Chapter 5: Protein
No food category generates more argument, more guilt, and more motivated reasoning than protein. Vegans and ranchers, nutritionists and animal rights activists, Silicon Valley technologists and Maasai herders — everyone has a position, and most positions involve telling someone else they're wrong.
This chapter isn't going to do that. It's going to trace the cost spectrum across protein sources, from the highest-impact to the lowest, showing how context shifts every food's position on that spectrum. The goal is clarity, not conversion.
Industrial Beef: The Outlier
Start with the biggest number. According to Poore and Nemecek's 2018 meta-analysis, beef produces roughly 60 kg of CO₂ equivalent per kilogram of protein — six times more than chicken, thirty-six times more than legumes [VERIFY]. It requires approximately 164 square meters of land per 100 grams of protein [VERIFY]. Its water footprint averages around 15,400 liters per kilogram [VERIFY].
Beef is the outlier. Not all meat. Beef specifically.
The reasons are biological. Cattle are ruminants — their four-chambered stomachs host bacteria that ferment cellulose, allowing them to digest grass. That fermentation produces methane, which the animals belch. A single cow produces 70 to 120 kg of methane per year [VERIFY]. Multiply by roughly one billion cattle on Earth, and you get approximately 14.5 percent of global greenhouse gas emissions from all livestock, with cattle accounting for about 65 percent of that [VERIFY].
Then there's the feed. In industrial systems (feedlots or CAFOs — Concentrated Animal Feeding Operations), cattle spend their final months eating grain, primarily corn and soy. Growing that feed requires land, water, fertilizer, and pesticides. The conversion ratio is poor: it takes roughly 6 to 8 kg of grain to produce 1 kg of beef [VERIFY]. You're eating grain filtered through a cow, with enormous losses at each step.
And the land. Approximately 80 percent of deforested Amazon land is used for cattle pasture [VERIFY]. Brazil is the world's largest beef exporter, and the link between Brazilian beef and deforestation is documented in detail by organizations like Trase, Mighty Earth, and Global Witness. Companies like JBS — the world's largest meatpacker, processing roughly 75,000 head of cattle per day [VERIFY] — have repeatedly been linked to supply chains that include illegally deforested land, despite pledges to the contrary.
Industrial beef, by virtually any measure, sits at the extreme end of the cost spectrum.
Pastured Beef: The Complication
But "beef" is not one thing.
A cow grazing on permanent pasture in the Scottish Highlands or the Argentine pampas is a different animal — systemically, if not taxonomically — from a feedlot steer in Kansas eating Iowa corn.
Grass-fed, pasture-raised beef has genuine advantages: no feed-crop footprint, no feedlot waste concentration, potential for rotational grazing that builds soil carbon. The Savory Institute and regenerative agriculture advocates argue that well-managed grazing can sequester enough carbon in soil to offset the methane emissions — making grass-fed beef carbon-neutral or even carbon-negative.
The evidence is contested. Peer-reviewed analyses, including a major review by Garnett et al. (2017) at the Food Climate Research Network, concluded that while grazing can increase soil carbon, the sequestration is "small, time-limited, reversible, and substantially outweighed by the greenhouse gas emissions" from the cattle themselves [VERIFY]. The Savory claims have been called "wishful thinking" by some scientists and "the best available management practice" by others.
What is clear: pastured beef uses more land per kilogram than feedlot beef (slower growth means longer lives means more pasture) and produces more methane per kilogram (slower growth means more lifetime belching). Its total carbon footprint may be comparable to or even higher than feedlot beef — the reduced feed-crop emissions are offset by increased land use and lifetime emissions.
The labor and justice picture is often better. Small-scale ranching supports rural communities, maintains cultural landscapes, and avoids the nightmarish conditions of feedlot operations. The welfare picture is incomparably better — a cow on grass has a life; a cow in a feedlot endures an existence.
