active in early 2025 | Samuel D. Sledge, CEO — FY2024-Q3 call, 2024-10-30: "we expect the fourth fleet to be deployed under contract by year-end, with the fifth fleet active in early 2025" | restated to "in 2025" on the FY2024-Q4 call, then "expected to be deployed under contract this year" on the FY2025-Q1 call — the slip is never called a delay | five FORCE fleets under lease by the FY2025-Q4 call, 2026-02-18 | withdrawn (date moved without saying so) |
| 6 | At least 220 MW of PROPWR capacity contracted by the end of 2025 | Samuel D. Sledge, CEO — FY2025-Q3 call, 2025-10-29: "we now have over 150 megawatts contracted with expectations to reach at least 220 megawatts contracted by the end of the year" | none | "Expanded total committed capacity to approximately 240 megawatts" — 8-K Ex. 99.1 of 2026-02-18 | beat — but see the caveat below |
| 7 | $350m lease financing facility, from a letter of intent to a signed facility | Caleb L. Weatherl, CFO — FY2025-Q3 call, 2025-10-29: "we've executed a letter of intent for a $350 million leasing facility with an investment-grade partner" | none | "a $350 million leasing financing facility secured in December with Stonebriar Commercial Finance" — FY2025-Q4 call, 2026-02-18 | met |
| 8 | At least two FORCE fleet lease buyouts during 2026, $40–50m | Caleb L. Weatherl, CFO — 8-K Ex. 99.1 of 2026-02-18 ("approximately $40 million to $50 million related to lease buyouts"), reaffirmed on the FY2026-Q1 call | none | "Prior guidance contemplated at least 2 fleet buyouts during 2026. We now expect to complete the first planned buyout this year at a cost of between $15 million and $20 million, with the second shifting into early 2027" — FY2026-Q2 call, 2026-07-29 | missed |
Scored: 8 rows — 3 beat, 3 met, 1 missed, 1 withdrawn. Current executives made 4 of the 8 (Sledge 2, Weatherl 2); the four departed rows are David Schorlemer's and are all beats or meets.
Two rows in flight, recorded but not scored
- FY2026 capital expenditures incurred $390–435m, first guided by Caleb L. Weatherl on the FY2025-Q4 call (2026-02-18), raised to $540–610m one quarter later (FY2026-Q1 call, 2026-04-30, "up from the $390 million to $435 million range highlighted in our fourth quarter earnings report") and trimmed to $525–595m on the FY2026-Q2 call. The first guide was abandoned by +24% to +40% within a single quarter, and the company said so explicitly, so this is not a withdrawn row under §25's definition — but the year cannot now land inside it. Recorded here because of what it says next to rows 1 and 4: PUMP's capex guidance has had no predictive value in either direction for three years running — 33% below the bottom of the first guide in 2024, 6% below in 2025, and 35% above it in 2026.
- "By the second half of 2026, we expect PROPWR to begin contributing meaningful earnings" — Samuel D. Sledge, CEO, FY2025-Q4 call, 2026-02-18. H1 2026 Power Generation adjusted EBITDA was −$6.0m (2026-09-23 dossier, from the Q2 2026 10-Q segment table). The Q3 2026 release, expected late October, is the first period in which this promise can be scored. It is already the name's S9 trigger.
The caveat on row 6, which is the most important line in this section
Row 6 is the only PROPWR promise that scores a beat, and it is scored on committed megawatts — which is one of the four quantitative metrics in the 2025 Annual Bonus Plan (DEF 14A of 2026-04-08, read by the 2026-09-23 pass) and which the company reports gross. The 2026-09-22 8-K headlines a ~230 MW Targa award and reports ~510 MW of total committed capacity against the ~350 MW stated on the Q2 2026 call — arithmetic that implies roughly 70 MW came off contract, disclosed as a subordinate clause with no number attached. The beat is real against the number management gave; whether it is real against the book is the question the late-October reconciliation has to answer. A promise measured on a gross number that management also controls is worth less than the table's verdict suggests, and S8's verdict below is written knowing that.
Dodge log — opened this run
PUMP had no dodge log. The two most recent calls were read in the §3.1 order (safe harbour skipped, CFO section skimmed for guidance changes, Q&A read in full).
