Chevron Corporation (CVX) — Investment Tree v1

✅ Update 2026-09-06 — the reserve table has been read directly (Tier A). The FY2025 10-K (accession 0000093410-26-000078) confirms 10.6B boe · 158% · 91%/95% verbatim, so the ~7.1-year reserve life stands on disclosed rather than derived data. Newly surfaced: proved undeveloped is 27-31% of total, making the developed reserve life only ~4.9-5.2 years; and Hess's ~13% share was independently audited by DeGolyer and MacNaughton with an unqualified opinion. The 16/25 band does not move — verification confirmed the numbers rather than overturning them.

## ⚠️ LONG-TERM UNSUITABILITY WARNING — READ BEFORE THE ESSAY Chevron scored 16/25 on the long-term durability test — one point below the 17/25 threshold this system requires before a ticker is treated as a candidate for a 5-10 year hold. 16/25 falls in the band "12-16 — Low durability: short-term opportunistic only." For a 5-10 year holder that means: this tree does not support treating CVX as a core long-term position. It supports, at most, an opportunistic cyclical position, entry-gated, in a low-single-digit percentage band. Section XII's final verdict says the same thing in the system's own decision vocabulary, and it does not conclude "quality compounder." This build proceeded past a failed gate on explicit owner authorisation, with the warning carried here rather than buried. ### The score's sensitivity — it is boundary-sitting, and the boundary is not settled 16/25 is one point below the threshold, on a number this build could not verify at Tier A. Two specific fragilities: 1. Q2 (moat trajectory) scored 3/5 is arguable. Guyana's Stabroek is genuinely top-decile cost and the AA−/Aa2/AA balance sheet is a real cyclical moat. Scoring Q2 at 4/5 on that basis alone yields 17/25 = "Medium — selective hold." 2. Q5 (reinvestment runway) rests on a Tier-B derived reserve-life figure. The ~7.1-year reserve life index is not company-disclosed — it is computed from a 10.6B boe proved-reserve figure that could not be read from the FY2025 10-K supplemental oil-and-gas table un-truncated, divided by a production rate that changed mid-year with the Hess consolidation. If that numerator is wrong, Q5 moves and the aggregate crosses the threshold. Read the result as "16/25, one point below, on a number we could not verify at Tier A" — not as a settled verdict. The highest-value verification owed on this tree is the FY2025 10-K supplemental oil-and-gas disclosure — the proved-reserves table and Chevron's own reserve-life and replacement figures. Until it is read, the difference between "short-term opportunistic only" and "selective hold" rests on an unverified number. What the score does not say. It is not a solvency or quality warning. Chevron is investment-grade at the top of its sector with 39 consecutive years of dividend increases and zero fatal flags. The score says something narrower: a ~7-year reserve life with sub-100% smoothed replacement makes the 5-10 year case dependent on continued acquisition, and that dependence is not what a "durable compounder" framing implies.

Stage 7 final essay. Bilingual companion: tree_v1_zh.md. Date: 2026-09-06 · Anchor price: $208.60 (2026-09-04 close) · Market cap: ~$410-413B · Shares: ~1.97B† (derived) · Dividend yield: ~3.4% Archetype: a cash-generative major on a seven-year reserve clock — the closest structural analogue is a high-quality royalty stream with a finite, renewable term, not a compounding franchise.

SOURCE QUALITY: Tier A on financials and production (Q2 2026 8-K, FY2025 8-K — both primary SEC filings). Tier B on the reserve position — the FY2025 10-K supplemental oil-and-gas table could not be fetched un-truncated, so proved reserves, both replacement series and the derived reserve-life index are all Tier B. Tier C on the long-run demand backdrop (no agency outlook fetched). Five figures are marked NOT VERIFIED and are listed in §XI. Evidence mode: GENERATE.

† = derived or estimated, not disclosed.

0. Company Fundamentals — what Chevron is and how it earns

An integrated oil major. It finds and produces hydrocarbons (Upstream, ~80% of earnings) and refines, markets and transports them (Downstream, ~20%).

The distinction that governs everything below: Chevron is a depleting-asset business with a dividend commitment. Every barrel sold must be replaced or the asset base shrinks. No other single mechanic explains as much of this analysis.

Tier A — Q2 2026 8-K · FY2025 8-K

† Period-label caveat, NOT VERIFIED. The "Q2 2026" flow column is more consistent with six months ended 2026-06-30 than with a standalone quarter, on three independent checks: segment earnings sum to $13.05B against $15.84B for all of FY2025; CFFO of $19.7B is 58% of a full prior year; and only the half-year reading reconciles with the stated ~18.9× TTM P/E. It changes how the earnings figures should be read. It does not touch the reserve arithmetic — production and realisations are rates, and the derived share count is period-independent.

