THE MORPHOLOGY OF FOOD

There is something poetic in the way the food chain is built. To become food, meat also has to eat something. A cow eats grass. A pig eats grain. A chicken eats meal. A salmon eats fishmeal made from anchovies that themselves ate plankton. Every level is a conversion: someone eats so that someone else can become food.

At the top level — humans. In North America a person consumes about 78 kg of meat a year. In Africa — about 9.6 kg. An eightfold gap, and it is not accidental: meat has always been a marker of prosperity. In China, Brazil, Southeast Asia consumption is rising quarter by quarter, somewhere thanks to Deng Xiaoping, somewhere as a consequence of the AI boom. The human diet is "improving."

In inverse proportion to this improvement, the composition of what is fed to animals is changing. Fishmeal in salmon feed — from 60% in 1990 to 12% in 2020. EPA+DHA¹ in the salmon itself — from 2.5 g to 1.8 g per 100 g. Woody breast² prevalence in broilers — from 5% in 2012 to 50% in some flocks by 2024.

Biologically the reactor is working fine. FCR³ in chicken has improved threefold over seventy years. The animal converts feed into protein more efficiently than ever. Only it is a different protein.

A loop? A paradox? More like shorting it.

A protein bubble that everyone sees in the data, but that no one talks about. The resource ceiling of feed protein has been visible since 1989 — when the world fish catch for reduction peaked and never grew again. Thirty-five years of plateau. Three price crises, each of which set a new baseline. And in parallel — a quiet degradation of the product that is not reflected in any headline, because the packaging says the same thing.

October 2012. Zacky Farms — the second-largest poultry producer in California, 70 years on the market — files for bankruptcy. The official reason from the company's statement: "historically high prices of corn and soybean meal." No financial crisis. No bloated debt. Just a Midwest drought that raised corn from $197 to $327 per ton — and the margin evaporated.

December 2008. Pilgrim's Pride — at the time the largest chicken producer in the United States — Chapter 11 bankruptcy. Corn +29%, soybean meal +63% in a single quarter. Feed costs grew by $800 million over the year. Here the picture is more complicated: the company was carrying $1.9 billion of debt from the purchase of competitor Gold Kist two years earlier, and the financial crisis had frozen credit. Grain finished off a company that was already vulnerable for three other reasons. The CEO said it precisely: "We were unable to raise prices fast enough to offset the volatility of the grain market."

2007–2010. Chile. The infectious salmon anemia virus (ISA) appeared on a single farm in July 2007. Within a few months it had hit 65% of Atlantic salmon farms in two regions. Production collapsed from 400,000 to 100,000 tons per year. $550–600 million in losses. 25,000 jobs. Salmon growing shoulder to shoulder in tight quarters showed zero resistance to the virus and dropped dead.

---

I. The Reactor

An animal is a conversion reactor. Feed goes in. Protein that we eat comes out. Everything else is servicing the reactor. Buildings, labor, veterinary, young stock.

Beef: $2.00/kg of feed — finishing phase only (90–120 days, FCR 6:1). Before that, 12–18 months on pasture: grass ~$0.01–0.04/kg. Feed inside feed. Sources: Southern Ag Today 2023; Norwegian Directorate of Fisheries 2024; USDA ERS 2024.

Feed is between 44 and 72% of the cost of meat depending on the species. But "feed" is not a homogeneous line. Inside it there is a hierarchy.

Broiler diet: 63% corn ($200/t) — energy. 26% soybean meal ($400/t) — protein. Plus synthetic amino acids: DL-methionine ($2,700/t) and L-lysine ($1,600/t) — small by share, but critical: without them soy protein is incomplete. All in, 1 kg of feed: $0.30–0.34. The protein component alone: ~$0.10–0.12 — that is one third of feed cost. Compound feed manufacturers: Cargill Animal Nutrition, CP Group (30%+ of the Asian market), ADM, De Heus. The poultry feed market — $223 billion (2025).

Pork is arranged similarly: 60–65% corn, 20–25% soybean meal, a small share of fishmeal in the starter diet for piglets. Feed: $0.25–0.28/kg. The protein share — ~30–38% of this sum.

A beef feedlot (US, 90–120 day phase): 65–70% corn, 15–20% dry DDG (a byproduct of ethanol production, 27–30% protein, its price tied to the corn market), 5–10% soybean meal or feed urea. Elanco's ionophore Rumensin: improves rumen fermentation, minus 0.15–0.25 to FCR for $0.003/kg. All in: $0.19–0.21/kg of feed. The protein share — ~18–26%. It looks small, but at an FCR of 6:1, 10 kg of feed goes into 1 kg of carcass — and two dollars on food alone.

