The whole entry collapses to a label rule: under 3 grams of protein per cup means the carton isn't doing milk's job. Fortified soy or pea match cow milk on protein and calcium when shaken; the rest are flavored, fortified water at two to four times dairy's price. Read the protein number, shake the carton, and check for B12 fortification if you're vegan.
Every plant milk is the same process: mill or soak the source ingredient, blend it with water, strain the solids out, then add stabilizers, oil for mouthfeel, and a fortification premix of calcium, vitamins, and sometimes B12. The ratios are extreme. A liter of commercial almond milk contains roughly twenty grams of almonds. A liter of oat milk uses an enzyme called alpha-amylase to chop the oat starch into shorter sugars — the trick that makes oat milk taste sweet without "added sugar" appearing on the label, and that pushes its blood-sugar response into roughly the same range as white bread Atkinson et al. 2008.
Soy and pea milk are the exceptions. Their manufacturing target is a protein-concentrated stream, not a flavored slurry — the finished beverage is engineered to land near cow milk's 3.3 grams of protein per 100 mL. Everything else in the category dilutes its source ingredient until the protein is mostly gone Vanga & Raghavan 2018.
The calcium that makes a fortified plant milk "equivalent to dairy" on the label isn't chemically bound to the milk the way cow milk's calcium is bound to its protein. It is a powder — usually calcium carbonate or tricalcium phosphate — suspended in the liquid, and it settles to the bottom of the carton between shakes. A pour from the top of an unshaken carton can deliver less than a third of the calcium printed on the label Singhal et al. 2017.
The type-by-type ranking
Three independent nutritional comparisons of the US plant-milk market — Vanga and Raghavan in 2018, Chalupa-Krebzdak and colleagues the same year, and Drewnowski's 2021 audit of 219 branded products — converge on the same rank order. Soy milk is the only plant milk that consistently matches cow milk's nutrient profile. Pea milk matches on protein and most fortified micronutrients. Almond, coconut, rice, oat, cashew, and hazelnut milks fall far enough behind on protein that the gap is structural, not a fortification problem Vanga & Raghavan 2018, Chalupa-Krebzdak et al. 2018, Drewnowski 2021.
Soy milk delivers 7 to 9 grams of protein per cup, all nine essential amino acids in roughly the proportions a human body needs, and a protein-quality score (DIAAS) around 0.91 — within striking distance of cow milk's 1.0 Mathai et al. 2017. Soy protein stimulates muscle protein synthesis at rest and after exercise about as well as casein, the slower of cow milk's two proteins Tang et al. 2009. Soy's isoflavones — plant compounds with weak estrogen-like activity — show modest cohort-level associations with lower postmenopausal fracture risk and small reductions in LDL cholesterol Zhang et al. 2007, Messina 2016.
Pea milk matches soy on protein quantity (8 grams per cup) with a DIAAS around 0.7 to 0.8 — slightly behind soy and cow milk, easily made up by drinking 20% more. Pea milk has no isoflavones, which some readers prefer.
Oat milk delivers 1 to 4 grams of protein depending on brand, with most commercial versions at 2 or 3. The beta-glucan that does cholesterol-lowering work in oat flour trials is mostly strained out during milk production: a cup of oat milk delivers roughly 0.4 to 1 gram of beta-glucan, well below the 3 grams per day the FDA requires for a qualified cholesterol-lowering claim Whitehead et al. 2014. The enzyme-broken-down starch pushes oat milk's glycemic response higher than most readers expect.
Almond milk delivers about 1 gram of protein per cup — sometimes less. A whole almond is mostly fat and a respectable amount of protein, but the milk is 2% almonds: most of the almond's nutrition is in the strained-out solids, sold separately as almond meal.
Coconut, cashew, hazelnut, and rice milks all sit at zero to 1 gram of protein per cup. Rice milk additionally inherits the rice grain's inorganic arsenic load, enough that the FDA explicitly advises against it as a primary milk for infants and young children FDA 2016.
For calcium specifically, the news is better than the protein news. Heaney and colleagues showed in 2000 that calcium carbonate-fortified soy milk, when properly shaken, delivers calcium your gut absorbs about as well as it absorbs the calcium in cow milk — fraction absorbed around 24 to 30% in both cases Heaney et al. 2000. A 2005 follow-up confirmed the same for tricalcium phosphate fortification Zhao et al. 2005. The fortification chemistry works. The problem is delivery: settling, partial fortification across the market, and skipping shake-the-carton.
For children, the clinical-outcome data are unusually clean. Morency and colleagues followed roughly 5,000 Canadian toddlers and found that children aged two to six who drank non-cow milk were shorter in dose-response: 0.4 centimeters shorter for each daily cup of plant milk, attributed to the lower protein and fat density of the substitute rather than to calcium Morency et al. 2017. The American Academy of Pediatrics, the Academy of Nutrition and Dietetics, and the American Heart Association issued a joint statement in 2019 recommending fortified soy as the only acceptable plant-milk substitute for cow milk in children under five Lott et al. 2019. The USDA's 2020-2025 Dietary Guidelines say the same for adults — fortified soy is the only plant milk classified as a dairy equivalent USDA & HHS 2020.
The buying rule
The category has dozens of products and the marketing all blurs together. Four numbers on the side panel will pick the right one for you in fifteen seconds.
If you're swapping cow milk for nutritional reasons — protein, calcium, B12 — fortified soy or pea is the like-for-like replacement and the only swap with regulatory backing. If you're swapping for taste or sustainability and you eat plenty of protein elsewhere, almond, oat, or coconut milk in your coffee is a defensible choice; just don't expect the carton to be doing anything nutritional beyond carrying a small dose of calcium and D.