FY2025's 3,723 MBOED average is not the current base. Hess consolidated only from 2025-07-18. The Q2 2026 figure of 4,070 MBOED is the clean run-rate, and it is the denominator this tree uses.

I. One-sentence verdict

Chevron at $208.60 is a fairly-priced, high-quality cash-return business whose proved asset base does not survive the holding period — and the price already assumes the replacement problem is solved.

II. Company snapshot

III. The five facts that drive everything

1. Proved reserves are ~7.1 years of production. 10.6B boe† against a 4,070 MBOED run-rate. Shorter than the 5-10 year holding period. Tier B, derived.

2. Smoothed reserve replacement has run below 100%. 5-year 91%, 10-year 95%. The headline 158% for 2025 is acquisition-driven — it is Hess, not exploration. Tier B.

3. FY2025 shareholder returns exceeded free cash flow. $27.1B returned against $20.2B adjusted FCF — a $6.9B balance-sheet-funded gap, in a trough year. Tier A.

4. Earnings are a first-order function of price, not operations. FY2025 net income fell ~30% ($17.661B → $12.299B) while production rose. Nothing operational broke. Tier A, two independent primary filings.

5. Lower-carbon capex is under 6% of the budget. ~$1.0B of $18-19B. Chevron is not hedging the transition; it is optimising hydrocarbons and returning cash. Coherent — and it leaves the equity carrying essentially undiluted long-run demand risk. Tier B.

The five read together: a business that generates a great deal of cash from an asset base it is consuming faster than it organically replaces, distributing that cash generously, with no internal hedge against the possibility that the product is less wanted in fifteen years.

IV. The H-0 thesis

At $208.60, Chevron is a high-quality cash-return vehicle whose 5-10-year equity case does not rest on the asset base it owns. The proved reserve position — 10.6B boe against a 4,070 MBOED run-rate, a derived ~7.1-year reserve life — is shorter than the holding period, and reserve replacement smoothed over five and ten years has run at 91% and 95%. The 158% print for 2025 was acquisition-driven. A holder is therefore underwriting management's continued ability to buy or find replacement barrels at acceptable returns — not the barrels currently on the books. And at ~14.2× the $28-30B adjusted FCF management guides for 2030, the market already prices that ability as demonstrated.

H-0 is deliberately two-clause, and this matters for reading every verdict below:

• Clause (a), positive: "high-quality cash-return vehicle." Supported by the AA− balance sheet, the sub-$50 breakeven, 39 years of dividend increases, and Tengiz delivered past nameplate.
• Clause (b), sceptical: "the case rests on replacement, and replacement is already priced." Supported by the reserve clock, the 91%/95% series, Guyana having been bought rather than found, and ~14.2× the 2030 plan.

A leaf confirming Chevron's balance-sheet quality supports H-0. A leaf confirming the reserve clock also supports H-0. H-0 is not "CVX is a good long-term hold" — it is a claim about what the holder is actually buying. That is why h0 can sit in the supported band on a ticker that failed its durability gate.

h0 derivation

h0 = Σ(tier_weight × verdict_value) ÷ Σ(tier_weight) over terminal leaves. Weights: crux ×5 · load-bearing ×3 · supporting ×1. Values: 强力支持 0.80 · 支持 0.625 · 部分支持 0.45 · 证据偏反 0.225 · 已证伪 0. 证据不足 excluded from both sides.

h0 = 20.700 ÷ 36 = 0.5750  →  58%      (band 55-74%: "supported")

Evidence coverage: 3 of 39 total tier-weight excluded as 证据不足 = 7.7%, far under the one-third threshold. No low-coverage flag.

Correlation discount. Three shared drivers are declared. Collapsing reserve-life-estimate-10.6b-boe (L1A 1.1 + L1E 5.3) gives 20.25 ÷ 35 = 57.9%; collapsing realised-crude-price (L1D 4.1 + L1B 2.2 + L1C 3.2) gives 20.075 ÷ 35 = 57.4%; both together give 19.625 ÷ 34 = 57.7%. The discount does not move the band — every variant rounds to 58%.

Crux sensitivity. Forcing L1A 1.1 to 已证伪 gives (20.700 − 2.250) ÷ 36 = 51%; forcing L1B 2.2 gives (20.700 − 3.125) ÷ 36 = 49%. Both cross the 55% boundary, which is what makes them cruxes rather than merely important.

