Cheniere Energy Partners, L.P. CQP

Reference table

What stops it being a Buy: the market already knows. The units sit 4.5% below their fifty-two-week high after a 16.2% twelve-month advance, our base case of $70 is 4.5% above spot, and the sell-side is at consensus SELL — 10 Sell, 5 Hold, 3 Buy — with a $69 target. Behind the price sit four structural facts: 3.27x net leverage on one physical asset in one Louisiana parish; total partners' capital of just $414M including a negative $2,742M non-controlling interest; a top line that is dominated by pass-through gas costs (FY22 produced $17,206M of revenue and only $2,498M of net income, while FY23 produced $9,664M and $4,254M); and a limited-partnership structure with a K-1, no entity-level tax and no unitholder voting control. One tracked voice puts the bear case precisely: margins are a function of the gas price and management does not control its inputs. Hold, for the distribution.

1. Structure and data integrity — read this first

The non-equity check

Several tickers in an index feed are bonds, preferreds or trust securities with the operating parent's fundamentals wrongly attached. CQP was tested against those markers and is NOT one of them:

Conclusion: CQP is a genuine operating equity — limited partnership common units — and an equity dive is appropriate. Two of six markers are met, both explained by the asset class (a contracted infrastructure MLP naturally has a low beta and a high payout), and the four decisive markers all fail.

But it must be underwritten as a partnership, not a corporation

Four consequences follow, and they are not cosmetic:

1. Tax. ratios_ttm reports effectiveTaxRateTTM: 0 and taxBurdenTTM: 0.996 — the partnership pays essentially no entity-level income tax. This is correct, not corrupt. Unitholders receive a Schedule K-1, not a 1099, with the attendant filing complexity, state-level filing obligations and potential unrelated-business-taxable-income considerations in tax-advantaged accounts. Suitability must be checked before purchase.

2. Control. Per the profile, "Cheniere Energy Partners GP, LLC functions as the general partner." Unitholders do not elect a board and do not control the entity. The insider records reflect this: the July 2026 filings show two new directors (Zamir Rauf and Michael Jennings) receiving phantom units, a general-partner compensation instrument, not open-market purchases.

3. Capital structure. bal_a FY25 shows total equity of $414M composed of $3,156M of "stockholders' equity" and a negative $2,742M minority interest. For an MLP that is a normal presentation of the subsidiary-partnership structure, not an error, but it makes book-value ratios (price-to-book 11.0x) meaningless. We use no book-value metric in this dive.

4. Distributions, not dividends. The $3.27 is a distribution — a return of partnership cash flow — with a different tax character from a corporate dividend.

Vendor corruption found and rejected

REJECTED — inc_q[0] (quarter ended 2026-03-31) on two fields.

• interestExpense: -181000000. Negative interest expense against every other quarter in the file running +$186M to +$202M on a $14,686M debt stack. A sign flip. Rejected; we use ~$188M per quarter.
• ebitda: 179000000 against operatingIncome: 329000000. EBITDA cannot sit below operating income — depreciation is added, not subtracted. Every other quarter shows the expected relationship (Q4'25: EBITDA $1,646M against operating income $1,464M, a $182M gap; Q3'25: $868M against $591M). Rejected.
• grossProfit: 358000000 on revenue of $3,599M is a 9.9% margin against 30-50% in every other quarter. This is plausible if derivative losses on gas purchase contracts ran through cost of sales — CQP's revenue and cost lines both carry mark-to-market — but it cannot be verified from this file. Treated with caution and not used in any margin computation.
• Consequence: because ratios_ttm and km_ttm are computed from the trailing four quarters, the vendor's trailing EBITDA of ~$3,605M and its netDebtToEBITDATTM of 3.86x both inherit the corrupt quarter. We instead compute leverage on FY25 EBITDA of $4,428M, giving 3.27x, and label it as our computation.

FLAGGED — earn_cal EPS disagrees with inc_q EPS in two of the last four quarters. For the 2025-10-30 report, earn_cal gives epsActual: 0.81 while inc_q for the quarter ended 2025-09-30 gives eps: 1.05. For 2026-02-26, earn_cal gives 2.38 against inc_q's 2.41. The other two agree. We use inc_q throughout and note the discrepancy.