HIGHER COST ←——————————————————————→ LOWER COST
Feedlot beef Grass-fed beef from Smallholder cattle
(Amazon soy feed, well-managed pasture in mixed farming
methane, water, (more land, but soil system (draft power,
deforestation, health, no feed crop, manure fertilizer,
worker exploitation) better welfare) cultural significance,
subsistence livelihood)
That last column matters. For roughly one billion people globally, cattle are not a consumer product — they're a livelihood, a savings account, a source of draft power, and a cultural institution. A Maasai herder's relationship to cattle is not the same as an American's relationship to a Big Mac. Any honest analysis must hold this distinction.
Chicken: Efficient and Terrible
If beef is the carbon outlier, chicken is the efficiency champion — at an appalling welfare cost.
Chicken produces roughly 6 to 7 kg of CO₂ equivalent per kilogram, about one-tenth of beef [VERIFY]. The feed conversion ratio is far better: roughly 2 kg of feed per kg of meat [VERIFY]. Water footprint: about 4,300 liters per kilogram, less than a third of beef [VERIFY]. Land use: far more modest.
By the environmental numbers, chicken is among the least impactful animal proteins. The industry has been remarkably effective at optimizing feed conversion, growing chickens from hatch to slaughter weight in approximately 42 days [VERIFY] — down from roughly 70 days in the 1950s.
That speed is the problem.
Modern broiler chickens have been bred to grow so fast that their skeletons can't keep up. Studies have found that roughly 25 to 30 percent of broilers suffer from leg problems, including lameness [VERIFY]. Their hearts and lungs struggle to support their unnaturally rapid growth — sudden death from heart failure (flip-over disease) is a recognized production loss. They live their short lives in densely packed barns, typically 20,000 to 30,000 birds per shed [VERIFY], on litter that may not be changed for the entire growing cycle.
Processing is its own cost dimension. Poultry processing plants run at extraordinary speed — the USDA allows line speeds of 140 birds per minute [VERIFY]. Workers, predominantly immigrants and people of color, suffer some of the highest injury rates in American manufacturing. Repetitive motion injuries, knife cuts, and exposure to chemicals are endemic. An Oxfam report documented workers wearing diapers because they weren't given bathroom breaks [VERIFY].
The chicken on your plate is cheap because the costs — animal suffering, worker injury, antibiotic resistance (roughly 70 percent of medically important antibiotics sold in the US go to livestock, with poultry a significant user [VERIFY]) — are externalized.
Eggs: A Spectrum of Welfare
Eggs compress the entire welfare debate into a single product.
- Battery cages (still legal in most US states [VERIFY]): 67 square inches per hen — less than a sheet of letter paper. The hens can't spread their wings. Some are de-beaked to prevent stress-induced pecking. This system produces the cheapest eggs.
- Cage-free: No cages, but often 20,000+ hens in a warehouse with no outdoor access. Better than battery, but the label overpromises.
- Free-range: Access to the outdoors — but the "outdoors" may be a small concrete patio accessed through a single door that most birds never find.
- Pasture-raised: Meaningful outdoor access on grass. The best commercial welfare standard. Also the most expensive — $6 to $8 per dozen vs. $2 to $3 for conventional [VERIFY].
- Backyard chickens: The gold standard if you have the space. Fresh eggs, plus the chickens eat kitchen scraps (waste reduction), control pests, and produce fertilizer. Zero transport, zero packaging.
Environmentally, the differences between egg production systems are smaller than the welfare differences. All egg production has a moderate footprint — roughly 4 to 5 kg CO₂eq per kilogram [VERIFY]. The environmental case for choosing pasture-raised over caged is less about climate and more about the kind of food system you want to support.
Fish: Wild Complexity
Fish might be the most confusing protein category because the range is so extreme. "Fish" includes sardines (tiny, abundant, low-impact) and bluefin tuna (apex predator, critically endangered, sold for millions at Tokyo auctions). Saying "eat more fish" is like saying "eat more plants" — it depends entirely on which one.
Wild-caught fish — the concerns:
- Bycatch: Industrial fishing operations catch and discard enormous quantities of non-target species. Bottom trawling for shrimp can produce bycatch ratios of 10:1 — ten pounds of unwanted marine life killed for every pound of shrimp [VERIFY]. This includes juvenile fish, sea turtles, dolphins, and sharks.