V. Investment tree (condensed)

H-0: a cash-return vehicle whose case rests on replacement — and replacement is priced          58%
│
├── L1A  Can Chevron keep the barrels?                                              支持
│   │    (depletion accounting × Ansoff replacement matrix)
│   ├── 1.1  CRUX  Proved reserve life ~7.1 yrs < hold horizon              ⚠️B  部分支持  ×5
│   ├── 1.2         Smoothed replacement 91%/95%, below 100%                ✅B  支持      ×3
│   └── 1.3         Forward runway is acquisition-dependent, not organic    ✅B  支持      ×3
│
├── L1B  What does the cash do, and is the payout funded?                           支持
│   │    (capital-allocation cascade × payout-funding test)
│   ├── 2.1         FY2025 returned $27.1B against $20.2B adjusted FCF      ✅A  支持      ×3
│   ├── 2.2  CRUX  Dividend survives Brent $55 without stress               ✅A  支持      ×5
│   └── 2.3         Repurchases at $208.60 are accretive                    ⛔B  证据偏反  ×1
│
├── L1C  Are the named growth assets delivering?                             部分支持
│   │    (project S-curve × critical-path milestones)
│   ├── 3.1         Guyana FPSOs 7 and 8 on the announced schedule          ⊗B  证据不足  (excl.)
│   ├── 3.2         Tengiz FCF to Chevron reaches ~$6B in 2026              ✅B  支持      ×1
│   └── 3.3         Permian plateau is a structural harvest decision        ✅B  支持      ×3
│
├── L1D  What does the price environment do to it?                          部分支持
│   │    (price-taker sensitivity decomposition × scenario grid)
│   ├── 4.1         Earnings are price-driven, not operational              ✅A  强力支持  ×3
│   ├── 4.2         Downstream materially dampens Upstream cyclicality      ⚠️B  部分支持  ×1
│   └── 4.3         Balance sheet converted to counter-cyclical buying      ⚠️B  部分支持  ×1
│
└── L1E  Is the product still bought in ten years?                          部分支持
    │    (substitution S-curve × sectoral demand decomposition)
    ├── 5.1         Oil demand ~100 Mb/d through 2035                       ⚠️B  部分支持  ×3
    ├── 5.2         Sub-6% lower-carbon capex leaves risk undiluted         ✅B  支持      ×3
    └── 5.3  STEEL  Short reserve clock insulates from transition           ⚠️B  部分支持  ×1

Verdict tally: 8 ✅ · 5 ⚠️ · 1 ⛔ · 0 ✗ · 1 ⊗

Two structural notes on this tree. L1C 3.1 is ⊗ because the test is pending, not because the evidence is weak — Guyana's geology is not in dispute, its delivery cadence simply has not been observed. And L1E 5.3 is a deliberate steelman: it argues the opposite of this tree's own thesis, at full strength, and half-lands. Full leaf detail with evidence, reasoning and falsification conditions is in leaves.md.

VI. Business model taxonomy

Not a compounder-under-review. A compounder reinvests into an expanding base at high incremental returns. Chevron returned $27.1B against $20.2B of adjusted FCF and cut Permian capex per boe ~25%. It is a harvest-and-return business that periodically buys inventory — a different animal with a different failure mode.

Not value-with-overhang. There is no discount to be found. Consensus targets sit on spot; this tree's expected value is within 3.2% of it.

The mispricing, stated precisely

There is no price discount here, and this tree will not manufacture one. Consensus PT $209-221 against $208.60; this tree's expected value $215.26. Two independent methods agree with the market to within 3%.

The claim is about kind, not magnitude: a fair price can encode a wrong shape.

The market prices Chevron's cash-generating capacity as a going concern of indefinite duration, when the disclosed proved base supports roughly 7.1 years. What sits beyond that horizon is not an asset the company owns; it is a capability the company must keep exercising.

Classified as a time-horizon mismatch expressed as an interpretation error: reserve life is public but published once a year in a supplemental table most models never touch, and "158%" is a better sentence than "91% over five years." Both are disclosed; only one describes the trend a decade-holder is exposed to.

And it has no closing catalyst — which makes it a sizing instruction, not a trade idea. It resolves the slow way: through delivery (Guyana lands, replacement recovers, and the market turns out to have been right to pay 14×) or through erosion (replacement stays at ~90% and the multiple compresses with no day on which the news broke).

VII. Valuation scenarios

Horizon: year-end 2029 (~3.3 years). These are scenario-conditional valuations, not price targets — each is inseparable from its conditions.