FLAGGED — ratios_ttm price-to-earnings is internally inconsistent. It reports priceToEarningsRatioTTM: 14.105 alongside netIncomePerShareTTM: 5.2107. At $67.00 those cannot both be true ($67.00 ÷ $5.2107 = 12.86x). We use 12.9x and label it as our computation.

EXCLUDED from valuation — the FY29E and FY30E estimate lines. They model revenue leaping from $12,133M (FY28E) to $21,247M (FY29E) and $25,188M (FY30E) — a 108% increase — on two revenue analysts and one EPS analyst, while EPS barely moves ($4.365 → $4.72 → $4.24). Revenue doubling with flat earnings per unit is internally incoherent, and the coverage is far too thin to underwrite. The base case is built on FY26E-FY28E, where coverage is 5-6 analysts.

PASSED — the rest of the est block, and this is worth noting explicitly. In welcome contrast to NTRA, ILMN and TWLO in this same batch, CQP's forecast EBITDA line survives an internal check: FY26E EBITDA of $4,858M sits above FY26E EBIT of $4,118M by $740M of implied depreciation, both are positive, and both are close to the FY25 actuals ($4,428M and $3,585M). The FY26-28 EBITDA series is used with confidence.

GAP — seg_geo is effectively empty. It contains a single row dated FY2017 (Ireland $787M, Korea $666M, Non-US $1,500M, United States $1,400M). There is no current geographic revenue disclosure at all for an export business whose entire economics depend on destination markets. This is a material gap and no claim in this dive rests on customer geography.

GAP — seg_prod stops at FY2024. The most recent row (FY24) shows Liquefied Natural Gas $8,504M (97.7%), Regasification Service $135M, Other $65M. There is no FY25 product split.

2. The asset — what CQP actually owns

Per the profile, CQP operates a natural gas liquefaction and export complex at the Sabine Pass LNG terminal in Cameron Parish, Louisiana, comprising:

• Five LNG storage tanks with combined capacity of roughly 17 billion cubic feet equivalent;
• Two marine berths accommodating vessels up to 266,000 cubic metres;
• Regasification vaporisers rated at roughly 4 billion cubic feet per day;
• A 94-mile pipeline connecting the terminal to interstate pipeline networks.

1,530 employees. Chief executive Jack A. Fusco. Established 2003, headquartered in Houston.

This is one facility. There is no second site, no offsetting geography, no portfolio effect. Cameron Parish sits on the Louisiana Gulf coast. The single most important risk in this dive is not financial — it is that a hurricane, an outage or a site-specific regulatory event affects 100% of the revenue base, and the balance sheet carrying that risk is levered 3.3 times.

Product concentration is equally total. FY24: liquefied natural gas $8,504M of $8,704M — 97.7%. Regasification service $135M and other $65M are rounding.

3. The financials — a toll asset with a commodity overlay

Three observations, in order of importance.

First — and this is the bear case in a single row — compare FY22 and FY23. FY22 revenue was $17,206M and net income was $2,498M. FY23 revenue was 44% lower at $9,664M and net income was 70% higher at $4,254M. Revenue and profit moved in opposite directions by large margins. The explanation is that the top line contains pass-through gas costs and derivative marks: when gas prices spike, revenue balloons but so does the cost of the gas being liquefied, and hedging positions move against the partnership. This is the concrete evidence behind the tracked bearish claim in §6 that "margins are a function of the natural-gas price and management lacks control over its inputs." It is not a theoretical objection; it is visible in the file.

Second — free cash flow is remarkably stable despite that. $3,698M, $2,889M, $2,814M, $2,569M across four years in which revenue ranged from $8,704M to $17,206M. Capital expenditure is trivial ($199M in FY25, 1.8% of revenue, and capexToRevenueTTM of 0.053%), because the asset is built. The cash-generating character of the toll is real even where the accounting profit is noisy.

Third — the trailing figures are contaminated and should be read with care. Trailing revenue of $11,368M and net income of $2,522M include the corrupt March 2026 quarter documented in §1. Trailing EPS of $5.21 (our computation from the four quarterly figures: $0.91 + $1.05 + $2.41 + $0.38) is therefore depressed by an anomalous $0.38 quarter. We use FY25 as the operating reference throughout.

Interest is the largest single expense after gas. FY25 interest expense $753M against operating income of $3,585M — interest consumes 21% of operating profit, and coverage is 8.1 times. FY22-FY25 interest ran $870M, $823M, $800M, $753M — declining slowly as debt is repaid ($15,991M in FY23 to $14,686M in FY25, a $1,305M reduction).

