Bunge Global Sa BG

Thesis

Bunge Global SA (BG) is one of the four 'ABCD' global agricultural traders, alongside ADM, Cargill, and Louis Dreyfus. The business buys oilseeds and grain from farmers, ships them, crushes them into meal and oil, and sells the output to food, feed, and biofuel customers. The 2024-2025 merger with Viterra extends origination scale, but the underlying economics remain those of a price-taker commodity middleman. The scorecard tells the story bluntly: 10-year average ROIC is 0.0%, FCF conversion 5y is -22.1%, owner earnings TTM is negative ($-0.17B), net debt to EBITDA is 21.99x (post-Viterra and inventory-financing distortion), interest coverage is 0.0x, and intrinsic value bands are -$20.88 (low), -$14.41 (base), -$14.41 (high) — all negative against a $124.61 price. The reverse-DCF implied growth is null because there is no positive base to grow. The scorer itself flagged 'insufficient history' and 'Short history (3y annuals); IV bands and 10y-ROIC less reliable; treat as exploratory,' with base CAGR clamped from -27.5% to -5.0%. Composite score is 69 and valuation 17 on partial data, but the qualitative shape is unambiguous: commodity inputs, commodity outputs, and a balance sheet whose leverage ratio is dominated by working-capital-style debt against pledged inventory. There is no price at which a long-term Buffett-Munger compounder thesis becomes meaningful here, because the underlying business does not compound — it oscillates. The right action is to file under Too Hard.

Moat

Bunge fails the moat test on every one of the five dimensions. Pricing power: none. Soybean meal, soybean oil, corn, and wheat are exchange-traded commodities with daily reference prices on CBOT and global cash markets. Bunge takes the spread between origination and processing, not a markup. Switching costs: trivial. Farmers sell to whichever elevator pays the best basis on a given day; food, feed, and biofuel customers buy from whoever ships the cheapest contracted spec. The Viterra merger consolidates origination but does not bind a counterparty. Network effects: weak. Logistics density (river terminals, rail, ports) creates regional advantages but does not compound globally — ADM, Cargill, COFCO, and Louis Dreyfus operate parallel networks, and origination share is contestable in every harvest. Intangibles: none of consequence. There is no consumer brand pull (Bunge is B2B), no patent estate, and no regulatory exclusivity. Cost advantages: real but bounded. Scale in crushing capacity, ocean freight chartering, and risk management does drop unit costs versus a regional player, but the four global majors have approximate parity, and Chinese state-linked COFCO has been narrowing the gap. None of this rises to the durable wide-moat shape Buffett describes for See's Candy [3] or BNSF [4] — businesses where customers cannot easily defect and where the structural cost or brand advantage is self-reinforcing across decades.

Apply the $10B / 5-year stress test: if a competitor or sovereign-backed entrant deployed $10B over five years to buy elevators in Brazil, the U.S. Midwest, and the Black Sea, would Bunge's economics meaningfully degrade? The answer is yes — and this scenario is not theoretical. State trading companies (COFCO, India's NAFED, sovereign Gulf buyers) have been doing exactly this for a decade. Damodaran [2] is explicit that competitive advantage must be both preserved and grown; in commodity intermediation, scale advantages tend to dissipate as new infrastructure comes online and as digital procurement platforms commoditize counterparty relationships further.

Erosion vectors are structural. (1) Renewable-diesel demand for soybean oil pulled crush margins to historical highs in 2022-2023; mean reversion is already underway. (2) U.S.-China trade-flow shifts have made Brazilian origination more valuable, but Brazil also rewards local consolidators (Amaggi, ALZ Grãos) that operate closer to the farmer. (3) Drought, flood, and geopolitical supply shocks are the source of episodic windfall years that bulls extrapolate; they are not durable moats. (4) The post-Viterra integration introduces execution risk that further widens the band of plausible outcomes.

Buffett's framing is helpful here: 'if a business requires a superstar to produce great results, the business itself cannot be deemed great' [3]. Bunge's good years are produced by trading desks and risk management teams reading basis, freight, and FX correctly — exactly the superstar-dependent shape Buffett warns against. There is no See's Candy, no Mayo Clinic, no toll-bridge here. The cost-advantage story, where it exists, is shared with three peers of comparable scale and contested by sovereign-backed entrants. Moat verdict: NONE.