Shared inputs. Shares ~1.97B†, derived two independent ways that agree: $12,100M ÷ $6.11 = 1,980M and $410-413B ÷ $208.60 = 1,966-1,980M. Band 1.96-1.98B — Tier B, derived, not disclosed. Driver is adjusted free cash flow, chosen because it is the best-sourced forward series available; the forward P/E and EV/EBITDA are both excluded for reasons given in §XI. Base multiple is read off today's price, not assumed: $411.0B ÷ $29B (2030 plan midpoint) = 14.2×. Cumulative dividends 2026 Q4 → 2029 Q4: $24.90 at 4%/yr growth, $23.15 if frozen.

Scenario 1 — Bull ($259.64 + $24.90 = $284.54, +36.4%, 25%)

Brent near $80. Guyana FPSOs 7 and 8 land on schedule. Organic replacement returns above 100% and the FY2026 10-K discloses a longer reserve life — L1A 1.1's falsification fires and durability crosses to 17/25.

$33B 2029 adjusted FCF × 15.5×  =  $511.5B  ÷ 1.97B  =  $259.64/sh  + $24.90 div  =  $284.54

Two independent things must go right: delivery on a cadence nobody has yet observed, and a market willing to pay 15.5× rather than 14.2×. Hence 25%.

Scenario 2 — Base ($198.98 + $24.90 = $223.88, +7.3%, 50%)

Brent at the plan's own $70. The 2030 guidance is delivered roughly on time. Replacement continues near 90-95%. The multiple does not move, because there is nothing to re-rate on.

$28B 2029 adjusted FCF × 14.0×  =  $392.0B  ÷ 1.97B  =  $198.98/sh  + $24.90 div  =  $223.88

The finding hiding in this scenario: management executes, the market keeps paying what it pays today, and the holder earns about 2% a year. That is what "the plan is already in the price" means arithmetically — and it is the successful scenario, not the bearish one.

Scenario 3 — Bear ($105.58 + $23.15 = $128.73, −38.3%, 25%)

Brent $55-60 sustained. Guyana slips a year or more. The Permian declines off its plateau. Buybacks stop; the dividend is held flat, not cut — 39-year record, sub-$50 breakeven.

$16B 2029 adjusted FCF × 13.0×  =  $208.0B  ÷ 1.97B  =  $105.58/sh  + $23.15 div  =  $128.73

Calibration: CVX fell ~59% peak-to-trough in 2020, so a −49% price leg is severe but not exotic. The multiple is deliberately not crushed — 13.0×, not 10× — because a distressed multiple would be inconsistent with AA− and a maintained dividend. This is a cash-flow bear with a mild haircut, not a solvency event.

Expected value and asymmetry

EV  = 0.25(284.54) + 0.50(223.88) + 0.25(128.73) = $215.26     →  +3.2% over ~3.3 yrs (≈ +1.0%/yr)

Asymmetry   upside $75.94 ÷ downside $79.87 = 0.95×            →  negatively skewed
2.0× at     (284.54 − P) = 2(P − 128.73)  →  P = $180.67       →  −13.4% from spot

Market-implied probabilities. Solving backwards from $208.60 with the base weight fixed at 50% gives 21 / 50 / 29 against this tree's 25 / 50 / 25. Four percentage points of bull weight is the entire disagreement — well inside the noise of any three-scenario construction. The honest conclusion is that this tree and the market agree.

Sensitivity. The share-count band moves the base case by ±$1.01 — immaterial. The multiple is nearly the whole answer: 13.0× takes the base total to $209.67 (+0.5%), 12.0× to $195.46 (−6.3%), 15.0× to $238.10 (+14.1%). Note also that 13.0× on $28B and 14.0× on $26B both produce $364.0B — the model cannot distinguish "the market pays less" from "the plan delivers less," and does not pretend to.

VIII. Competitive analysis

The European discount is not a signal for CVX. SHEL at 9.4× and TTE at 8.4× trade where they do substantially because of European fiscal exposure, windfall-tax risk and heavier transition capex — none of which applies to a US-domiciled major spending 5.3% of capex on lower-carbon. Reading "CVX is expensive versus SHEL" would be a category error.

The genuine comparable is XOM, and there the picture inverts: 5.3× against 10.1× EV/EBITDA is a near-2× gap between two integrated US majors on the same metric. NOT VERIFIED and deliberately unresolved — no EBITDA figure was available at any tier, so the gap cannot be decomposed into differing net-debt treatment, differing definitions, or a real spread. No scenario uses it.

The comparison that would actually matter

Multiples answer "what does the market pay for a barrel of integrated-major earnings." They do not answer a duration question. The relevant peer comparison is reserve life and replacement — and peer reserve data was not fetched in this build (NOT VERIFIED).