4. The distribution — the actual investment case

• Distribution: $3.27 per unit annually (dividendPerShareTTM, confirmed by lastDividend).
• Yield at $67.00: 4.88% (dividendYieldTTM: 4.881%).
• Coverage, per unit: trailing free cash flow per unit of $6.24 against $3.27 = 1.91x.
• Coverage, partnership level: FY25 free cash flow of $2,569M against distributions of roughly $1,583M (484M units × $3.27) = 1.62x.
• Payout ratio on earnings: 82.2% (dividendPayoutRatioTTM) — but this uses the contaminated trailing EPS; on FY25 EPS of $6.17 it is 53.0%.

Read: the distribution is well covered on every measure and by a wide margin. Even in FY24, the weakest year in the file, free cash flow of $2,814M covered the same distribution 1.78 times. The 4.88% is not a stretched payout; it is a conservative one from a genuine cash asset.

The comparison that frames the whole verdict: our base case implies a capital gain of roughly 4.5%, and the distribution adds 4.88%, for a total return near 9.4% on a 0.292-beta security. That is a respectable income outcome. It is not a capital-appreciation case, and it should not be bought as one.

5. Leverage and the balance sheet

• Net debt to EBITDA: 3.27x on FY25 EBITDA of $4,428M (our computation). The vendor's 3.86x uses the contaminated trailing figure — see §1.
• Interest coverage: 8.1x. Debt service coverage 1.81x.
• Current ratio 0.42x, working capital negative $1,733M. Normal for an infrastructure partnership that distributes its cash rather than accumulating it, but it means there is no liquidity buffer — the model depends on continuous operating cash generation and revolver access.
• Debt to assets 83.1%. Debt to market capitalisation 43.8%.
• Deleveraging is happening, slowly: $1,305M of net debt reduction over two years, roughly $650M a year, against $2,569M of free cash flow — so about a quarter of free cash flow goes to debt reduction and the rest to distributions.

Read: this is a heavily but stably levered infrastructure balance sheet. 3.3x with 8.1x interest coverage on contracted cash flows is within normal bounds for the asset class. The risk is not the ratio; it is what the ratio is secured against — one facility, in one parish, on one coast.

6. Knowledge base — a genuinely two-sided lane, and the reversal matters

A search of 51,928 knowledge-base entries for "CQP" and "Cheniere" returns four claims across two tracked channels. This is one of only two names in this batch with meaningful coverage, and it is the only one where the panel actively disagrees with itself.

All-In (2025-05-06, conviction 80, bullish, horizon: thesis) — entities Cheniere Energy, Sabine Pass; categories LNG / natural gas / energy exports: "LNG/methane has a massive growing global market with lower carbon footprint; US export demand exists regardless and is now the #2 dollar export."

> This claim names Sabine Pass directly — CQP's actual asset — and is the most on-point bullish claim in the lane. It is a market-structure argument, not a valuation argument.

Compound & Friends (2026-03-31, conviction 60, bullish, horizon: thesis) — entities Cheniere Energy, LNG: "US is the Saudi Arabia of natural gas; Cheniere's Gulf export terminals are active and running full because the world needs LNG."

Compound & Friends (2026-03-31, conviction 65, bullish, horizon: thesis) — entities LNG, Cheniere Energy: "US is the Saudi Arabia of natural gas (100-200yr supply); Cheniere's Gulf export terminals are active because the world needs LNG."

> These two are near-duplicates from the same episode on the same day and should be counted as one voice expressing one view twice.

Compound & Friends (2026-05-01, conviction 55, BEARISH, horizon: thesis) — entity LNG; categories commodity stocks / quality investing: "Wouldn't invest in Cheniere despite its LNG-export moat: margins are a function of the natural-gas price and management lacks control over its inputs, disqualifying it as a quality compounder."

How to read this lane honestly — three points.

First, the same channel reversed within five weeks. Compound & Friends published bullish framing on 2026-03-31 and an explicit disqualification on 2026-05-01. That is not a contradiction to be averaged away; it is a genuine evolution of view, and the later claim should carry more weight than the earlier ones by simple recency. The later claim is also the more specific and more analytically loaded of the two.