FY2026-Q2, 2026-07-29 — three dodges, two of them on the same topic:
- PROPWR contract economics · Derek Podhaizer (Piper Sandler) · asked for "pricing term, the return profile" on the newly contracted 110 MW and when contracts extend into the 10-year-plus range · Travis Simmering answered on grid availability, operator behaviour and "optionality" and gave no price, no term and no return · turn 23.
- PROPWR contract duration and pricing structure · Alexis Bruneau (Goldman Sachs) · asked for average contract duration and pricing structure on the new capacity, split between oil-and-gas and industrial · Samuel D. Sledge: "I think we're at a point right now in the life cycle of PROPWR where some of that's just a little bit too competitive to disclose", then "almost all these deals as long term in nature, most of them multi-year" · turn 27. Management said why it would not answer, which is better than pretending to, and the question still went unanswered — the second analyst in one call to ask it.
- PROPWR cash needs beyond 2026 · Saurabh Pant (Bank of America) · asked explicitly for colour "beyond" the 2026 guidance on cash needs over the next 12 months · Caleb L. Weatherl answered on 2026's guide, the CAT facility and current liquidity; Sledge answered on strategy; the years after 2026 were not addressed · turns 7–8.
FY2026-Q1, 2026-04-30 — none found. Saurabh Pant asked the same financing question (turn 9) and got a direct answer on cost per megawatt and the funding tools (turns 10–11).
Thread to pull: PROPWR contract pricing, duration and returns was avoided twice in one call by two different analysts. That is one quarter, so it is a note and not yet a finding under §3.1's two-or-three-quarter rule — check it on the Q3 2026 call, and note that it bears directly on the S7 valuation table, whose EBITDA-per-megawatt input the 2026-09-23 pass had to mark [background] because no public figure could be located. The one number that would replace that assumption is the number being declined.
Verdict
reliable. 8 rows: 3 beat, 3 met, 1 missed, 1 withdrawn — misses and withdrawals are 2 of 8, well short of the half that would make it over-promises, and short of the "nearly every" that would make it conservative. Two qualifications the owner should carry:
- The record splits by business. Every beat and meet is about the completions business or the balance sheet — spending less, running the fleets promised, signing the financing facility. Both broken rows are about the new business: the fifth FORCE fleet's date, and the fleet buyouts. The capex guide that was raised 35% is PROPWR's.
- Half the scored record belongs to a CFO who left, and the promises the current team has made are younger and mostly not yet scoreable. The verdict is not new team — thin record, because Sledge has been CEO throughout and made rows 5 and 6 himself; but the CFO seat has turned over twice inside the window and turns over again in accounting terms on 2026-10-30.
Because the verdict is reliable and not over-promises, the S7 valuation table's base case is not haircut for management's guide (§25). It remains haircut in a different way, which the 2026-09-23 pass already wrote: its EBITDA-per-megawatt input is this pipeline's assumption, not a company disclosure.
S2 — capital allocation and returns (2021–2025)
All figures from SEC XBRL companyfacts, cik 1680247, as filed. The FY2023–FY2025 columns come from the FY2025 10-K (0001680247-26-000028); earlier years from the 10-K of the year concerned.
Where five years of operating cash flow went
| $m | 2021 | 2022 | 2023 | 2024 | 2025 | 5y total |
|---|---|---|---|---|---|---|
| Operating cash flow | 154.7 | 300.4 | 374.7 | 252.3 | 231.6 | 1,313.8 |
| Capital expenditure (cash paid) | (143.5) | (319.7) | (370.9) | (140.3) | (186.3) | (1,160.7) |
| Acquisitions, cash paid | — | (38.6) | (22.2) | (21.0) | — | (81.9) |
| Buybacks | — | — | (51.7) | (59.1) | — | (110.8) |
| Dividends | — | — | — | — | — | — |
| Long-term debt raised | — | 30.0 | — | — | n/d | |
| Long-term debt repaid | — | — | — | — | (3.6) | |
| Stock-based compensation (non-cash, for scale) | 11.5 | 21.9 | 14.5 | 17.3 | 16.9 | 82.1 |
Uses total $1,353.4m against $1,313.8m of operating cash flow: PUMP spent everything it generated and a little more, and it has paid no dividend in its history.