The consequence is specific and should not be overstated: without the peer distribution, this tree cannot say whether ~7.1 years is unusually short or simply normal. Majors typically run 8-12 years, which would put Chevron at or below the low end — but "typically" there is Tier C training knowledge, not a fetched comparison, and it is not permitted to carry a verdict. The leaves are written accordingly: L1A 1.1 claims reserve life is shorter than the holding period — an absolute statement on Chevron's own numbers — never that it is shorter than peers.

Where Chevron genuinely leads

Two places, both Tier A, neither a multiple: credit quality (AA−/Aa2/AA at 13.1% net debt, sustained while absorbing ~$10.0B of Hess debt and still cutting total debt $40.8B → $37.1B) and the 39-year dividend-growth record, which no name in the table approaches. Both feed the L1B 2.2 crux, and both are why H-0 carries an explicitly positive clause.

IX. Forward triggers and red flags

T1 — FY2026 10-K supplemental oil-and-gas disclosure · the decisive one

Nodes: L1A 1.1 (crux), L1A 1.2, L1E 5.3, H-0 clause (b) · Window ~Feb-Mar 2027 · Source: SEC EDGAR, Chevron 10-K supplemental oil-and-gas disclosure

• Fires up (reserve life ≥10 yrs or 5-yr organic replacement ≥100%): L1A 1.1 moves toward 已证伪, durability crosses 16/25 → 17/25, and a full Stage-4 rerun is required.
• Fires down (life confirmed ≤8 yrs, organic replacement confirmed <100%): L1A 1.1 upgrades ⚠️ → ✅, h0 rises ~2 points, and the durability band is settled rather than provisional.

This is the highest-value observation in the file. Everything else here is secondary.

Catalyst proximity: nothing within 24 weeks. The first meaningful observation (T3/T4) is roughly five months out; T1 follows. And per §VI the duration mispricing has no closing catalyst at all — this list resolves the thesis, not the price.

Red flags

X. Long-term durability assessment

16 / 25 — below the 17/25 threshold. Band 12-16: "Low durability — short-term opportunistic only." Fatal flags: 0.

Full reasoning in durability_test.md. The three things worth carrying into a reading of this essay:

Q5 is where the reserve position lands, and it cuts both ways. Against: a ~7.1-year reserve life with 91%/95% smoothed replacement means the current proved base does not survive the holding period on its own. For: Guyana and Tengiz are a credible, funded decade-long answer — an 8th FPSO targeted by 2030, TCO FCF ramping ~$5B → ~$6B, 2-3%/yr growth guided to $28-30B adjusted FCF by 2030 at $70 Brent, breakeven below $50. Scored 3/5 because the runway exists but is inorganic and finite on current disclosure.

No fatal flag fires. Not chronic ROIC < WACC (Chevron was more disciplined than peers through 2014-2020 and carries the sector's strongest balance sheet). Not near-term disruption >80% (the transition is multi-decade; petrochemical, aviation, marine and heavy-industrial demand is sticky). Not balance-sheet survivability (13.1% net debt, AA−/Aa2/AA).

The score is boundary-sitting, and the boundary is not settled. Scoring Q2 at 4/5 on Guyana's top-decile cost position alone yields 17/25 = Medium. And Q5 rests on a Tier-B derived reserve-life figure the 10-K could not confirm. See the warning at the top of this document — it is not a formality.

Open item raised during this build

durability_test.md Q1 makes a period-incomparable claim. It states that Downstream earnings rose "$3.0B → $4.9B while Upstream fell as realisations dropped." In the only period pair available, realisations rose (US $48.13 → $70.80), and the two periods are of different lengths (§0 period caveat). The integrated hedge is a real industry mechanic — this evidence does not demonstrate it, which is why L1D 4.2 resolves 部分支持 rather than 支持.

The 16/25 aggregate is left exactly as authored. Q1's 4/5 does not turn on the hedge demonstration; it rests on demand persistence and 140+ years of operating history, both of which stand. Recorded as an editorial note for the next reviewer rather than silently re-scored mid-build.

XI. Tracking dashboard

Falsification status: 16 conditions armed · 0 triggered · 0 expired. Nothing is due before 2027-02-28. Full table in dashboard.md.

What xii_score < h0 means here

Unusually for this corpus, the scorecard sits below thesis confidence (47% vs 58%). In most trees the gap runs the other way and names headroom that catalyst-firing could close. Here the inversion says the analysis is more confident than the investment case is attractive — there is no headroom to close, because the tree is confident about a conclusion that does not favour ownership.