Management & Capital Allocation

Capital allocation at Bunge is constrained by the nature of the business. Most reinvestment is mandatory maintenance and safety capex on crushing plants, terminals, and barges; growth capex in the last cycle has gone into oilseed crush expansion, biofuel-adjacent renewable-diesel feedstock capacity, and the transformative Viterra acquisition. Each of these decisions is rational on its face, but each is also a leveraged bet on continuing through-cycle margins in a commodity business where the median 10-year ROIC has been zero per the scorecard.

Reinvestment: The Viterra deal (announced 2023, closed 2025) added grain origination and handling assets at a cycle that may or may not have been the top. The strategic logic — combining Bunge's downstream processing with Viterra's upstream origination — mirrors the integration logic that Cargill and ADM have long pursued. The execution risk is real: post-merger leverage is the proximate cause of the 21.99x net debt to EBITDA ratio in the scorecard, though that figure is partly an artifact of inventory financing being scored as debt and depressed TTM EBITDA in a normalizing margin year.

Acquisitions: The Viterra deal is by far the largest in Bunge's history and dominates any management grade. The price paid (~$8.2B equity plus assumed debt) was struck when crush margins and grain handling spreads were elevated. If through-cycle margins normalize lower than 2022-2023, the deal will look expensive; if scale benefits accrete over 5+ years, it could be defensible. There is no way to know yet, which is itself a reason for low conviction.

Debt: Leverage rose materially to fund Viterra. Interest coverage in the scorecard reads 0.0x, which reflects depressed TTM EBIT against rising interest expense — a flag, even allowing for measurement noise on a freshly-merged entity. A commodity processor running tight coverage into a margin downcycle is the textbook recipe for forced asset sales or equity issuance at the wrong price.

Buybacks: Bunge has historically returned capital through buybacks during good years. The 10-year share count change is -3.39% per the scorecard — modest, and notably the buybacks were concentrated in 2022-2024 when the stock traded at higher multiples. Buying back stock above intrinsic value during cyclical peaks is the canonical capital-allocation mistake Buffett warns about. We cannot calculate average P/IV precisely because IV is negative, but qualitatively the timing was poor.

Dividends: Steady quarterly dividend, well within free cash flow in good years; this is the least controversial element of the policy.

Communication: 10-K and 10-Q disclosures are detailed and segment-level (Agribusiness, Refined and Specialty Oils, Milling, Sugar and Bioenergy historically). Tone is professional and quantitative. There is no obvious promotional language. Management does not pretend the business is a compounder; they describe it accurately as cyclical and basis-driven.

The honest grade has to weight the Viterra deal heavily because it dominates the next decade of outcomes. Conditional on integration going well and through-cycle margins holding, this is a B-tier capital allocator. Conditional on integration friction or margin reversion, it is a D. Splitting the difference and adding a discount for the buyback timing: Capital allocator: C.

Industry Structure

Threat of new entrants — Low to Moderate. Building global origination, crushing, and logistics infrastructure requires billions in patient capital and decades of relationships, which deters greenfield entry. However, sovereign-backed entrants (COFCO most prominently) have demonstrated that capital-and-policy combined can compress the entry timeline. The barrier is real but not Buffett-grade.

Bargaining power of suppliers — Moderate. Suppliers are millions of farmers globally. Individually they have no power, but collectively, in concentrated growing regions like the U.S. corn belt or the Brazilian Cerrado, they can shift sales to whichever elevator offers the best basis on a given day. Co-ops (CHS, etc.) further compress trader margins on the origination leg.

Bargaining power of buyers — Moderate to High. Buyers are large and sophisticated: feed compounders (Tyson, JBS), food processors (Unilever, Nestlé), biofuel refiners, and renewable-diesel producers. They contract on negotiated specs, multi-source aggressively, and have full transparency into exchange-quoted reference prices. There is essentially no scenario in which Bunge can raise its take-rate by fiat.

Threat of substitutes — Moderate and Rising. Soybean oil faces substitution from palm, canola, sunflower, and used cooking oil (UCO) on a relative-price basis. The renewable-diesel boom has made this dynamic louder: a feedstock-agnostic refinery will buy whatever is cheapest per gallon of finished fuel. Plant proteins and synthetic alternatives are smaller threats over a 10-year horizon but real ones.