- Stock depletion: Roughly 34 percent of global fish stocks are overfished, and another 60 percent are fished at maximum capacity [VERIFY, FAO 2022]. Sustainable fisheries exist — but they require strict management, quotas, and enforcement.
- Habitat destruction: Bottom trawling scrapes the ocean floor, destroying coral, sponge beds, and benthic ecosystems. It's been compared to clear-cutting a forest to catch a deer.
Farmed fish (aquaculture) — different concerns:
- Feed: Many farmed fish (salmon, trout) are carnivorous and fed pellets made from wild-caught fish. It can take 2 to 5 kg of wild fish to produce 1 kg of farmed salmon [VERIFY]. You're not reducing pressure on ocean ecosystems; you're redistributing it.
- Pollution: Concentrated fish farms produce waste — nutrients, antibiotics, sea lice treatments — that pollutes surrounding waters. Norwegian salmon farms have devastated wild salmon populations in some fjords through sea lice transmission [VERIFY].
- Antibiotics: Chilean salmon farming uses enormous quantities of antibiotics — approximately 500 tons per year [VERIFY], contributing to antibiotic resistance.
The species that work: Sardines, anchovies, herring, mackerel — small, fast-reproducing forage fish that are abundant when well-managed. They're low on the food chain (low bioaccumulation of mercury and PCBs), nutritious (omega-3 rich), cheap, and can be fished sustainably. Also: mussels, clams, oysters — filter-feeding bivalves that require no feed, actually clean the water they grow in, and have among the lowest environmental footprints of any animal protein.
Monterey Bay Aquarium's Seafood Watch provides the best accessible guide — green/yellow/red ratings by species and catch method. It's not perfect, but it's the most rigorous consumer-facing tool available.
Lentils and Beans: The Quiet Champions
If this book were assigning winners, legumes would take the prize. And then nobody would buy the book, because nobody wants to be told to eat more lentils.
The numbers are almost embarrassingly good. Lentils produce roughly 0.9 kg CO₂eq per kilogram [VERIFY]. Water footprint: about 5,800 liters per kilogram, but mostly green water (rainfall) [VERIFY]. Land use: minimal. They fix nitrogen in the soil through symbiotic bacteria, reducing the need for synthetic fertilizer and actually improving soil fertility for subsequent crops.
Nutritionally: high in protein (about 25 percent by dry weight [VERIFY]), iron, folate, and fiber. Cheap — among the least expensive protein sources on Earth. Long shelf life — dried lentils last years. Low waste — minimal processing, minimal packaging, no cold chain required.
They score well on labor and justice too. Lentils and beans are typically grown by smallholder farmers in India, Turkey, Canada, and East Africa. The supply chains are less concentrated, less exploitative than commodity meat or commodity crops. No child labor scandals. No worker injury epidemics.
So why aren't they the default?
Partly culture — Western diets are built around meat as the centerpiece of the plate, with legumes relegated to side-dish status or invisible as an ingredient. Partly marketing — nobody runs Super Bowl ads for lentils. Partly convenience — dried beans require soaking and cooking time. Partly the gas — let's be honest, the oligosaccharides in beans cause flatulence, and this is a real barrier for some people (though your gut microbiome adapts within a few weeks of regular consumption [VERIFY]).
India, where lentils (dal) are a dietary staple for hundreds of millions of people, demonstrates that a legume-centered diet is culturally viable, nutritionally adequate, and environmentally superior. It's not a novelty — it's a proven system.
Tofu and Tempeh: The Soy Distinction
The soy conversation requires one essential fact: roughly 77 percent of global soy production goes to animal feed [VERIFY]. Another 6 percent or so is consumed directly by humans as tofu, tempeh, soy milk, and edamame. The rest goes to soy oil and industrial uses.
When someone says "soy drives deforestation," they're technically correct — soy production in the Brazilian Cerrado and parts of the Amazon is a deforestation driver. But the soy that becomes your tofu is almost certainly not that soy. Human-grade soybeans for tofu are typically non-GMO varieties grown in the US, Canada, or Japan, processed in dedicated facilities. The soy that drives deforestation feeds chickens, pigs, and cattle in Europe and China.
