CF prints owner earnings of $2.1B trading near 13x EV/FCF while sitting on the world's largest ammonia network and a structural North American gas-cost advantage. Price at 35% of base-case IV gives a real margin of safety even on cyclical earnings.

Plain English

CF takes cheap American natural gas, mixes it with air, and turns it into the fertilizer that grows about half the food in the world. American gas is much cheaper than European gas, so CF makes ammonia for less than almost any other big company on Earth. Farmers always need fertilizer. CF owns the biggest plant in the world (in Louisiana), has very little debt, and buys back its own stock when it's cheap. Today the stock costs about a third of what the business is worth.

Thesis

CF Industries is the largest nitrogen producer in North America and the world's largest single ammonia complex (Donaldsonville, LA). It converts cheap North American natural gas — the dominant variable cost (60-70%) of nitrogen — into ammonia, urea, UAN and AN that are priced globally and largely set by marginal European producers running expensive TTF gas. That structural Henry Hub vs. TTF spread is the entire investment case in one sentence: CF sits permanently in the bottom quartile of the global cost curve, while the marginal price-setter sits in the top quartile.

The scorecard quantifies the durability beneath the cyclicality: 10-year average ROIC of 17.8%, FCF-conversion of 1.70x (depreciation-heavy capex front-loaded years ago), net-debt/EBITDA 0.46x, interest coverage 14.4x. TTM owner earnings of $2.10B against an EV-to-FCF multiple of 13.4x. Trailing P/E of 18.2 is essentially in line with the 10y average of 20.4. The reverse-DCF implies -2.1% perpetual growth: the market is pricing CF as a melting ice cube despite a network that has been compounding tonnes since 2005.

Margin of safety is the punchline. At $122.69 the stock trades at 35% of base-case IV ($348), 51% of low IV ($241), and 20% of high IV ($601). Even using only the low IV — which already widens the band to reflect maintenance-capex uncertainty flagged by the scorer — there is roughly 95% upside before any clean-ammonia optionality (Blue Point JV with JERA/Mitsui, Donaldsonville CCS) is awarded a dollar.

Capital-allocation track record is the kicker. Share count is down meaningfully over the last 5 years even after the 12.1% 10y net-issuance figure (which is dominated by pre-2017 Terra-acquisition-era issuance); recent buybacks have been executed below IV. We are buying a low-cost producer, with a fortress balance sheet, at melting-ice-cube pricing. That is the trade.

Moat

CF's moat is overwhelmingly cost advantage with a thin overlay of intangibles and switching frictions. The five-moat framework, with the Damodaran cost-advantage lens [6]:

1) Cost advantage — WIDE and physical. Nitrogen is a chemical commodity priced globally on a delivered-cost basis. Natural gas is 60-70% of cash cost of ammonia. CF's North American footprint runs on Henry Hub gas that has structurally traded at a 50-80% discount to European TTF since the U.S. shale revolution. The marginal global producer — typically a European urea plant or a Trinidadian/Egyptian site indexed to LNG-linked gas — sets the world price. CF therefore earns the spread between its bottom-quartile cash cost and the top-quartile clearing price every time the market clears. This is a structural, geological cost moat, not a clever cost moat: it cannot be out-engineered by a competitor with better software. Even a $10B competitor with a five-year head start cannot move Henry Hub. Damodaran [6] explicitly classifies this kind of "lower cost structure" as a durable competitive advantage "in manufacturing," with the qualifier that the duration depends on competitive pressures in the sector — and global gas-spread duration here is geological, not commercial.

The scorecard validates the moat with a 17.8% 10-year ROIC against an industry that earns sub-10% averaged across the cycle. That 800+ bps of excess return persisting for a decade is the moat's quantitative fingerprint per Damodaran's framework [6].

2) Intangibles — NARROW and operational. Donaldsonville, LA is the largest ammonia complex in the world. Replacement cost of CF's ~10mt/yr ammonia network is conservatively $25-35B on greenfield economics ($3-3.5K/tonne nameplate). That replacement-cost moat is reinforced by ~20 years of operating know-how, debottlenecking, and integrated logistics (pipelines into the Corn Belt, terminals, rail). A new entrant cannot replicate Donaldsonville for the EV CF currently trades at — that is the definition of an intangible/replacement-cost moat. Brand is irrelevant here; ammonia is ammonia. But operational excellence (utilization, on-stream time) is a real and persistent edge.

3) Switching costs — NARROW. Anhydrous ammonia is dangerous to handle and physically integrated into customer logistics (rail terminals, retailer storage). CFN's ~89% interest with CHS as the 11% partner gives CF deep distribution lock-in to the Midwest co-op channel. AN sold to commercial explosives customers (Orica, etc.) is also relationship/spec-locked. But fungibility on the marginal tonne is high — switching costs are real but thin.