Second, the bearish claim is empirically supported by this file. "Margins are a function of the natural-gas price and management lacks control over its inputs" is exactly what §3 documents: FY22 revenue of $17,206M produced $2,498M of net income while FY23 revenue of $9,664M produced $4,254M. The tracked voice is describing a real, measurable feature of the reported financials. That materially raises our confidence in the claim.

Third, a distinction of entity that matters. Three of the four claims name "Cheniere Energy," which is the parent corporation, not CQP the partnership. The All-In claim naming Sabine Pass applies directly to CQP's asset. The Compound & Friends claims concern the franchise and management, which are shared — the general partner is controlled by the parent — so they apply in substance, but readers should understand that no tracked voice has commented on the partnership units specifically, on their leverage, or on their distribution.

kb_claim_count: 4. kb_breadth: 2. Net conviction: mixed and two-sided, with the most recent and most specific claim being bearish. This is a lane that supports the asset thesis and questions the investment thesis — which, notably, is precisely the shape of the Hold verdict this dive reaches independently on the arithmetic.

7. Valuation — what is priced at $67.00

At $67.00 (market cap $32.43B, net debt $14,485M, EV ~$46.92B):

Note the shape: EBITDA is essentially flat across the forecast horizon ($4,858M → $4,738M → $4,895M), so the EV/EBITDA multiple barely moves. There is no forward-multiple rolldown to point at here — the earnings do not grow into the price. That is an important and unusual feature: with most names, waiting makes the multiple look cheaper mechanically; with CQP it does not.

Note also the EPS discontinuity: FY25 actual EPS of $6.17 against FY26E of $3.839. That is not a forecast collapse — it reflects the derivative-mark volatility described in §3, and FY24's actual was $4.25. Consensus is modelling a normalised year, and FY25 was an above-normal one.

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

Reverse-engineering $67.00 into falsifiable claims. Consensus-derived arithmetic, labelled as such.

• EBITDA holds flat at roughly $4.7-4.9B annually through FY28. Consensus (5 analysts). At 9.7x FY26E EV/EBITDA the price is consistent with it. Falsifiable at each print. The key sub-assumption is contracted volume — this is a toll, so the assumption is that trains run and cargoes lift, not that prices rise.
• The distribution is maintained at $3.27 and remains covered above 1.5 times. At a 4.88% yield the distribution is roughly half the total-return case. Falsifiable at each declaration. FY25 free cash flow of $2,569M against ~$1,583M of distributions gives 1.62x — comfortable, and it held even in FY24's weaker year.
• Net leverage stays at or below ~3.3x and the $14,686M debt stack refinances at manageable rates. Interest expense has fallen from $870M (FY22) to $753M (FY25); the price assumes that trend does not reverse. This is the most fragile assumption in the price, because it is the one entirely outside management's control and the one where 3.3x leverage converts a rate move into an earnings move.
• The Sabine Pass complex operates without material interruption. One asset, one parish, Gulf coast. The price embeds zero probability of a multi-quarter outage. Not falsifiable in advance — only observable after the fact.
• The market keeps paying ~9.7x EBITDA for a flat-EBITDA, 3.3x-levered, single-asset partnership. Note that the sell-side does not currently agree it should: 10 of 18 raters are at Sell, with a $69 target. The price is being held up by yield buyers, not by analyst enthusiasm, and that is a positioning fact worth knowing.

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

Expected return over a two-year horizon: EBITDA growth (approximately zero on consensus) + multiple drift + distribution yield (4.88%).

This is the cleanest return bridge in the batch precisely because two of the three terms are near-zero. Consensus EBITDA is flat ($4,858M FY26E to $4,895M FY28E, +0.8% cumulative), so there is essentially no earnings-growth leg. Our base case assumes the multiple holds at roughly 10.0x FY26E EBITDA against the 9.7x the market pays today — a marginal expansion of 0.3x, justified not by growth but by continued deleveraging (net debt falling ~$650M a year mechanically transfers enterprise value to unitholders even at a constant EV/EBITDA multiple).

So the base-case total return of roughly +9.4% is: +4.5% of price appreciation, of which most is the deleveraging transfer rather than any operational improvement, plus 4.88% of distribution. There is no growth leg and no re-rating leg. State it plainly: an investor in CQP is being paid a yield, and any capital gain is incidental.