Did the buybacks more than offset the compensation?
No, and then the question was overtaken.
- Diluted weighted-average shares: 102.7m (2021) → 105.4m (2025), +2.7%. Five years, $110.8m of buybacks, and the diluted count is higher than when it started.
- Shares outstanding: 114.5m (2022) → 109.5m (2023) → 103.0m (2024) → 104.3m (2025).
- Then 122.8m issued at 2026-06-30 — +18% in six months, after the 15,000,000-share offering at $10.00 in January 2026 and the $690m zero-coupon convertible in May 2026 (2026-09-23 dossier).
Buybacks at what price
The FY2025 10-K discloses no share count for the programme, only that "During the year ended December 31, 2025, the Company made no share repurchases" and that $89.2m remained authorized of the $200m programme — which confirms $110.8m spent in 2023–2024. The only average price the filings print is the Item 5 issuer-purchases table in the FY2024 10-K: Q4 2024, 423,476 shares at an average of $7.98, the last purchases the programme made.
A blended price has to be inferred, and the inference is bounded rather than exact: shares outstanding fell 11.5m from 2022 to 2024 while stock compensation was issuing shares, so gross repurchases were at least 11.5m and the blended price is at most $110.8m ÷ 11.5m ≈ $9.6, probably lower. Against $10.03 at 2026-09-22 that is roughly break-even — and eleven months after the last repurchase the company sold 15,000,000 new shares at $10.00. It bought back ~11.5m shares at about $9.6 or less and reissued more than that at $10.00.
What the acquisitions were worth
$81.9m of cash acquisitions in 2022–2024 (AquaProp, Silvertip and the wireline assets) produced:
- a $23.6m goodwill impairment in 2024, taken against the wireline segment (FY2024 8-K Ex. 99.1 footnote 2: "Represents noncash impairment of goodwill in our wireline operating segment");
- contingent-consideration reversals of −$2.6m (2024) and −$4.9m (2025) running back through the adjusted-EBITDA bridge, which is the earnouts not being earned.
Separately, and much larger: $188.6m of property and equipment impairment in 2024, the conventional Tier II diesel fleet, written off inside the hydraulic-fracturing segment. Add the two and $212.2m of the capital deployed in this window was written off in a single year.
Return on the capital
Invested capital taken as equity plus long-term debt less cash; NOPAT as operating income at a 21% statutory rate. Both endpoints stated, as §25 asks.
| 2021 | 2022 | 2023 | 2024 | 2025 | |
|---|---|---|---|---|---|
| Operating income ($m) | (68.7) | (2.6) | 130.3 | (167.0) | 6.4 |
| NOPAT at 21% ($m) | (54.3) | (2.0) | 103.0 | (131.9) | 5.0 |
| Invested capital ($m) | 714.4 | 905.2 | 1,010.0 | 810.8 | 844.1 |
| ROIC | (7.6)% | (0.2)% | 10.2% | (16.3)% | 0.6% |
Five-year average ≈ −2.7%, and the single positive year is 10.2%. The incremental return is reported here as an arithmetic fact rather than a signal: NOPAT moved from −$54.3m to +$5.0m while invested capital moved from $714.4m to $844.1m, which computes to +46% — but the base year is a loss year and the denominator is flat only because $212.2m of the assets bought with this capital were impaired in 2024. The honest summary is that $1.24bn of capital expenditure and acquisitions over five years left invested capital roughly where it started and operating income at approximately zero.
Verdict: destroys
Five years, $1.16bn of capex and $82m of acquisitions, average ROIC about −3% and never above 10.2%; $212.2m impaired in one year; $110.8m of buybacks at ~$9.6 or below followed by an equity issue at $10.00 and an 18% increase in the share count in six months; no dividend ever.
Two things this verdict is not saying. It is not a kill: S2's kill test stands as the 2026-09-23 pass wrote it — the company is pre-funded and survives the next twelve months on $905m of liquidity. And it is not a judgement on PROPWR, which has consumed $198.4m of capital expenditure incurred in 2025 and $118.4m in H1 2026 and has produced −$11.6m of adjusted EBITDA in 2025 and −$6.0m in H1 2026; that capital is too young to have a return. What the record says is that the management team now allocating $525–595m a year has a five-year history of allocating capital at a negative return, and that is the context in which the PROPWR programme is being financed. Per §25 a capital destroyer is not a kill on its own: it gets a ⚠️ line under S2 and a kill criterion, both written into the scorecard by this run.