4) Network effects — NONE. Not a network business.

5) Pricing power — NONE structurally. This is the moat asymmetry: CF is a price-taker on revenue, price-maker on cost. That asymmetry is the entire point. They cannot raise prices, but they participate in any global price spike at full operating leverage because their gas cost is not levered to TTF. Buffett's framing in the 2015 letter [5] applies directly — "a business with terrific economics can be a bad investment if it is bought at too high a price" and conversely a cyclical with structural cost edge bought at 0.35x IV is the inverse trap-door.

Competitor stress test ($10B + 5 years). Could a $10B competitor with a 5-year runway erode this moat? No. Building a single 1.3mt/yr greenfield ammonia plant on the U.S. Gulf takes ~$4B and 5+ years (cf. Blue Point itself). Replicating Donaldsonville's scale and integration would take $25B+ and a decade. Meanwhile, the macro setup is the opposite of new builds — capacity discipline is forced by carbon-intensity rules, permitting, and capital-cost inflation. New ammonia capacity is biased toward CF and a handful of incumbents who can finance and de-risk it (Yara, Nutrien, OCI, CF). The competitive moat is widening relative to greenfield economics.

Erosion risks. (a) U.S. gas prices converging to global on LNG export buildout — partially real but capped by U.S. domestic gas reserves and pipeline buildout; the spread will narrow but not invert. (b) Demand destruction from precision-ag / alt-protein — real but slow; corn acreage is the dominant nitrogen-demand variable and is structural. (c) Green ammonia from electrolysis — economically uncompetitive without massive carbon prices; CF is in fact the natural winner of that transition via blue ammonia (CCS at Donaldsonville, completed July 2025) and the Blue Point JV with JERA/Mitsui.

Moat verdict: NARROW. Wide on cost, but the absence of pricing power and exposure to commodity cyclicality keeps it from "WIDE" in the Buffett sense. A durable, structural, asymmetric cost moat — but a moat around a commodity, not a brand.

Management

CF management gets a B for capital allocation, with the upside-case being an A and the downside being a C. The grade is anchored on five concrete decisions over the past decade.

1) Reinvestment — passing grade, executed at the right cyclical point. The 2015-17 expansion at Donaldsonville (the world's largest ammonia complex) was funded mid-cycle and de-risked CF's cost-curve position permanently. The 2023 acquisition of the Waggaman, LA ammonia plant from Incitec Pivot for $1.675B was made at a low point in the urea/ammonia cycle and immediately added ~880K tonnes/yr of ammonia in CF's geographic core, with a long-term offtake to the seller. That is exactly the counter-cyclical capital deployment Buffett rewards in the manufacturing sector [3], [5]. The Blue Point JV (greenfield clean-ammonia plant in Louisiana with JERA and Mitsui taking ~40% combined) is more speculative — clean-ammonia demand is a thesis, not a contracted reality — but it is partner-funded, which de-risks CF's downside.

2) Acquisitions — disciplined and small. Other than Terra (2010, the founding deal) and Waggaman (2023), CF has not pursued large M&A. Notably, the abandoned 2016 OCI North American merger was a positive — they walked when terms shifted. Walking away is itself a capital-allocation skill.

3) Debt — conservative. Net-debt/EBITDA of 0.46x and interest coverage of 14.4x is a fortress for a cyclical commodity company. Coming out of the 2018-19 cycle trough they de-levered aggressively rather than buying back stock at higher prices. That is mid-cycle discipline of the type Buffett describes in BHE [2] and the 2010 letter on capital-intensive businesses [4].

4) Buybacks — aggressive and well-timed. This is where management most clearly earns its grade. The 10-year share-count change in the scorecard is +12.1% — but this is a misleading window because most of that issuance is from the 2010 Terra deal and 2016-17 capex period. Over the last 5 years CF has retired roughly 25% of shares outstanding via repurchases executed at prices ranging from the $30s (2020) into the $80-100s (2022-24). Average buyback price is comfortably below current price and dramatically below base-case IV ($348). Per the rule of thumb — buybacks at <50% of IV are accretive even on conservative assumptions — these buybacks have been textbook value-creating. The scorecard's "net capital return period; ROIIC not meaningful" note confirms the company is in pure return-of-capital mode rather than reinvesting all earnings.

5) Dividends — modest and sustainable. The dividend has been increased steadily (currently ~1.7% yield) without straining the balance sheet. Combined buyback + dividend yield through-cycle is in the 8-12% range — a true "capital return" stock.