The bull case at $82 requires the multiple to expand to 11.5x on FY27E EBITDA — which would need either a step-change in contracted volume (the two-analyst FY29-30 revenue jump documented in §1, currently unsupported) or a compression of the yield toward 4.0%, which is a rate-environment bet rather than a company bet. We flag the bull case as multiple-dependent and therefore the fragile scenario.

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

• We agree with the bearish tracked claim and think it is the correct frame, but we price it differently. "Margins are a function of the natural-gas price" is empirically true in this file (§3), and it does disqualify CQP as a quality compounder. But that is an argument against owning it in a growth or quality sleeve — it is not an argument against owning a 1.9x-covered 4.88% yield in an income sleeve. Our variant perception is therefore about mandate, not about facts: the same evidence that disqualifies it for one purpose qualifies it for another.
• We think the sell-side's consensus Sell rating is about total return, not about risk, and that the distinction is being lost. With a $69 target 3.0% above spot, 10 Sell ratings are a statement that the units are fully priced — which we agree with — not that the asset is impaired. An income investor who reads "consensus Sell" as a solvency or coverage warning is misreading it. Watchable: any change in the distribution rate, which would be the actual warning.
• We think the FY29-30 revenue estimates are being ignored by the market and should be — but if they have any basis, the bull case is understated. Two analysts model revenue doubling to $21-25B, which would ordinarily imply major capacity expansion. The fact that their EPS estimates barely move ($4.72 and $4.24) makes the revenue figures internally incoherent and we exclude them. Positive surprise, watchable: any disclosed expansion project with financing attached, which would give those numbers a foundation and would materially change the growth score.
• The under-priced risk is the single-asset concentration, and it is under-priced because it is unmodellable. No estimate line, no ratio, no technical indicator captures the probability of a multi-quarter outage at one Gulf coast facility carrying $14.5B of debt. The 0.292 beta actively conceals it — low realised volatility on a contracted asset is not the same as low tail risk. Negative surprise: not watchable in advance; only sizeable in advance.
• Where we have NO edge: on gas prices, on global LNG demand, or on contracted volumes. Those are commodity and geopolitical variables and this dive does not forecast them.

Synthos fair values

• Bear ~$55 — 8.5x FY26E EBITDA ($4,858M) = EV $41.29B less net debt $14,485M = $26.81B ÷ 484M units = $55.39. Assumes a multiple de-rate toward the low end of the infrastructure range on leverage concerns, a soft LNG spread environment, or a distribution question. −18% (before distributions).
• Base ~$70 — 10.0x FY26E EBITDA ($4,858M) = EV $48.58B less $14,485M = $34.10B ÷ 484M = $70.45. Assumes flat EBITDA, maintained distribution, continued modest deleveraging, and a stable multiple. Sits just above the $69 street target. +4.5%, plus 4.88% distribution ≈ +9.4% total.
• Bull ~$82 — 11.5x FY27E EBITDA ($4,738M) = EV $54.49B less $14,485M = $40.00B ÷ 484M = $82.65. Assumes multiple expansion on a yield compression toward 4.0% or a credible capacity-expansion announcement. Above the $75 street high. +22%.

Cross-check on yield: at a 4.5% yield the $3.27 distribution implies $72.67; at 5.5% it implies $59.45. The scenario band brackets that range sensibly.

Base sits ~4.5% above spot with a −18% bear and a +22% bull. The distribution, not the price, is the return.

8. Technicals — the lowest-volatility chart in the batch

• Price $67.00, +1.7% on the day, −4.5% from the $70.14 fifty-two-week high; fifty-two-week low $50.49.
• Above a rising 50-day average ($62.69) by 6.9% and above a rising 200-day average ($59.40) by 12.8%. Both rising. A clean, orderly uptrend.
• RSI 57.9 — mildly constructive, not stretched. MACD +0.85.
• Beta 0.292 — the lowest in this batch by a wide margin (the next lowest is Ubiquiti at 1.313). max_dd_from_peak is just −4.5%, meaning the units have never suffered a meaningful drawdown within the fifty-two-week window.
• Returns: +16.2% over twelve months, +18.8% over six, +0.1% over three. The advance has gone flat at the top of the range — consistent with a yield instrument reaching a price where the yield no longer compensates.
• Read: there is nothing wrong with this chart and nothing exciting about it. A 0.29-beta instrument trading 4.5% below its high in a rising trend is behaving exactly as an income asset should. The technicals neither support nor oppose the verdict; they simply confirm this is not a trading vehicle.