Industry rivalry — High. Four global majors (ADM, Bunge, Cargill, Louis Dreyfus) plus COFCO compete on basis, freight, and risk management. Margins compress in surplus years (e.g., the soybean glut periods) and expand in scarcity years. Through-cycle, no participant has demonstrated structurally higher returns; Damodaran's framing applies — apparent advantage is the consequence, not the cause, and in commodities the consequence does not durably persist [2].

Value pool location and trajectory. The value pool sits primarily in (a) crush margins, which are oversupplied during normal years and tight during demand spikes (renewable diesel pulled them tight in 2022-2023, with mean reversion now visible); (b) basis trading, which is essentially a short-volatility activity that pays well in stable years and produces losses in dislocated ones; (c) ocean freight and logistics arbitrage, which is highly cyclical. The trajectory of the pool is roughly flat to modestly down in real terms as digitization narrows information asymmetries and as biofuel mandates become a known variable rather than a surprise tailwind.

Verdict. The industry has high entry barriers in absolute terms but the value pool is not durable, returns are cyclical with median through-cycle ROIC near zero, and pricing power belongs to the exchange, not the participants. Industry Verdict: Average — closer to Poor than to Good, but not the worst structure in the economy because the entry barriers do prevent total commoditization.

Inversion (Bear Case)

I am a short-seller, and the bear case here writes itself.

The single event that kills this. A normalization in soybean crush margins back to the 2018-2020 range, simultaneous with a soft Brazilian or U.S. harvest cycle that compresses basis, simultaneous with the renewable-diesel feedstock arbitrage closing as biofuel mandates plateau and as used-cooking-oil and tallow imports continue to surge. Any one of these is survivable. All three together — which is the modal scenario for 2026-2028 — produces an EBITDA collapse against the post-Viterra debt stack and forces either an equity raise at a depressed multiple or asset sales at distressed prices. The scorecard already shows the warning lights: net debt to EBITDA at 21.99x, interest coverage 0.0x, owner earnings TTM negative at -$0.17B. Bulls will argue these are noisy post-merger artifacts. They are also exactly what the early innings of a cyclical downturn look like.

Why the moat is narrower than bulls think. The bull narrative leans on Viterra-driven scale: more origination, more crush, more vertical integration, lower unit cost. The flaw is that scale in commodities does not produce durable pricing power — it produces marginally lower unit cost in a market where everyone else is also scaling. ADM is integrating its own deals, COFCO is buying everything Brazilian that moves, and Cargill remains private and patient. The Damodaran point [2] cuts in the bear's favor: competitive advantage must be continuously reinvested in or it dissipates, and in commodity processing the reinvestment treadmill is brutal. Buffett's See's Candy framing [3] makes the contrast vivid: See's grew volume 2% a year for 50 years and produced extraordinary returns because the moat was a brand, not a cost curve. Bunge has no brand. Its customers do not know its name. Its suppliers will sell to whoever pays the best basis tomorrow morning.

Why management is worse than it appears. Management timed buybacks during the cyclical peak (the 10-year share count is down only 3.39%, but the bulk of repurchases clustered in 2022-2024) and timed the Viterra acquisition during what may prove to have been peak crush margins. Both decisions are standard cyclical-management mistakes — buying high and using leverage to do it. The communication is professional, but professional communication does not change the math. The scorecard's interest coverage at 0.0x and net debt to EBITDA at 21.99x reflect choices, not just measurement noise. Bulls will say the leverage will delever as Viterra synergies arrive; bears will say synergies arrive late, partial, and into a worse macro than was underwritten.

What bulls are extrapolating that won't hold. Bulls extrapolate (1) elevated crush margins into perpetuity, (2) renewable-diesel demand growing linearly through 2030, (3) Viterra synergies achieved on schedule, (4) Brazilian origination remaining a structural advantage, and (5) interest rates falling enough to ease the post-merger debt stack. Each of these is plausible individually; the joint probability is low. Mean reversion in commodity margins is the single most reliable empirical fact in this industry. Renewable-diesel mandates are politically negotiable and have already shown signs of plateauing. Synergies in cross-border ag mergers historically arrive at 50-70% of guidance. And rates are not in management's control.