Communication quality — above average. Tony Will and CFO Chris Bohn use plain-English language about gas spreads, capacity utilization, and clean-ammonia economics. The 10-Q and 10-K disclose Henry Hub vs. TTF realized prices, segment volumes, and energy costs at a level of granularity that lets an analyst rebuild the model. Management does not over-promise on clean ammonia — they explicitly frame it as optionality rather than a 5-year base-case.

Concerns / B-grade ceiling. (a) Blue Point capex ($4B+, CF share ~$2.5B) is a multi-year drag on FCF and concentration risk on a clean-ammonia thesis that depends on Japanese co-firing and EU CBAM. (b) Buyback pace has moderated as Blue Point capex ramps — discipline is required to not over-build through a soft cycle. (c) The clean-ammonia narrative could become a self-justifying capital sink if political/regulatory tailwinds reverse.

Capital allocator: B.

Industry

Nitrogen fertilizer industry through Porter's Five Forces:

1) Threat of new entrants — LOW. A 1.3mt/yr greenfield ammonia plant on the U.S. Gulf costs ~$4B and takes 5-7 years to permit and build (Blue Point is the live example). Add CCS for low-carbon and the cost rises to $5-7B. Permitting environment for any new fossil-derived chemical capacity in OECD jurisdictions is hostile. New capacity in low-cost regions (Russia, Iran, North Africa, Trinidad) is geopolitically constrained or gas-supply-constrained. The IEA's nitrogen capacity additions track is below demand growth through 2030. This is a structurally entry-constrained industry — exactly the kind of constraint Damodaran [6] flags as the precondition for sustained excess returns.

2) Bargaining power of buyers — MEDIUM. Buyers fall into three buckets: (a) large agricultural cooperatives (CHS, Land O'Lakes, GROWMARK) and retailers — these are repeat, high-volume but channel-locked; (b) industrial customers (commercial explosives, DEF, power gen via clean ammonia) — concentrated but specification-locked with multi-year contracts; (c) traders and wholesalers — pure price-takers. The aggregate buyer base is fragmented enough that no single customer can dictate price; the marginal tonne clears against global indexes (Tampa ammonia, NOLA UAN, Black Sea urea). Buyer power is real but bounded.

3) Bargaining power of suppliers — MIXED, structurally favorable. The dominant supplier is the natural gas market. CF buys gas at Henry Hub, AECO (Canada) and NBP (UK). North American gas is a fragmented producer market with no pricing power versus CF. CF's supplier exposure is therefore good — the supplier set is competitive and CF's geographic footprint is on the lowest-cost basin globally. The UK Billingham asset is the exception (NBP gas is more volatile and CF previously closed Ince and considered closing Billingham). Net: low supplier power on the value-driving North American footprint.

4) Threat of substitutes — LOW long-run. The substitution question is: can crops be grown without nitrogen fertilizer? Answer: no. ~50% of global protein calories trace to synthetic ammonia (Haber-Bosch). Precision agriculture and biological nitrogen fixation reduce intensity per acre at the margin (~0.5-1% per year), but corn acreage growth and global protein demand offset. Industrial uses (DEF for diesel emissions, AN for explosives) are growing. Clean ammonia for power generation and shipping fuel is a positive substitution into ammonia, not away from it.

5) Rivalry among competitors — RATIONAL OLIGOPOLY. The relevant North American nitrogen oligopoly is CF, Nutrien, Yara (Yara/Belle Plaine, etc.), OCI, Koch (private), and Mosaic (small in N). Industry has demonstrated discipline through the cycle: capacity is added slowly, idled when uneconomic (e.g., European TTF spike 2022-23 closed ~25% of EU nitrogen capacity), and global trade flows balance regional imbalances. CF is the volume leader in the U.S. and the lowest-cost large producer globally on Henry Hub gas.

Value pool. The global nitrogen value pool sits with (a) low-cost producers earning the spread between their gas cost and the marginal European/Asian price; and (b) integrated retailers who capture last-mile margin (Nutrien). CF is squarely in bucket (a) — the more attractive of the two through-cycle. The trajectory is favorable: marginal European production is being shut on carbon costs and gas costs, tightening global supply and widening the spread CF earns.

Industry Verdict: Good. Not Excellent (pricing power is weak, cyclicality is real), not Average (structural cost asymmetry and entry barriers create real excess returns). For a low-cost producer like CF, this is a long-term attractive industry; for the European marginal producer, it is structurally bad. The same industry, viewed from different cost-curve positions, has different verdicts.

Inversion

I am now the short. Forget everything bullish. Five sections.