Salmon is a different order of numbers. Fishmeal (12% of the diet, $2,000/t), fish oil (6%, $2,700/t), soy concentrate (22%, $620/t). Three key ingredients total: $0.54 out of $0.71/kg of feed — 76%. Three quarters of the cost of feed is protein and specific fats. The salmon feed producers who control ~85% of the market: Skretting (Nutreco/SHV Holdings), Cargill Aqua Nutrition, BioMar (Denmark).

The more expensive the animal protein — the higher the share of protein ingredients in its feed. Salmon at the top: three quarters of the cost of feed is protein. That is exactly why a shock in the fishmeal market hits salmon disproportionately harder than it hits chicken.

---

II. Seventy Years

FCR — Feed Conversion Ratio. How many kilograms of feed are needed for the animal to gain one kilogram of live weight. At a feed price of $0.32/kg, the difference between FCR 4.70 and FCR 1.64 is $0.82 versus $0.52 per kilogram. Across three billion kilograms of global chicken production — that is a difference on the scale of a small country's GDP. Since the 1950s this has been the main KPI of all industrial animal farming.

For chicken, in 1925 FCR was 4.70. In 1957 — 2.80, the bird weighed 586 g after 84 days. Havenstein took preserved 1957 genetics, grew them alongside modern lines on the same ration — a fivefold difference in weight. This is not an improvement of the same bird. It is a different biological object. By 2020, FCR 1.64, bird 3.1 kg in 42 days. 80–90% of the progress is genetics: three companies (Aviagen, Cobb, Hubbard) control 80% of global broiler genetics. The key mechanism: the gut shrank from 9% to 5.2% of live weight — less organ, less energy on its work, more on muscle. The biological ceiling: ~1.35–1.40. The world record — New Zealand: 1.38.

For beef the path is slower: ~10:1 in the 1950s, ~5.5–6.0:1 in the American feedlot today. In the EU, Russia, China — 6.0–7.0:1, because hormonal implants and beta-agonists are banned there (they add +15–20% to gain). One package labeled "beef" — different biochemical processes behind it. The ceiling without hormones: ~5.5:1.

For pork industrial FCR 2.5–3.0:1 — the best among land animals. A short cycle (6 months), monogastric⁴, a highly concentrated diet. The biological ceiling: ~2.0–2.2:1.

For salmon the steepest story. In the 1980s FCR 6–7:1 on wet feed — chopped fish. The main jump happened not because of genetics: in 1990–1995 the industry switched from wet feeds to extruded dry pellets. FCR immediately jumped to 2.0–2.5:1. Not fifty years of work — five years and a change of format. Today Norway 1.25, Chile 1.35, Canada 1.49. The biological ceiling: ~1.0–1.1.

What happens when you approach the conversion ceiling closely? Biology sends the bill.

The broiler's breast muscle grew from 12% of body in the 1970s to 25% in the 2020s. The capillaries did not keep up with the volume: tissue hypoxia at a depth of more than two centimeters. The muscle degenerates — it is replaced by collagen and fat. This is called woody breast. 5% prevalence in 2012, 29% by 2015, up to 50% in some flocks by 2024 (Frontiers in Physiology, 2024). It cannot be eliminated without slowing growth: heritability h²=0.50⁵, genetic correlation WB↔WS r=0.88⁶. Affected meat is not sold as "chicken breast" — it goes into McDonald's and KFC nuggets, ready meals and preserves. Without labeling. Muscle protein with a high-quality amino acid profile is replaced by collagen, whose profile is fundamentally different: almost no tryptophan⁷, little methionine⁸.

Salmon has a different degradation. In 1990, 90% of Norwegian feed was made of marine ingredients. In 2020 — 22%. It is important not to mix up causation: FCR in salmon improved in 1990–1995 because of the switch to extruded dry pellets — that is a separate technological event. The replacement of fishmeal by soy and rapeseed is a different decision, made later for economic reasons. Two different processes in one time window. The result of the second: EPA+DHA¹ content dropped from ~2.5 g/100 g in the 1990s to ~1.8 g/100 g by 2024 — data from the Norwegian Institute of Nutrition. The protein is the same. What they charge a premium for is not. The label is the same.

---

III. Three Cases

Zacky Farms, October 2012. A Midwest drought. Corn from $197 to $327/t (+66%) in a few weeks. In the state of Georgia alone, poultry farmers received $430 million of additional feed costs per year (University of Georgia, 2012). Zacky Farms had neither bloated debt nor bets on grain futures. Margin simply evaporated. A 70-year-old company, 1,500 employees — bankruptcy on October 8, 2012. The management's statement: "historically high prices of corn and soybean meal."