Earnings quality
Written as §25 requires: each finding with its locator; a clean result says so. The non-GAAP bridges compared are the FY2023 column in the 8-K Ex. 99.1 of 2025-02-19, and the FY2024 and FY2025 columns in the 8-K Ex. 99.1 of 2026-02-18.
Four flags.
- The headline free-cash-flow measure has been redefined twice in three years, each time to exclude the cash outflow of the moment. The 2025-02-19 release reconciles to "Free Cash Flow adjusted for Acquisition Consideration" — free cash flow excluding cash paid for acquisitions, in the years the company was buying businesses. The 2026-02-18 release reconciles instead to "Free Cash Flow for Completions Business" — free cash flow with PROPWR's operating and investing cash flows added back. For FY2025 that is $190.2m against group free cash flow of $81.8m: the headlined number is 2.3x the company's actual free cash flow and it excludes precisely the business consuming the capital. Both definitions are disclosed and reconciled in the same exhibit; neither is hidden. The flag is the pattern.
- A "nonrecurring" add-back present in three years of three. Retention bonus and severance expense is added back at $2,297k (2023), $2,324k (2024), $2,633k (2025) — steady and rising. Other general and administrative expense, net ("nonrecurring professional fees") is likewise in all three years: $2,969k, $1,782k, $339k.
- Equipment failures moved into depreciation, where the EBITDA add-back already lives. The FY2025 release's footnote 1 states: "The write-offs of remaining book value of prematurely failed power ends and other components are recorded as depreciation in 2025. In order to conform to current period presentation, we have reclassified the corresponding amounts of $12.4 million from loss on disposal of assets to depreciation for the year ended December 31, 2024." The same year's FY2024 column is therefore restated from a $7,451k loss on disposal (2025-02-19 release) to a $4,925k gain (2026-02-18 release). Adjusted EBITDA is unchanged either way — both lines are add-backs — but recurring write-offs of prematurely failed equipment now sit inside D&A, which is where a reader stops looking. Loss/(gain) on disposal itself is in all three years and was $73.0m in 2023.
- The capital-expenditure figure management guides on is not cash. PUMP guides capital expenditures incurred: FY2025 $281.2m incurred against $186.3m paid, a $94.9m gap the Q3 2025 call attributes to "PROPWR-related CapEx that has been financed and paid directly by the financing partner and unpaid CapEx included in accounts payable and accrued liabilities". The gap is disclosed in every release; a reader who takes the guided number as cash is out by a third.
Clean, with the number:
- Stock-based compensation is small. $16.9m in 2025 = 1.3% of revenue (1.2% in 2024, 0.9% in
- and 20.7% of group free cash flow. Free cash flow after SBC is 79% of reported — comfortably above the half that §25 makes a flag.
- Cash conversion is conservative, not aggressive. Operating cash flow against adjusted EBITDA: 93% (2023), 89% (2024), 111% (2025). Accruals ((net income − operating cash flow) ÷ average total assets) are −20.5%, −28.9%, −18.4% — strongly negative in all three years, the direction that means reported profit lags cash, not leads it. Operating cash flow against net income is not meaningful here: net income was $0.8m in 2025 against $174.9m of depreciation.
- Inventory is falling faster than revenue: $17.7m → $16.2m → $13.3m against revenue of $1,630m → $1,444m → $1,269m.
One watch item, short of a flag. Days sales outstanding: 53.1 (2023) → 49.5 (2024) → 57.7 (2025) on receivables of $237.0m → $196.0m → $200.8m. Receivables rose in dollars in 2025 while revenue fell 12%. §25's test is two periods running and 2024 fell, so this is a note — but it is the line to watch on the Q3 2026 release, because the 2026 revenue line is falling faster still (H1 2026 revenue −16% year on year).