Everything marked NOT VERIFIED in this build

1. The FY2025 10-K supplemental oil-and-gas table. Could not be fetched un-truncated. Proved reserves (10.6B boe), both replacement series (91%/95%) and the derived ~7.1-yr reserve life are all Tier B. This is the number that decides the durability band.
2. Whether the 91%/95% series are organic-only or total-including-acquisitions. The ambiguity runs one way: if total, organic replacement is worse.
3. The period label on the "Q2 2026" flow column. Three independent checks say six-month cumulative, not a standalone quarter. Changes how earnings read; does not move the reserve arithmetic.
4. The ~13.2× forward P/E. Implies ~$15.80 EPS and ~$31.1B net income at ~1.97B shares, against FY2025's $12.299B. Not reconcilable at a constant share count. Excluded from every scenario.
5. Peer reserve life and replacement ratios. Not fetched, so this tree cannot say whether ~7.1 years is an outlier or the norm.

Also absent and stated rather than estimated: no ROIC/WACC series (§XII Q13 is scored on the absence of a chronic-destruction record, not a positive finding), no EBITDA figure (the CVX-vs-XOM EV/EBITDA gap is flagged, not decomposed), and no quantified climate-litigation exposure (RF7 exists as an unmodelled risk rather than a leaf).

XII. Investment Scorecard

Format B (15 questions — long-term hold). Per MANUAL_en.md Part K.6. Verdict values: ✅ = 1.0 · ⚠️ = 0.5 · ✗ = 0. Evidence grades suffixed per the M1 convention (A = primary filing, B = verified secondary/guidance, C = training knowledge). Weight tiers: Critical (5×) · Load-bearing (3×) · Important (2×) · Confirming (1×).

Section A — Business quality (Questions 1-5)

Q1 ⚠️B — Partial. The moat is cost position and asset quality, not brand or switching costs. Widening: Guyana's Stabroek is top-decile cost with >11B boe block-wide; the AA− balance sheet is itself a moat in a cyclical industry. Narrowing: the Permian has plateaued five quarters, and no oil producer has pricing power — the moat can never become a spread over the commodity. And the strongest element was bought for ~$53B in stock, not built. Score: 0.5 × 5 = 2.5

Q2 ⚠️B — Partial. CFFO of $33.9B and adjusted FCF of $20.2B in a trough year is genuine cash-generative strength. But FY2025 net income fell ~30% on price alone, and no ROIC or WACC series was obtained at Tier A/B in this build — "high returns on capital" is not established here, only un-contradicted. Score: 0.5 × 5 = 2.5

Q3 ✅B — Yes. 2-3%/yr production growth guided to 2030, funded by the Guyana FPSO programme and Tengiz debottlenecking. The Q2 2026 run-rate of 4,070 MBOED is already well above FY2025's 3,723. Visibility over three years is good. Score: 1.0 × 3 = 3.0

Q4 ⚠️B — Partial. For: 39 consecutive years of dividend increases, best-in-class ratings, and more discipline than peers through the 2014-2020 destruction cycle. Against: FY2025 returned $27.1B against $20.2B adjusted FCF — above 100% of free cash flow, balance-sheet funded — and the ~$53B all-stock Hess deal is too recent to judge on returns and was not counter-cyclically timed. Score: 0.5 × 3 = 1.5

Q5 ✗A — No. A price-taker by construction. FY2025's ~30% earnings decline came entirely from realisations while production rose. This is the tree's only 强力支持 leaf (L1D 4.1), and it resolves this question negatively with two independent primary filings. Score: 0 × 2 = 0.0

Section B — Balance sheet & capital allocation (Questions 6-9)

Q6 ✅A — Yes. Net-debt ratio 13.1%; ratings AA−/Aa2/AA; total debt cut $40.8B → $37.1B while absorbing ~$10.0B assumed from Hess; capex-plus-dividend breakeven guided below $50 Brent. In the bear scenario buybacks stop, but the dividend and the ratings hold. Score: 1.0 × 2 = 2.0

Q7 ⚠️B — Closer to no than to yes. $12.1B of buybacks in FY2025 and $3.1B in the 2026 column, executed at a price equal to ~14.2× the guided 2030 FCF, with consensus targets on top of spot and this tree's asymmetry at 0.95×. The corresponding leaf (L1B 2.3) records ⛔ 证据偏反 — evidence leans against accretion. It is scored ⚠️ rather than ✗ only because the three-value scorecard rubric has no cell for "leans against but not decisive," and 25% bull probability is not a rounding error. Score: 0.5 × 2 = 1.0

Q8 ✅B — Yes. FY2025 CFFO $33.9B against organic capex $17.3B gives FCF $16.6B, adjusted $20.2B, on net income of $12.299B — conversion above 1.0×. Worth naming the mechanism: high FCF-to-net-income conversion in an E&P partly reflects depletion charges running ahead of replacement spend. That is a real strength for cash today and is precisely the accounting shadow of the reserve clock. Score: 1.0 × 1 = 1.0