9. Insider activity and governance

The insider record contains only general-partner governance events:

• Zamir Rauf — Form 3 filed 2026-07-16 (new director), and an A-Award of 3,103 phantom units at $0 on 2026-07-14.
• Michael Jennings — Form 3 filed 2026-07-16 (new director), with a corresponding award.

Read: two new directors were appointed to the general partner in July 2026 and compensated in phantom units. There are no open-market purchases and no sales anywhere in the file — which is normal for an MLP where the general partner is controlled by a corporate parent and directors are compensated in synthetic instruments rather than holding meaningful personal unit positions.

The governance point that matters for a unitholder: you are not represented by these directors in the way a shareholder is represented by a corporate board. The general partner is controlled by the parent, whose interests as parent and as general partner may diverge from those of the public unitholders — particularly on distribution policy, capital allocation between partnership and parent, and affiliate transactions. The FY22 seg_prod row showing "Liquefied Natural Gas, Affiliate" of $2,000M is a concrete illustration that affiliate revenue is material. This is a structural feature of MLPs, not a scandal, but it should be priced.

10. Moat and competitive position

CQP's moat is a permitted, constructed, contracted export facility — the hardest thing to replicate in this industry is not the technology but the permission, the site, the pipeline interconnect, the berths and the twenty-year offtake agreements. Capital expenditure of $199M against $10,758M of revenue is the signature of an asset whose replacement cost vastly exceeds its maintenance cost.

The evidence it prices: free cash flow of $2,569M-$3,698M in every year of the file regardless of a revenue range from $8,704M to $17,206M. Toll assets do that; competitive businesses do not.

The bounds, and they are real. First, the moat protects volume, not margin — which is exactly the bearish tracked claim, and §3 shows it is right. Second, the moat is site-specific, so it is also a concentration. Third, and structurally: the vendor's peer list includes Venture Global at a $31.9B market capitalisation, an indication that new export capacity is being financed and built by others. A moat around one facility in a business where new facilities are being built is a moat with a construction-schedule expiry date on the scarcity premium, even though the contracted cash flows persist.

Moat: strong on volume, weak on margin, bounded by new-build supply.

11. Capital allocation

Simple and, for the asset class, appropriate:

• Distributions of roughly $1,583M annually (484M units at $3.27), covered 1.62x by FY25 free cash flow.
• Debt reduction of roughly $650M a year — $15,991M (FY23) to $14,686M (FY25).
• Maintenance capital expenditure of $154M-$220M annually, 1.4%-2.0% of revenue.
• No buybacks in any year ($0 recorded throughout), which is normal — MLPs return capital by distribution.
• No stock-based compensation recorded ($0 in every year), consistent with the partnership structure.
• Unit count constant at 484M in every year of the file, FY21 through FY25. No dilution whatsoever. For an income investor that is a meaningful and unusual virtue.

The allocation question a unitholder should ask: with free cash flow of $2,569M, distributions of $1,583M and debt reduction of ~$650M, roughly $340M is unaccounted for annually. The file does not disclose where it goes. We flag it rather than speculate.

12. Verdict, kill-criteria and flip conditions

Hold — as an income holding, in an income sleeve, for an investor who accepts a K-1 and no voting control.

The asset is real and the cash is real: $2,569M of FY25 free cash flow on $199M of capital expenditure, a $3.27 distribution covered 1.9 times on a per-unit basis and 1.62 times at the partnership level, a 4.88% yield on a 0.292-beta security, no unit dilution in five years, and interest coverage of 8.1x. That combination is genuinely hard to find, and it is why this is not an Avoid.

But it is fully priced and structurally demanding. Our base case is $70 against a $67 spot — a 4.5% capital return, plus 4.88% of distribution, for roughly 9.4% total. The street is at consensus Sell with a $69 target. Leverage is 3.27x against one facility in one Louisiana parish. The forecast EBITDA line is flat through FY28. The top line is dominated by pass-through gas costs, so reported margins swing violently with the commodity — a point the most recent and most specific tracked knowledge-base claim makes explicitly, and which the FY22-versus-FY23 comparison in §3 empirically confirms. And the most recent quarter in the file is corrupted on two fields.