EQ summary line for the scorecard: EQ: 4 flags — headline FCF measure redefined twice in three years (completions-only FCF $190m vs group $82m), severance added back 3 of 3 years, failed-equipment write-offs reclassified into D&A, guided capex is "incurred" not cash ($281m vs $186m). SBC 1.3% of revenue and accruals −18% of assets are clean. DSO +8 days in 2025 — watch.
What this run changed on the scorecard
- S8 ⬜ → ✅ reliable — 8 rows: 3 beat, 3 met, 1 missed, 1 withdrawn; current team made 4 of 8 (CFO seat turned over twice in the window).
- S2 keeps its ⚠️ glyph — the kill test verdict is unchanged — and gains the capital-allocation table and the verdict destroys, with a new kill criterion.
- S3 keeps its ✅ (gate-minimum) glyph and gains the
EQ:summary above. - Dodge log opened, three entries, all FY2026-Q2.
- S7 valuation base case is not haircut: S8's verdict is reliable, not over-promises.
- ⚠️ failed-gates and ⏳ open lines updated.
Full notes, 23 Sep 2026 12 min read
PUMP — dossier, 2026-09-23
Slot names-insurance-2, third member, taken as a Tier 0 promotion (priority 2, queue.json, for the 8-K filed 2026-09-21) in place of LNC. Model-drafted unattended by the cloud research routine. Nothing here is human-verified. Every figure carries the document it was read from; anything from background knowledge is marked [background]. Quotes were located by exact substring search in a text extraction of the filing and copied out — never re-typed.
This name has no cluster pass. cluster-energy-materials-industrial is still pending in schedule.yaml, so S0, S4, S6 and S7 are ⬜ and were not improvised — the same position AX was in on 2026-09-22. There is no Stage 0 lens, no industry map and no scuttlebutt on ProPetro anywhere in this repo.
1. Why it was promoted, and what the promotion actually caught
The Tier 0 sweep promoted PUMP on an 8-K filed 2026-09-21. That filing is an Item 5.02:
"Celina Davila notified ProPetro Holding Corp. (the “Company”) of her intention to resign from her position as Chief Accounting Officer and principal accounting officer of the Company, effective October 30, 2026" — 8-K filed 2026-09-21 (event 2026-09-16), Item 5.02
with the CFO stepping in: "the Company's existing Chief Financial Officer, Caleb Weatherl, will serve as interim principal accounting officer". The filing states the resignation was not the result of any disagreement with the Company on any matter relating to the Company's operations, policies, or practices, including any matters concerning the Company's accounting principles or practices, financial statement disclosure, or internal control over financial reporting — the full statutory denial, which is what a clean departure looks like.
The next day, 2026-09-22, a second 8-K landed, and it is the more interesting one. Reading them together is the value the promotion produced.
2. The PROPWR contract arithmetic
The 2026-09-22 Item 7.01 announces that PROPWR has contracted approximately 230 MW to a subsidiary of Targa Resources — behind-the-meter power for Permian gas processing, "with full deployment expected in early 2028". Then, in the same paragraph:
"total capacity committed under contract for PROPWR now stands at approximately 510 MW" — 8-K filed 2026-09-22, Exhibit 99.1, first paragraph
followed by the qualifier that this also reflects previously announced oil and gas power capacity that is no longer committed under contract, and by "This recontracting allows PROPWR to redeploy the associated capacity to Targa" and to free megawatts for potential data centre deployments in 2027 and beyond.
The baseline, from the Q2 release eight weeks earlier:
"This brings total capacity committed under contract to approximately 350 megawatts." — 8-K filed 2026-07-29, Exhibit 99.1, Recent PROPWR Highlights
350 + 230 = 580. The company reports 510. On the company's own two disclosed numbers, roughly 70 MW — about a fifth of the pre-existing book — came off contract, and the figure is larger if any of the "over 100 megawatts" then in advanced oil-and-gas negotiations was signed in the interim. The release attaches no number to it. The headline is the gross addition.
Caveat, stated plainly: this is arithmetic on two press-release figures, each of which is "approximately". It is not a reconciliation, and the company has not published one. The scorecard's next action is to demand one at the Q3 release.
Two further facts make the gap matter:
- Deployment is in early 2028. The 230 MW produces no revenue for roughly five quarters after this one.