Q9 ⚠️B — Partial, on the stated test. The dividend alone costs ~$14.03B/yr† ($7.12 × ~1.97B shares) = 69% of FY2025 adjusted FCF and 85% of reported FCF — above the 50% threshold this question sets, in a trough year. Against the guided 2030 band of $28-30B it falls to ~48%. Coverage is not in doubt; the payout ratio at trough earnings is. Score: 0.5 × 1 = 0.5

Section C — Valuation (Questions 10-12)

Q10 ✗B — No. Scenario-weighted expected value $215.26 against $208.60 = +3.2% over ~3.3 years, with 0.95× asymmetry — you risk $1.05 for every $1.00 of upside. Consensus independently agrees at $209-221. The stock is fairly valued, which is a different finding from cheap. Score: 0 × 5 = 0.0

Q11 ✗B — No, because there is no discount to close. The 2030 plan is already in the price at ~14.2×, and §VI establishes that a duration mispricing has no closing catalyst by construction — it resolves through delivery or erosion, not through an announcement. T1 (the 10-K reserves table) resolves the thesis, not the price. Score: 0 × 3 = 0.0

Q12 ✅B — Yes, appropriately priced. 13.2× forward P/E and 5.3× EV/EBITDA sit mid-pack: cheaper than XOM on EV/EBITDA, dearer than the European majors on P/E for defensible jurisdictional reasons, with the sector's longest dividend-growth record. The market has priced this business correctly, and that is the point — Q10 and Q12 are not in tension, they are the same finding read from two sides. Score: 1.0 × 2 = 2.0

Section D — Long-term hold assessment (Questions 13-15)

Q13 ⚠️B — Partial, and the grade reflects evidence quality, not doubt. No ROIC or WACC series was obtained at Tier A/B in this build. What exists: the durability test's Q3 fatal-flag check for chronic ROIC < WACC did not fire, on the basis that Chevron was more disciplined than peers through 2014-2020 and carries the sector's strongest balance sheet. That is an absence of a chronic-destruction record, not a positive return finding, and it is scored as such. Score: 0.5 × 2 = 1.0

Q14 ⚠️B — Partial, and this is the binding question. Guyana and Tengiz are a real, funded runway: an 8th FPSO targeted by 2030, TCO FCF ramping ~$5B → ~$6B, 2-3%/yr growth to $28-30B adjusted FCF at $70 Brent. But with a ~7.1-year reserve life and 5-/10-yr replacement at 91%/95%, the runway depends on continuing to buy or find replacement barrels — a harder promise than reinvesting in an existing base. Score: 0.5 × 2 = 1.0

Q15 ⚠️B — Limited. Guyana's undeveloped resource beyond the sanctioned FPSOs is the only material call option in the portfolio. Lower-carbon ventures at under 6% of capex are not optionality at scale, and there is no adjacent-market sleeve of consequence. Score: 0.5 × 1 = 0.5

K.3.5 Weighted score derivation

Weight distribution: 3 Critical (5×) · 3 Load-bearing (3×) · 6 Important (2×) · 3 Confirming (1×). Theoretical maximum: 3×5 + 3×3 + 6×2 + 3×1 = 15 + 9 + 12 + 3 = 39.

xii_score = 18.5 / 39 = 47%

xii_score: 47% — the "wait or skip" band (45-64%). This value is the raw K.3.5 weighted score with no haircut applied; it matches the INDEX_META block exactly.

Where the score is lost: Q10 and Q11 both resolve ✗ for the same underlying reason — the stock is fairly priced and there is no discount to close — and together they zero out 8 of the 39 available points. Q5 (pricing power) zeroes another 2. That is 26% of the scale surrendered to facts that are not in dispute: a price-taker, fairly valued, with no catalyst.

Scorecard summary

xii_score 47% · h0 58% · durability 16/25 · fatal flags 0 · asymmetry 0.95×

Final verdict

AVOID as a 5-10 year core holding. Hold-with-sizing only as an opportunistic cyclical position — entry-gated, low-single-digit percentage band.

This cannot conclude "quality compounder," and the framework does not permit it to. Chevron failed the durability gate at 16/25 in the band "short-term opportunistic only." Its scorecard sits at 47%, in "wait or skip." Its risk/reward at spot is negatively skewed at 0.95×. Its expected return over 3.3 years is +1.0% a year. And the analysis is more confident than the investment case is attractive — the rare inversion where xii_score < h0.