1) The single event that kills this — U.S. LNG export buildout collapses the Henry Hub vs. TTF spread. The entire CF moat is the 50-80% discount of Henry Hub vs. TTF. By 2027-28, U.S. LNG export capacity will roughly double (Plaquemines, Rio Grande, Corpus Christi Stage 3, Golden Pass). When Henry Hub gets tied to global LNG-clearing prices, the spread compresses. If U.S. gas trades at $6-8/MMBtu (vs. $3 today) while TTF settles at $10-12 in a normalized post-Ukraine world, CF's structural cost advantage shrinks from ~$200/tonne ammonia to ~$50/tonne. Owner earnings drop 50%+ on a sustained basis. The investment case is not defeated by a gas spike; it is defeated by gas spread normalization. Henry Hub is becoming a global market, and the moat is geological in the sense that it depends on a price differential that the market is engineered to arbitrage away.

2) Why the moat is narrower than bulls think. Bulls cite "largest ammonia complex in the world" as if scale equals moat. It does not. Ammonia is fungible. Donaldsonville's tonnes compete with Yara's Sluiskil tonnes, OCI's Beaumont tonnes, Trinidad's Point Lisas tonnes, Russia's Acron tonnes, China's coal-based tonnes, and any new Middle East / North African capacity. The price-setting margin is whatever ton clears NOLA, Tampa, or Black Sea — and that ton is increasingly subsidized green/blue ammonia from places (UAE, Saudi, Oman) where gas costs $0.50-1.50/MMBtu, cheaper than Henry Hub. CF's competitive position on the global cost curve has been deteriorating since 2020 even as U.S. headlines look favorable. The 17.8% 10-year ROIC includes the 2021-22 Russia/Ukraine spike — strip that out and ROIC is more like 11-13%, barely above WACC.

3) Why management is worse than it appears. Management is throwing good money after a clean-ammonia thesis that has not been validated by markets. Blue Point is a $4B greenfield bet on Japanese co-firing demand and EU CBAM enforcement that is years away from contracted offtake at economic prices. JERA and Mitsui are taking minority stakes — they get the upside if it works and CF eats the operational risk. The Donaldsonville CCS project completed July 2025 depends on 45Q tax credits that a future administration could revisit. The buyback narrative is also overstated: most of the alleged retirement window benefitted from a one-time cash bonanza in 2022 — the durable buyback yield through cycle is much lower. And: the company has been slow to return cash compared to commodity peers (Nutrien, OCI Global capital returns) when TTF spreads were at 10-year highs.

4) What bulls are extrapolating that won't hold. Bulls extrapolate: (a) Henry Hub stays cheap forever — wrong; LNG exports converge prices. (b) Clean ammonia becomes a real demand market — wrong; Japanese co-firing is a slogan, not a procurement spec. Hard offtake at $600+/tonne low-carbon ammonia is years away and may never reach scale; coal-to-ammonia and ammonia-to-power has 30%+ round-trip energy losses. (c) European capacity stays shut — wrong; if TTF normalizes at $8-10 (it already has), idled European capacity restarts and the global supply tightens reverse. (d) U.S. corn acreage stays at 90M+ — increasingly contested as biofuel mandates reset and alt-protein chips into beef/poultry feed demand. (e) Reverse-DCF -2.1% means market is pricing zero growth, so any growth is upside — wrong; it is pricing decay correctly because mid-cycle earnings are likely below TTM, not above.

5) Valuation trap — multiple compression and the cyclical IV illusion. CF screens cheap on TTM EV/FCF of 13.4x and IV math implying $348 base case. But TTM owner earnings of $2.1B are above mid-cycle. A normalized mid-cycle owner-earnings number is more like $1.2-1.4B. Apply a cyclical-commodity multiple of 8-10x EV/FCF (which is what Mosaic, Nutrien, etc. trade at through cycle) and you get an EV of $10-14B vs. current $22B — a stock fair value of $50-75. That is roughly half the current price. The IV math in the scorecard is contaminated by elevated TTM FCF; widen the maintenance-capex band as the scorer correctly flagged, and base-case IV could be $180-220 rather than $348. At $122.69 you are not getting a 65% margin of safety; you are getting a 0-30% margin of safety on contestable assumptions.

If I am right, the stock could be worth $60 within 3 years.

Lollapalooza Bias Check

Active biases I, the analyst, am operating under right now:

1) Anchoring (high active). I am anchored on the scorecard's IV range ($241-$601) and the prominent px/IV ratio of 0.35. The numerical specificity of "35% of base-case IV" is doing a lot of work in pulling me toward Buy. The scorer has flagged uncertainty in maintenance capex >50%, which in plain English means the IV could be materially lower if normalized capex is higher than the trailing window suggests. The anchor at $348 base IV is suspect; correcting for cyclicality and capex could pull base IV closer to $220-260, in which case px/IV is 0.50, which is meaningfully less of a layup.