- The equipment is already bought. On the Q2 call the CFO said, defending the capital plan: "we've already ordered or have delivered 1.1 gigawatts of equipment". Against ~510 MW under contract that is roughly 590 MW with no customer, and at management's own guidance of $1.4-1.5m per megawatt inclusive of balance of plant, about $825-885m of capital.
And the reason to watch the megawatt number specifically, from the proxy:
"included Adjusted EBITDA, FCF, PROPWR committed megawatts, and a Sustainability and Safety metric" — DEF 14A filed 2026-04-08, CD&A, 2025 Annual Bonus Plan metrics
90% of each named executive's 2025 annual incentive was tied to quantitative metrics, up from 80% in 2024, and committed megawatts is one of the four.
3. The largest customer sold its shares and is letting the contract run out
From Note 10 of the Q2 10-Q:
- ExxonMobil held 16,600,000 shares — 13.54% on the 2026 proxy's 5% holders table — inherited through Pioneer's 2024 merger. It sold the entire position in May 2026: "Following such sale, ExxonMobil no longer holds shares of the Company and is no longer a related party." A Schedule 13D/A was filed 2026-05-20 for an event date of 2026-05-18.
- The XTO sub-agreement covering two committed FORCE electric fleets (with an option for a third): > "We expect this agreement will expire in late 2026. At this time, we do not expect such agreement > to be renewed or extended" > — Q2 2026 10-Q, Note 10, ExxonMobil and Pioneer and the 10-Q adds that if ProPetro cannot procure additional XTO work it must redeploy the affected fleets, with an adverse effect until it does.
- Revenue from ExxonMobil (including Pioneer and XTO) was $39.3m in Q2 2026 and $111.2m over H1, against $71.2m and $144.2m a year earlier — −45% in the quarter.
Set against management's FY2025 risk summary, which lists "Reliance upon a few large customers may adversely affect our revenue and operating results." as the first risk under Customers, Suppliers and Competition, this is that risk running its course in public.
Share supply. ExxonMobil's 16.6m-share block came to market in May, eight weeks after the January offering of 15,000,000 shares at $10.00 (upsized from 12,500,000, Goldman Sachs lead book-runner, off a Form S-3, proceeds earmarked for growth capital for additional power generation equipment). That is roughly 31.6m shares of supply in five months against 122,823,917 shares issued at 2026-06-30. The stock closed $10.03 on 2026-09-22 — level with the January offering price eight months later.
4. Kill test (S2) — the numbers, which are ugly and survivable
Segment detail from the Q2 2026 release; balance sheet from the Q2 10-Q.
| Six months ended 30 June ($m) | 2026 | 2025 |
|---|---|---|
| Service revenue, total | 576.5 | 685.6 |
| — Hydraulic fracturing | 386.6 | 515.1 |
| — Wireline | 119.3 | 101.4 |
| — Cementing | 59.8 | 69.1 |
| — Power generation | 11.5 | — |
| Adjusted EBITDA, total | 81.2 | 122.3 |
| — Power generation | (6.0) | (2.9) |
| Capital expenditures incurred | 155.7 | 111.7 |
| — Power generation | 118.4 | 60.9 |
| Free cash flow | (30.4) | 40.4 |
Quarterly: revenue $306m in Q2 (from $271m in Q1), net loss $8.1m (from $3.6m), adjusted EBITDA $44.8m (from $36.4m).
Balance sheet at 2026-06-30: cash $784m, total liquidity $905m (including $121m of undrawn ABL, no ABL borrowings), long-term debt $764.9m plus $21.4m current maturities plus $10.9m interim debt, shareholders' equity $957.4m, total assets $2,059.8m. Debt was ~$122m at 2025-12-31 — the $690m convertible (0% coupon, May 2026, no dilution below $29.49 after the capped call) is the difference. Caterpillar Financial borrowings $130m, facility upsized to $167m.
FY2026 capital expenditure guidance $525-595m, of which $400-450m PROPWR and $125-145m completions. That is roughly five times annualised H1 EBITDA.