Pilgrim's Pride, December 2008. The same mechanism — a feed shock of $800 million over the year — but the company was carrying $1.9 billion of debt from the purchase of a competitor, the financial crisis had frozen credit, consumer demand fell. Grain finished it off, but didn't kill it alone. An important distinction from Zacky: with high financial leverage, feed kills faster. The CEO put it honestly: "We were unable to raise prices fast enough to offset the volatility of the grain market."

Chile, 2007–2010. Not an external economic shock — a systemic one.

The farms were ultra-dense — because the economics required maximum output per unit of licensed area. It is the same logic as FCR optimization: squeeze the maximum protein out of the minimum resource. It was precisely the high stocking density that gave ISA its speed of spread — within a few months the virus had hit 65% of farms in two regions. In parallel: salmon on a high share of plant feed have a different immune profile than on a marine one. When the industry replaced fishmeal with soy to reduce costs, it simultaneously reduced the fish's immune resilience.

After the collapse, during recovery, the industry was forced to raise the share of fishmeal and fish oil in feed — to support the immunity of the restocked population. That is to say, a disease that arose in part because of feed optimization, when exiting it, created additional demand for that same scarce marine protein. The price mechanism is the same: Chile's production — 27% of the world's salmon — fell by 75%, and world prices jumped immediately. Norwegian operators reaped windfall profits. This is exactly the same zero elasticity of supply as El Niño with fishmeal.

Zacky Farms: an external shock killed the company through feed. Pilgrim's Pride: the same shock finished off a weakened company faster. ISA in Chile: the internal logic of optimization created a biological fragility that collapsed at the first shock. Three different paths to the same place.

---

IV. 75 Years: Three Stable Periods

Over seventy-five years, feed protein prices had three calm periods — about forty years in total. All the rest of the time — crises, each of which shifted the baseline of prices upward without reversal.

1950–1971: about twenty years of silence. The Green Revolution, cheap energy, stable world demand for meat. Corn $55–80/t, fishmeal ~$100/t.

1972–1974: the first big shock. Three independent blows at once. The USSR unexpectedly entered the world market with the largest grain purchase in history — about $750 million in a single season because of an internal drought. In parallel, the El Niño of 1972 destroyed the Peruvian anchovy catch: in 1970 Peru caught 12 million tons, two years later — seven times less. In parallel, the 1973 oil shock raised fertilizer and logistics costs. Corn +150%, fishmeal +250% — from $100 to $350/t. Dutch farmers declared a "protein crisis" and switched to soybean meal from Latin America. It was precisely then that the structural dependence on Argentine soy was laid down — which came back to bite in 2022.

1975–1993: eighteen years of stability. Latin America scaled up soy and fishmeal production. Prices hold, the meat industry builds growth models.

1994–1996: droughts in the United States, record-low stocks, corn hits highs. But the 1997 Asian financial crisis doused the wave — demand in Asia collapsed, prices rolled back.

2000–2006: six years of the last calm. Corn averaging $86/t ($2.18/bushel), soybean meal $187/t. Record harvests, low Chinese demand, cheap energy. It was from these prices that Pilgrim's Pride built its financial model when buying Gold Kist in 2006. (those who shorted them saw the same chain)

2007 — and since then: a structural shift. Not a crisis that will end. A different baseline forever. The US ethanol mandate redirected a third of the corn harvest into fuel. The developing world — China, India, Brazil — began to eat meat on a scale that had never existed before. Financial investors discovered grain futures as an asset class. Corn: average 2007–2025 $190/t ($4.83/bushel), soybean meal $362/t — plus 121% and plus 94% to the 2000–2006 level. No return in eighteen years. 2012: drought — corn to $327/t ($8.30/bushel). 2022: Argentina La Niña — soy harvest from 44 to 25 million tons in one season. 2022–2023: Ukraine — corn to $315/t, wheat +77%. After each peak prices roll back, but the baseline stays higher than the previous one.

---

V. The Protein Inside the Feed

Feed is not just grain. Inside it is the protein without which growth does not happen and FCR does not work. This protein has three sources, each with its own ceiling.

Soybean meal trades on the CBOT. About 70% of world production goes into animal feed. The US, Brazil, Argentina control 80%+ of world exports. Supply growth: 0.8–1%/year. The drought in Argentina in 2022–23: −45% of the harvest in a single season. This is not an anomaly — this is a recurring pattern. Another factor: the Chinese state mandate to reduce the share of soybean meal in feeds from 20–25% to 10% by 2030. This creates a multi-million-ton gap inside a single country — which has to be closed with something else.