What Chevron is: an excellent operator of a depleting asset base, with the best balance sheet in its sector and a 39-year dividend record, priced correctly for what it is.

What it is not: an asset you can buy at $208.60 and stop thinking about for a decade. The proved reserves run out inside the holding period. What replaces them has to be bought or found, repeatedly, at acceptable returns — and at ~14.2× the 2030 plan, you are paying as though that were already settled.

Sizing discipline (percentage bands only — this system records no position values):

This is a framework output, not advice to any person. It records what the system's own rules produce from the evidence assembled. The decision is the reader's.

2-minute pitch

Chevron is the best-capitalised integrated oil major in the world — AA−/Aa2/AA, 13.1% net debt, thirty-nine consecutive years of dividend increases, a capex-plus-dividend breakeven below $50 Brent, and two genuinely world-class growth assets in Guyana's Stabroek and a Tengiz project that not only landed but was debottlenecked past its own nameplate. All of that is true, and none of it is the question.

The question is what a ten-year holder is actually buying. Proved reserves are 10.6 billion barrels of oil equivalent against a run-rate of 4,070 thousand barrels a day — roughly seven years of production, shorter than the holding period. Reserve replacement over five and ten years has averaged 91% and 95%; the headline 158% for 2025 is the Hess acquisition, not exploration. So the barrels that produce the cash flow in 2035 are not on the books today. They have to be bought or found, repeatedly, at acceptable returns.

And the market pays roughly fourteen times the $28-30 billion of free cash flow management guides for 2030 — as if that replacement capability were already demonstrated. Consensus targets sit on top of spot at $209-221. This tree's expected value is $215.26. Everyone agrees on the price; the disagreement is about the shape. At $208.60 the risk/reward is 0.95× — you risk slightly more than you stand to make — and there is no catalyst to close a gap that is about duration rather than cheapness.

Current position: watch, 0%. Entry only at $180.67 or below, where the same scenarios produce 2.0× asymmetry. And the single thing that would change this analysis is not the oil price — it is the FY2026 10-K supplemental oil-and-gas table, which would confirm or refute the reserve-life figure the entire case rests on.

Relevant risk types (per K.4)

• Commodity price — the dominant first-order risk. FY2025 proved it in isolation: ~30% earnings decline, production rising, nothing operational broken.
• Depletion / inventory — the risk this tree is built around. A ~7.1-year proved reserve life with sub-100% smoothed replacement makes the terminal value contingent on a capability, not an asset.
• Execution concentration — with the Permian harvesting, essentially all guided growth rests on two assets: one bought in an all-stock deal, one in Kazakhstan.
• Sovereign / jurisdictional (Tengiz) — RF4. Real, and Tier C in this build, so recorded as a red flag rather than allowed to discount a verdict it cannot support.
• Capital-allocation drift — FY2025 returned above 100% of FCF. Defensible once; the pattern is what to watch (RF3).
• Long-run demand / transition — with lower-carbon capex under 6%, the equity carries this essentially undiluted. Multi-decade, not near-term — the disruption fatal flag did not fire.
• Evidence risk — unusually, a first-class risk here. The load-bearing reserve figure is Tier B and derived. The analysis could be wrong because the input is wrong, and that possibility is on the face of this document rather than in a footnote.

When NOT to buy (anti-pattern check, per K.5)

• ✗ Do NOT buy this as a long-term core holding at any price while durability sits below 17/25. The gate exists precisely to stop a strong balance sheet and a long dividend record from being read as long-term durability. They are not the same thing, and here they diverge.
• ✗ Do NOT buy at $208.60 for the asymmetry. 0.95× is below 1.0 — negatively skewed. The threshold for a core position in this system is 2.0×, reached at $180.67.
• ✗ Do NOT buy the dividend without pricing the payout. FY2025 returned $27.1B against $20.2B of adjusted FCF. The dividend is well-covered at the guided 2030 cash flow and was not covered by FY2025's. A 3.4% yield funded partly from the balance sheet in a trough year is not the same instrument as a 3.4% yield funded from operations.
• ✗ Do NOT read 158% reserve replacement as reserve health. It is the Hess acquisition. The five- and ten-year averages — 91% and 95% — are the trend, and they are the numbers a decade-holder is exposed to.
• ✗ Do NOT wait for a catalyst that does not exist. This is a duration mispricing with no closing event. It resolves through delivery or erosion over years, not through an announcement. Sizing and patience are the only available responses.
• ✗ Do NOT treat "fairly valued" as a reason to buy. Two independent methods agree the price is right. A correctly-priced asset with negative skew, no catalyst, and a failed durability gate is a pass, not a hold.