Tofu's environmental footprint is modest: roughly 2 to 3 kg CO₂eq per kilogram [VERIFY], moderate water use, minimal land. It's a complete protein with all essential amino acids. Tempeh — fermented soybeans — is nutritionally superior (higher protein bioavailability, B12 from fermentation, easier to digest) and has a similarly low footprint.
The processing is simple. Tofu has been made by essentially the same process for over 2,000 years: soak soybeans, grind, boil, coagulate with mineral salts, press. This is processing — but it's processing the way bread-baking is processing. Comparing tofu to a Beyond Burger because both are "processed plant protein" is like comparing bread to a Pop-Tart.
Insects: The Future Nobody Wants
More than 2 billion people already eat insects regularly — primarily in parts of Asia, Africa, and Latin America [VERIFY]. Roughly 2,000 species of edible insects have been documented. Insect farming is not a future technology — it's an ancient practice being rediscovered.
The efficiency is remarkable. Crickets require about 1.7 kg of feed per kg of body weight, compared to 2 kg for chicken and 8 kg for beef [VERIFY]. They reproduce faster, require less water and land, and produce a fraction of the greenhouse gas emissions. About 80 percent of a cricket is edible, compared to roughly 55 percent of a chicken and 40 percent of a cow [VERIFY].
In Western markets, insect protein is growing — cricket flour in protein bars, mealworm snacks, black soldier fly larvae as animal feed. The EU approved several insect species for human consumption between 2021 and 2023 [VERIFY].
The barrier is psychological, not nutritional or environmental. Western disgust toward insects is cultural, not universal. If you eat shrimp (an arthropod) and honey (insect vomit, essentially), the logical case against crickets is weak. But food is not logic. It's culture, memory, and identity. The transition, if it happens, will be slow.
Lab-Grown Meat: The Energy Problem
Cultivated meat — muscle tissue grown from animal cells in bioreactors — was the darling of food-tech investment from 2015 to 2023. The promise: real meat without the slaughter, the methane, the land, the water. Upside Foods and Good Meat received the first USDA approvals for cultivated chicken in June 2023 [VERIFY].
The reality check came from the energy numbers.
A 2023 study from UC Davis found that, at current production scale, cultivated meat's carbon footprint could be 4 to 25 times higher than conventional chicken, depending on the energy source and purification requirements [VERIFY]. The bioreactors require pharmaceutical-grade sterile conditions, growth media that currently use fetal bovine serum or expensive recombinant proteins, and enormous amounts of energy to maintain temperature and agitation.
If the energy comes from renewable sources, the footprint drops substantially. If growth media becomes cheaper (a major research focus), costs could fall from the current ~$50-100/kg to something approaching conventional meat [VERIFY]. But "could" and "does" are different words.
The most honest assessment: cultivated meat is a technology with theoretical promise and current limitations. It may eventually be competitive on cost and carbon if several engineering breakthroughs converge. Or it may turn out to be the food equivalent of flying cars — perpetually five years away.
The Protein Spectrum
Here's the broad picture, from highest to lowest total system cost:
HIGHER COST ←————————————————————————→ LOWER COST
Industrial Farmed Chicken Eggs Tofu/ Small Lentils/
beef salmon (efficient (moderate tempeh pelagic beans
(extreme on (feed, calories, footprint,(low fish (lowest
nearly every antibiotics,terrible welfare impact, (sardines, footprint,
dimension) sea lice) welfare) spectrum) soy ≠ mussels, N-fixing,
Amazon) oysters) cheap,
nutritious)
But remember: this is the median. The range within each category is enormous. Backyard eggs from your neighbor's chickens have a lower footprint than imported organic lentils flown from Turkey. Wild Alaskan salmon from a well-managed fishery scores differently than Chilean farmed salmon pumped full of antibiotics.
Context. Always context.
The single most impactful change most Western consumers can make is not going vegan — it's reducing beef specifically. Swapping beef for chicken, pork, or legumes even two or three times per week captures the majority of the environmental benefit. Going from daily beef to weekly beef is a larger reduction in carbon footprint than going from weekly chicken to fully vegan [VERIFY].
Not perfection. Direction.