Rendering — the processing of slaughter by-products into meat-and-bone, blood, and poultry meal. A $23.5 billion/year market, growing 1–2%/year. Rigidly tied to livestock numbers. The US cattle herd in 2025 is at its lowest since 1952. Tyson is closing plants. ASF 2018–20 wiped out 25% of the world pig herd. The rendering flow shrinks together with the herd.

Fishmeal trades through IFFO. Production: 5.4–5.6 million tons/year — a plateau for more than thirty years. The world catch of fish for reduction fell from 35 million tons in 1989 to ~17 million tons by 2022. 39% of fishmeal is already produced from heads, backbones and processing waste — not from whole fish. Peru provides ~20% of world production from a single species, the Peruvian anchovy, which does not spawn in warm water. The price elasticity of fishmeal supply, according to the World Bank, is close to zero: quotas are limited by the biology of the populations, not by the market price. This means: every next percent of deficit produces a disproportionately large price increase.

Total supply growth from all three sources: a maximum of 1–1.5%/year. Fishmeal — zero.

---

VI. Who Else Eats This Protein

Fishmeal is the scarcest feed protein. To understand the scale of the problem, one has to see who exactly is competing for it.

Sources: Science Advances 2024 (global FMFO use review); FAO State of World Fisheries 2024; IFFO Changing Demands 2022; Rabobank "Hooked on Scarcity" Sept. 2025. Share data — 2021.

The key point of this table: pet food and nutraceuticals pay more for fish oil than aquafeed does — they have higher margins, they can absorb a price rise. Aquaculture often cannot. That is exactly why the structural deficit of fish oil will hit aquafeed first: the oil will flow to where they are willing to pay more for it.

---

VII. The Arithmetic of Deficit

Put it all together.

FCR in every species is approaching the biological ceiling. This means: further growth of meat production can no longer be achieved through efficiency — more animals are needed. More animals — more feed. More feed — more protein on the input.

World chicken production is growing ~3%/year. FCR improvement — ~0.5–0.8%/year. Net new demand for feed: +2.2–2.5%/year. With production growing from 130 to 145–150 million t by 2033 and a moderate FCR improvement from 1.64 to ~1.55 — in total ~12 million additional tons of feed are needed. Of these, 20% is protein — plus 2.4 million tons just for chicken.

For salmon: FCR is already close to the ceiling (Norway 1.25 at a biological minimum of 1.0–1.1). Further reduction of the fishmeal share in the diet is impossible without loss of survival and immunity (the ISA lesson in Chile). Production growth of ~5–6%/year = direct growth of fishmeal demand with no compensation from efficiency.

Rabobank (September 2025, "Hooked on Scarcity"): additional aquaculture demand for fishmeal by 2033 — +1.2–1.3 million tons, that is 25% of current world production. Structural deficit — by 2028. Nature Food (2025): in order not to halt the growth of aquaculture, 1.8 million tons of alternative ingredients are needed annually. For fish oil there is a separate story: Rabobank's forecast of a 20,000-ton shortage in 2025 did not materialize — Peru had a good season. But from 2026 production is falling again. The structural problem has not gone anywhere — the date has shifted.

Total demand for feed protein — 3–4%/year. Total supply growth from the three sources — a maximum of 1–1.5%/year. The gap accumulates every year.

---

— The Consequence

The biological ceiling of FCR and the structural deficit of feed protein are two descriptions of the same problem. When chicken FCR reaches 1.38, further growth of production requires more animals, which means more protein on the input. The three sources of this protein grow at 0.8–2%/year. Fishmeal — zero for thirty years.

Every time a deficit with zero supply elasticity appears in this system, the price jumps disproportionately. 1972: fishmeal +250%. 2014: +66%. 2023: +66%. Every time a new baseline is set. This is exactly how Zacky Farms goes bankrupt. This is exactly how Pilgrim's Pride collapses. This is exactly how in Chile, after the ISA crisis, recovery creates even greater demand for that same scarce protein.

A source of protein without a biological ceiling is needed. One that does not depend on the climate of three countries. Not tied to the quotas on a single species of fish. Scalable — increasing volume in proportion to demand, not in spite of the weather.

The most expensive segment is aquafeed and premium pet food. It is precisely there that willingness to pay for protein is highest, sensitivity to the quality of the amino acid profile is highest, and the deficit of the coming years is highest. That is where alternative protein will enter first — because there it is already price-competitive today.

---

Summary. Protein and fatty acids per 1 kg of product

Sources: Norwegian Institute of Nutrition (EPA+DHA in salmon 1990–2020); MDPI Food Sciences 2024 (woody breast); USDA FoodData Central.

---