Verdict: ProPetro does not have a solvency problem in the next twelve months — it raised approximately $1.5bn over eighteen months and is spending it on a fleet it has largely ordered. It has a return problem: whether 1.1 GW of equipment finds contracts at the assumed economics before the completions business stops funding the gap. The share count is the pressure gauge: 104.3m issued at 2025-12-31, 122.8m at 2026-06-30 (+18% in six months), with the converts above.
5. Filings read, and the red-flag sweep (S3)
Three biggest risks, in management's own ordering, FY2025 10-K Risk Factor Summary, p.iii, under "Risks Inherent in Our Business and Industry":
- "Our business and financial performance depends on the historically cyclical oil and natural gas industry and particularly on the level of capital spending of E&P companies within the United States and in the Permian Basin..."
- "Many of our power generation services involve long sales cycles."
- "Our PROPWR SM business line exposes us to the significant risks and uncertainties associated with establishment of a new line of business, and such business line may not achieve the results we anticipate."
Two of the three are the new segment. Also on the list, and live this week: "We rely on a few key employees whose absence or loss could adversely affect our business."
Auditor: RSM US LLP, PCAOB ID 49, report dated 2026-02-19, ICFR opined effective at 2025-12-31. The sole critical audit matter is the power generation arrangements — the complexity of management's evaluation of the contracts, whether an embedded lease conveyed the right to control the use of an identified asset, and the estimation of standalone selling prices for non-lease components. The one judgement the auditor singled out is the accounting for the business the Chief Accounting Officer is leaving in the middle of.
Red-flag sweep (efts.sec.gov full text, CIK 0001680247, 2023-01-01 to 2026-09-23):
| Query | Hits | What they are |
|---|---|---|
"changes in accountants" | 0 | — |
"non-reliance" | 1 | a May 2026 8-K exhibit (financing documentation), not a non-reliance determination |
"material weakness" | 10 | 10-K/ARS boilerplate and the Section 404 risk factor |
"restatement" | 24 | clawback-policy exhibits and S-8 plan documents |
"going concern" | 5 | credit-agreement covenant language in 8-K exhibits |
⚠️ The filing index does carry NT filings: an NT 10-K at 2024-03-01 (the FY2023 10-K was then filed 2024-03-13, so a Rule 12b-25 extension was used and met), plus NT 10-Q and NT 10-K filings in 2019-2020. There is no 10-K/A or 10-Q/A anywhere in the index. Recorded as history, not as a live finding — but a name with a late-filing record, a finance-team change and a CAM on its newest accounting deserves the note.
Also: the common stock was registered on NYSE Texas on 2026-08-20 (Form 8-A12B plus an exchange certification). Allstate did the same thing in 2026. Noise, recorded once.
6. Ownership (S5), and the short position
3spread, 13F report date 2026-06-30 (threespread.py holders PUMP; 179 filers against 253 — the quarter is incompletely ingested and no exit may be read from an absence).
| Manager | Shares | $m | QoQ |
|---|---|---|---|
| BlackRock | 9,105,609 | 130.6 | +534,291 |
| Sourcerock Group | 7,713,646 | 110.6 | +471,833 |
| D. E. Shaw | 6,034,264 | 86.5 | +1,290,711 |
| Encompass Capital | 4,574,591 | 65.6 | NEW |
| Cooper Creek Partners | 4,301,575 | 61.7 | +1,335,014 |
| Valiant Capital | 4,242,475 | 60.8 | +199,482 |
| VR Advisory Services | 4,097,401 | 58.8 | +1,632,488 |
| State Street | 3,488,719 | 50.0 | +433,995 |
| Citadel Advisors | 3,209,545 | 46.0 | +1,678,364 |
| Millennium Management | 3,080,656 | 44.2 | +1,527,787 |
| Wellington | 2,290,361 | 32.8 | +2,057,195 |
| Assenagon | 2,608,110 | 37.4 | NEW |
Almost every line is an add, and the adds are event-driven and multi-strategy money, not long-only value. That is what a register looks like when a 13.5% strategic block clears and the story changes from Permian frac to contracted power.
On the other side: short interest 16,221,951 shares at the 2026-08-31 settlement, up 1,069,516 from 2026-08-14, 5.4 days to cover. On 122.8m shares outstanding that is 13.2% — against ALL's 2.4% and AIG's 1.7% in the same slot. This is a genuinely contested security. (Equibles also returns a model estimate of ~15.9m shares for the 2026-09-15 settlement, explicitly flagged by the tool as near a coin flip on direction; it is a prediction, not a FINRA figure, and is not used above.)