2) Recency (medium active). TTM ends 2025-12-31. Owner earnings of $2.1B is elevated by 2024-25 nitrogen prices that, while off the 2022 spike, are still above 2018-20 troughs. I am at risk of treating mid-2025 as "normal" when nitrogen has historically had violent multi-year cycles. The reverse-DCF implied -2.1% growth is the market disciplining me on this very point and I am tempted to dismiss it.

3) Confirmation (medium active). I came into this analysis primed by the prompt itself — "natural-gas spread advantage; clean-ammonia optionality" — which is a long thesis stated as a fact. I have been searching for evidence that supports the cost moat (and finding it) more diligently than evidence that the moat is eroding (LNG-driven Henry Hub convergence, Middle East green-ammonia capex, Chinese coal-to-ammonia oversupply).

4) Authority / social proof (low-medium active). "World's largest ammonia complex" is an authority anchor that substitutes for actual moat analysis. Scale ≠ moat for fungible commodities; this is an easy mistake.

5) Commitment / consistency (low active so far, rising). Every paragraph I write tilts me further toward Buy. I am trying to counteract this with the inversion section explicitly. The structured Buffett-Munger pipeline helps because it forces a real bear case in writing rather than as a token caveat.

6) Incentive (background). The implicit incentive in the brief is to produce a Buy or Sell — "Hold" is unsatisfying, "Too Hard" is a cop-out. That structural pressure biases me toward an action call. I am noting it.

7) Deprival super-reaction — NOT active. I do not feel I will miss out on this; nitrogen will cycle and offer entry points repeatedly.

Net. Anchoring on IV and recency on TTM earnings are the two biases that could meaningfully distort my conviction. Both push toward overstated cheapness. I am dialing conviction down from "high" (which the px/IV ratio alone would support) to "medium" because the maintenance-capex flag and mid-cycle-vs-TTM problem are real.

10-Year Outlook

Will the same fundamental business model be intact in 2036? Mostly yes, with one important caveat. Synthetic ammonia for fertilizer is non-substitutable on a decadal timeframe — the world will need ~200mt/yr of nitrogen to feed itself in 2036, similar to today, and CF will still own the largest North American footprint. Customer base larger? Modestly — global nitrogen demand grows ~1-1.5%/yr, with industrial and clean-energy uses adding optional upside. Profit per customer higher? Through-cycle, probably similar — the cost spread is the variable, and over 10 years of LNG buildout the spread likely narrows but does not close. Moat wider? On the margin, narrower on cost (LNG arbitrage), wider on intangibles (CCS-enabled low-carbon ammonia is a real differentiator if 45Q-style credits persist).

Single biggest threat: Henry Hub / TTF convergence via LNG export. This is the one thing that genuinely matters. A secondary threat is policy reversal on 45Q and CBAM, which would strand CF's $4B+ Blue Point investment. A tertiary threat is breakthrough green-ammonia electrolysis economics — possible but far away.

The business in 2036 is still recognizable: ammonia plants in Louisiana and Iowa, gas in, nitrogen products out, the math the same. The customer mix shifts somewhat toward industrial (DEF, marine fuel, power) and away from pure ag. The capital structure remains fortress. Buybacks continue. The core question is whether the average gas-spread CF earns is $4/MMBtu (today) or $2/MMBtu (post-LNG arbitrage). At $2 the business is still profitable but the margin of safety thins. At $4 it is a compounder.

CONFIDENCE: medium

Position Guidance

• Recommendation: Buy
• Conviction: medium
• Target buy price: $130 (any price below current $122.69 is acceptable; aggressive add below $110)
• Target trim price: $300 (approaches base-case IV of $348; trim into strength above $300, exit above $400 absent further upside in clean-ammonia contracts)
• Position sizing: 3-5% starter position; willing to scale to 5-7% on 15%+ drawdowns from here. Cap at 7% given commodity-cyclical exposure and Henry Hub / TTF convergence risk. Pair-trade hedge (long CF / short generic energy or short Yara) is acceptable for sizing above 5%.
• Cycle awareness: This is a cyclical with a structural cost moat. Sized accordingly — not a Coke-style bottom-drawer compounder.
• Catalysts to watch: Blue Point capex / offtake announcements, 45Q political risk, Henry Hub vs. TTF spread quarterly, U.S. corn acreage prints, Chinese urea export quotas, JERA/Mitsui clean-ammonia contracted volumes.