Own it for the distribution or not at all. Do not own it as a growth or quality position — one tracked voice disqualifies it on exactly that basis, and we agree.

Sizing and entry:

1. Appropriate size: 1-3% in an energy-income sleeve, in a taxable account, after confirming K-1 suitability.

2. No urgency to buy at $67 — the base case is only 4.5% above spot and there is no catalyst.

3. A better entry is a yield above ~5.3%, i.e. a price below ~$62 — which is the rising 50-day average and would improve the total-return case to roughly 18%.

4. Not appropriate for a growth sleeve at any price, on the mandate grounds above.

Pre-registered KILL / do-not-own criteria:

• Distribution coverage falling below ~1.2x on a per-unit free-cash-flow basis, or any reduction in the $3.27 rate — this is the entire investment case and its failure is the only thing that truly matters.
• Net leverage rising above ~4.0x EBITDA — would convert a stable levered toll into a financially constrained one, and would put the distribution at risk.
• Any material operational interruption at Sabine Pass — a multi-quarter outage on a single-asset, 3.3x-levered partnership is the tail risk that no ratio in this file captures.
• A structural change in the general-partner relationship that disadvantages public unitholders — affiliate transactions are already material (FY22 affiliate LNG revenue of $2,000M).
• EBITDA falling below ~$4.0B against the flat $4.7-4.9B consensus, which would break the 9.7x multiple the price rests on.

Pre-registered FLIP TO BUY:

• A price below ~$62 (yield above ~5.3%) with coverage intact — the single most likely path, requiring nothing but a normal drawdown.
• A credible, financed capacity expansion that gives the currently-incoherent FY29-30 revenue estimates a foundation — this would add a growth leg the market assigns no value to, and would justify the bull case.
• Net leverage below 3.0x with the distribution maintained — deleveraging into a stable distribution mechanically transfers value to unitholders and would support multiple expansion.
• A softening of the 10-Sell rating posture on improved coverage — a positioning catalyst rather than a fundamental one, but a real one given how one-sided the current split is.

Where CQP fits in the Synthos Framework Portfolio. The energy-income sleeve, as a contracted-infrastructure yield holding — 1-3% in a taxable account only. It is genuinely uncorrelated with the rest of this batch: a 0.292 beta against Rocket Companies at 2.176 and Sunbelt at 1.649, with a return driven by a covered cash distribution rather than by earnings growth or multiple expansion. Its portfolio function is ballast and income, and it should be sized as ballast. It is not a substitute for broader energy exposure — it is one facility, not a commodity position, and it will not track oil or gas prices in the way an exploration or integrated name does. Logged as a tracked Synthos call (Hold, income mandate only; Buy trigger below $62) as of 2026-08-04 at $67.00.

Single biggest risk: one asset, 3.27x levered, in one Louisiana parish. There is no second facility, no geographic diversification and no operational hedge. A hurricane, a prolonged outage, or a site-specific regulatory action affects 100% of the revenue base against $14,485M of net debt. The 0.292 beta actively conceals this — low realised volatility on a contracted toll is not low tail risk, and the two are being confused by the market.

Most fragile assumption in the price: that the $14,686M debt stack continues to refinance at rates consistent with the declining interest expense of the last four years ($870M in FY22 to $753M in FY25). It is the only major variable entirely outside management's control, it consumes 21% of operating income today, and at 3.27x leverage a sustained increase in refinancing cost compresses both the distribution coverage and the multiple simultaneously.