Insiders: 13 directors and executive officers hold 1,894,695 shares, about 1.5% (DEF 14A 2026-04-08). The 5% table on that proxy shows ExxonMobil at 16,600,000 / 13.54% and Sourcerock at 6,846,650 / 5.58%; the first of those is now gone.
The Sixth Street row is not a holding. 2,086,479 shares reported as NEW, byte-identical to Charles Schwab Investment Management's — the defect documented in ledger/research/README.md, now at a seventh issuer after QCOM, ACGL, KNSL, AX, ALL and AIG. Seven for seven on every SEC name this repo has run the tool against since 2026-09-22.
7. What management said, and what the tape did
From the Q2 call (2026-07-29), turns 9-20 of 80:
- On liquidity, the CFO: whole-year capex of $525-595m, with the expanded Caterpillar facility and completions free cash flow covering "a significant portion", against $905m of liquidity — so cash "exceeds the cash we need for the CapEx this year for what we've announced by hundreds of millions of dollars". Approximately $1.5bn raised over eighteen months. Cost per megawatt guidance of $1.4-1.5m unchanged, and "we've already ordered or have delivered 1.1 gigawatts of equipment".
- On the trajectory, the PROPWR lead: "350 megawatts today, headed towards 450 very soon, and several hundred megawatts at the data center", filling "almost all of, you know, middle of 28, maybe all of 28".
- On data centres, the CEO: the large long-term deals "have taken a few twists and turns that were unexpected", the counterparties are still engaged, and he still expects the overwhelming majority of future capacity to end up on data centre sites. One 60 MW hyperscaler deployment is live and operating.
So the September announcement lands against a July trajectory of 350 → 450 → "several hundred" data centre megawatts, and what arrived was 230 MW of midstream capacity plus a recontracting that took the net to 510. No data centre contract has been announced.
The tape. PUMP closed $12.22 on 2026-09-09 and $10.03 on 2026-09-22 — −18% in nine sessions, with the decline beginning before either September 8-K. On the Targa announcement day itself the stock fell from $10.16 to $10.03 on 4.4m shares. A 230 MW award from an investment-grade counterparty produced no positive reaction. This dossier records the sequence and does not claim a cause; the broader oilfield-services tape was not examined.
8. Evidence that is thin, and access limits
- No cluster pass, so no S0, S4, S6 or S7.
cluster-energy-materials-industrialispending. There is no Stage 0 score, no value chain, no scuttlebutt and no thesis for this name in this repo. The kill criteria in the scorecard were written from this pass's evidence because a promoted name needs them, but they are not a substitute for the cluster's thesis. - No scuttlebutt of any kind. No WebSearch was spent on PUMP; no analyst-coverage count, no price targets, no channel checks, no Glassdoor.
- One transcript, and only turns 9-20 of 80. The rest of the Q2 Q&A — including anything said about the ExxonMobil cliff or the oil-and-gas recontracting — was not reached. The Q1 2026 call was not read.
- EBITDA per megawatt in the S8 model is this pass's assumption [background], not a company disclosure. Management guides to cost per megawatt and refers to "earnings per megawatt" without a figure this run located. It is the most load-bearing number in the model and it is not sourced.
- The 70 MW figure is arithmetic on two "approximately" press-release numbers, not a reconciliation. The company has published none.
- The 13D/A filer was not confirmed. The 2026-05-20 Schedule 13D/A on PUMP is dated to the same window as ExxonMobil's disclosed sale and the primary document was not opened this run, so the edge in
relationships.yamlcites the filing, not the filer's identity. - Form 4 transaction codes not itemised; no insider buy/sell direction is asserted, including for the departing CAO.
- No investor presentation read, though Exhibit 99.2 of each results 8-K is one and it is free on EDGAR. The "earnings per megawatt" figure is most likely in it. That is the cheapest gap to close next time and it costs nothing.
- Next-earnings date is derived, not scheduled — late October 2026, from the 2025-10-29 and 2026-07-29 results 8-Ks.