Valuation trap. P/E TTM of 15.42 versus 10-year average of 9.42 means the multiple has expanded against earnings that are themselves elevated by recent cycle conditions — classic dual compression risk. Reverse-DCF implied growth is null because the base is negative. Intrinsic value bands per the scorecard are -$20.88 (low) to -$14.41 (high), all negative, against a market price of $124.61. Even allowing that the IV calculation is unreliable on a 3-year history (the scorer flagged this), the directional read is unambiguous: there is no scenario in this scorecard in which the stock is structurally cheap. A multiple compression to 8x on $5 of normalized through-cycle EPS would value the stock around $40. A worse cycle and a forced equity raise could take it lower.

If I am right, the stock could be worth $50 within 3 years.

Lollapalooza Bias Check

Multiple biases are tugging on me as the analyst, and the discipline is to name them.

Authority bias. The scorecard itself, with its composite of 69, has an authoritative numerical feel that can override the narrative read. The scorer's own notes explicitly flag this — 'insufficient history,' 'IV bands and 10y-ROIC less reliable; treat as exploratory,' base CAGR clamped from -27.5% to -5.0% — and the right move is to weight the scorer's own caveats more heavily than its composite. The composite is an aggregation of partial signals, not a verdict.

Anchoring. The current price of $124.61 anchors any thinking about 'cheap' or 'expensive.' The right anchor is through-cycle normalized earnings power, which I do not have a confident estimate of given the post-Viterra restructuring and 3-year financial history.

Recency bias and narrative bias. The renewable-diesel boom of 2022-2023 produced a salient bull story, and crush margins during that period are the most vivid recent data points. The 10-year ROIC of 0.0% is the more boring, more honest signal, and recency tends to suppress it.

Confirmation bias. Once the narrative crystallizes as 'commodity-cyclical, no moat, Too Hard,' I will preferentially weight evidence consistent with that frame. The mitigant is to ask honestly: is there a version of this where the Viterra integration produces a durable structural step-change in returns? The honest answer is 'maybe, but I cannot underwrite it from a 3-year history.'

Deprival super-reaction. The fear of missing a cyclical bottom is real. If crush margins re-tighten in 2027 due to a drought or a renewable-diesel mandate expansion, the stock could double from here. The mitigant: my mandate is compounders, not cyclical trades. Missing a cyclical double in a non-compounder is not a mistake against the mandate.

Incentive bias (in management, observed). Management's compensation is partly tied to returns on the Viterra deal succeeding, which biases their disclosures and capital-allocation choices toward narratives in which the deal works. I should not take their framing of synergy timing at face value.

The bias most likely to lead me astray is recency — extrapolating the renewable-diesel windfall as structural. The strongest counter-discipline is the 10-year ROIC of 0.0% and the negative IV across all three scenarios.

10-Year Outlook

On a 10-year horizon, will this be the same fundamental business model? Yes — Bunge will still be buying oilseeds and grain from farmers, processing them, and selling meal, oil, and grain to food, feed, and biofuel customers. The Viterra integration adds origination depth but does not change the shape.

Will the customer base be larger? Probably, in absolute terms, as global protein demand grows with emerging-market income. But Bunge's share of that demand will be contested by ADM, Cargill, COFCO, Louis Dreyfus, and regional consolidators in Brazil and India. There is no reason to expect Bunge's share to grow durably.

Will profit per customer be higher? Almost certainly not in real terms. The long-run economic shape of commodity intermediation is convergence to cost of capital, and Bunge's 10-year average ROIC of 0.0% is the empirical confirmation of that theory.

Will the moat be wider? No. There is no plausible mechanism by which a price-taking commodity processor's moat widens over a decade. Scale advantages can be matched, logistics networks can be built, and technology (digital procurement, on-farm data) is more likely to commoditize the trader function further than to entrench it.

The single biggest threat: a structural decline in U.S. and Brazilian crush margins driven by (a) renewable-diesel demand maturing, (b) Chinese demand growth slowing, and (c) Black Sea and South American supply elasticity reasserting. Stack on top: a higher-for-longer interest rate regime against the post-Viterra debt stack, and the 10-year base case is mediocre at best.

The 3-year financial history is too short to underwrite a high-confidence 10-year forecast in a business this cyclical. The scorer's own note — 'Short history (3y annuals); IV bands and 10y-ROIC less reliable; treat as exploratory' — is doing a lot of work here. Combined with the structural commodity-cyclical shape, this is not a compounder I can hold with confidence over a decade.

CONFIDENCE: low