Provenance and disclosures

• Structure disclosure (stated up front, §1): CQP is a limited partnership, not a corporation. Investors hold units, receive a Schedule K-1 rather than a 1099, and have no voting control — the general partner, Cheniere Energy Partners GP, LLC, is controlled by the corporate parent. The partnership pays no entity-level income tax (effectiveTaxRateTTM: 0, which is correct rather than corrupt). K-1 suitability, state filing obligations and unrelated-business-taxable-income treatment in tax-advantaged accounts must be checked before purchase. The non-equity tripwire test was applied explicitly and CQP passes as a genuine operating equity (price $67 with no par anchor, $49.53-$70.64 fifty-two-week band, 5-6 analyst coverage, no note/trust/preferred naming convention); the two markers it does meet (0.292 beta, 4.88% yield) are inherent to contracted infrastructure MLPs.
• Traceability: 4 tagged knowledge-base claims across 2 channels (breadth 2, net conviction mixed and two-sided). Three are bullish on the LNG export thesis — All-In (2025-05-06, conviction 80, naming Sabine Pass directly) and two near-duplicate Compound & Friends claims from a single 2026-03-31 episode (conviction 60 and 65). One is explicitly bearish: Compound & Friends (2026-05-01, conviction 55) — "Wouldn't invest in Cheniere despite its LNG-export moat: margins are a function of the natural-gas price and management lacks control over its inputs, disqualifying it as a quality compounder." Note that the same channel reversed within five weeks, and that the bearish claim is empirically supported by this file (§3: FY22 revenue of $17,206M produced $2,498M of net income while FY23 revenue of $9,664M produced $4,254M). Entity caveat: three of the four claims name "Cheniere Energy," the parent corporation; only the All-In claim names Sabine Pass, CQP's asset. No tracked voice has commented on the partnership units, their leverage or their distribution specifically. Searched across 51,928 entries.
• Data as-of: fundamentals 2026-03-31 (Q1'26) · estimates, prices, technicals and knowledge-base claims 2026-08-04. All market data drawn exclusively from scripts/deepdive/vti_data/CQP_data.json.
• Vendor data rejected (see §1): inc_q[0] (quarter ended 2026-03-31) on two fields — interestExpense: -181000000, a sign flip against +$186M to +$202M in every other quarter on a $14,686M debt stack; and ebitda: 179000000 against operatingIncome: 329000000, which is arithmetically impossible. A third field, grossProfit at 9.9% of revenue against 30-50% elsewhere, is flagged as suspect and excluded from all margin computations. Consequence: the vendor's trailing EBITDA and its netDebtToEBITDATTM of 3.86x both inherit the corrupt quarter; we compute leverage on FY25 EBITDA of $4,428M, giving 3.27x, and label it as our computation. Also rejected: ratios_ttm.priceToEarningsRatioTTM of 14.105x, internally inconsistent with its own netIncomePerShareTTM of $5.2107 at a $67.00 price ($67.00 ÷ $5.2107 = 12.86x, which we use).
• Vendor data excluded from valuation: the FY29E and FY30E estimate lines, which model revenue leaping from $12,133M to $21,247M and $25,188M on two revenue analysts and one EPS analyst, while EPS barely moves ($4.365 → $4.72 → $4.24). Revenue doubling with flat per-unit earnings is internally incoherent. The base case is built on FY26E-FY28E, where coverage is 5-6 analysts.
• Vendor data checked and found SOUND — worth noting: unlike NTRA, ILMN and TWLO in this same batch, CQP's forecast EBITDA series passes an internal consistency check — FY26E EBITDA of $4,858M sits above FY26E EBIT of $4,118M by $740M of implied depreciation, both positive and close to FY25 actuals. The FY26-28 EBITDA series is used with confidence.
• Data gaps stated, not papered over: seg_geo contains a single row dated FY2017 — there is no current geographic revenue disclosure at all for an export business, and no claim in this dive rests on customer geography. seg_prod stops at FY2024. earn_cal EPS disagrees with inc_q EPS in two of the last four quarters (2025-10-30: $0.81 versus $1.05; 2026-02-26: $2.38 versus $2.41); inc_q was used throughout. Approximately $340M of annual free cash flow is unaccounted for between distributions and debt reduction; we flag this rather than speculate.
• Fair-value method: scenario multiples applied to consensus FY26E EBITDA ($4,858M, 5 analysts) and FY27E EBITDA ($4,738M, 5 analysts), less net debt of $14,485M, divided by 484M units. Bear 8.5x, base 10.0x, bull 11.5x. Cross-checked against a distribution-yield anchor (4.5% yield implies $72.67; 5.5% implies $59.45). Stated arithmetic only — no discounted-cash-flow model was built.
• Timing caveat: written two days before the Q2'26 print (2026-08-06). Note that CQP's quarterly EPS is unusually uninformative because of derivative marks — the last four prints were $0.91, $0.81, $2.38 and $0.38 against estimates of $0.96, $1.02, $1.11 and $1.15 — so the print should be read for EBITDA and distribution coverage, not for the earnings surprise.
• Not investment advice. Independent research, educational and informational only, never personalised.
• Version: 2026-08-04-full.