When the rule changes
Tree-nut allergies rule out almond, cashew, and hazelnut milks. Soy allergy rules out soy milk. The widely repeated worry that soy milk feminizes men or harms women through its isoflavones does not survive the evidence: Messina's 2016 review of the clinical and epidemiological literature found no effect on male testosterone, no thyroid effect at normal intake, and no increase in breast cancer risk Messina 2016. The soy panic is one of nutrition's persistent myths and shouldn't keep anyone from the protein leader of the category.
Older adults already losing muscle should be aware they have removed a free 8 grams of protein per cup from their day if they swapped cow milk for almond or coconut. Either go back to dairy, switch to fortified soy or pea, or add the protein back somewhere else in the day Mariotti & Gardner 2019.
What people get wrong
"Plant milks are basically the same as cow milk now." Only soy and pea are. The other types deliver between zero and three grams of protein per cup against cow milk's eight. No amount of fortification adds protein — manufacturers don't sprinkle free amino acids into the carton. Drewnowski's audit of 219 US products found the majority failed a simple four-nutrient adequacy check against cow milk even after fortification Drewnowski 2021.
"Soy milk feminizes men." No. The clinical and epidemiological literature, surveyed comprehensively by Messina in 2016, finds no effect on testosterone, no thyroid disruption at dietary intake, and no breast cancer risk increase in women Messina 2016. Postmenopausal cohorts that eat the most soy have lower fracture risk, not higher Zhang et al. 2007.
"The additives are poisoning your gut." Mostly hype. The viral concern traces to a single 2015 mouse study showing that two specific emulsifiers — carboxymethylcellulose and polysorbate-80 — disrupted gut bacteria and promoted inflammation in mice fed high doses Chassaing et al. 2015. Those particular compounds aren't the ones most plant milks use. The actual additives — gellan gum, guar gum, locust bean gum, sunflower lecithin — haven't shown the same effects at human consumption levels. The honest answer is "we don't know enough to be sure," not "this is harming you."
"Vitamin D is vitamin D." Not quite. Plant milks are often fortified with D2 (from yeast or fungi), while cow milk uses D3. At equivalent doses on the label, D3 raises your blood vitamin D level roughly twice as efficiently as D2 Tripkovic et al. 2017. If you depend on your plant milk for vitamin D — especially in winter or at higher latitudes — look for a brand that fortifies with D3.
Where people go wrong in practice
The silent swap. Someone reads that almond milk is healthier, switches their cereal and coffee over, changes nothing else in their diet, and quietly drops 10 to 15 grams of daily protein and 300 to 500 milligrams of daily calcium without noticing. For most adults this is invisible. For an older woman eating at the low end of protein intake, this is the difference between staying above and slipping below the threshold associated with preserving lean muscle Mariotti & Gardner 2019.
The B12 gap. Vegans who switched to plant milk specifically because they read it would cover B12, and then bought a "minimal ingredients, just oats and water" brand that doesn't fortify. B12 deficiency builds slowly and the neurological symptoms — pins-and-needles, balance off, mind a bit slower — lag the actual deficiency by years. By the time the symptoms arrive, some of the damage may not reverse Allen 2009. Read the label or take a tablet.
The sweet "Original" carton. "Original" or "Vanilla" almond and oat milks routinely carry 5 to 15 grams of added sugar per cup. Three coffees a day for a year is roughly four to seven pounds of sugar consumed from milk alone. The "Unsweetened" version of the same brand has none of it and tastes nearly identical in coffee.
The unshaken carton. The calcium settles. Pour from the top of a carton that's been sitting in the fridge undisturbed and you may be drinking a third of the calcium the label promises Singhal et al. 2017. Shaking the carton once before each pour brings it back to label spec. This is free.
The price and the shelf
Plant milks retail at roughly two to four times the per-liter price of cow milk in most Western supermarkets. A daily-cup habit comes out to $50 to $150 a year more than cow milk. Refrigerated open shelf life is comparable to dairy at about a week; shelf-stable Tetra Pak versions keep unopened for around a year, which is the main practical reason to keep one in the pantry.
Soy and pea milk are usually the cheapest of the plant milks per gram of protein delivered — by a wide margin, because the others deliver so little protein. Almond milk is the most expensive per gram of any nutrient. If sustainability is the reason you're switching, most plant milks have a meaningfully lower greenhouse-gas footprint than dairy; almond milk's exception is its water use in California-grown supply, and oat milk is usually the cleanest on both counts.
Home-made almond or oat milk is straightforward — soak, blend, strain — but the result is unfortified, has a few days of shelf life, and doesn't carry the calcium or B12 a store-bought fortified version does. Useful for taste, not for nutritional swap.
What you don't notice happening
For most adults the swap from cow milk to almond is invisible. The protein and calcium your day was getting from milk is small enough relative to everything else you eat that nothing changes — you feel the same, your strength is the same, your bone scan in twenty years looks the same. This is the honest answer for the majority of the catalogue's readers.
For a smaller group it matters quietly, in ways that take years to show up.
The older woman whose protein intake was already at the low end of adequate — and who switched to almond milk in her coffee and cereal because her doctor mentioned cholesterol — loses muscle a little faster than she would have. The grip strength that drops earlier, the stairs that get harder a year sooner than they would have, the fall in her seventies that comes from a body that has less of itself to catch with: none of these show up in a way that points back at the milk. It is a slow subtraction Mariotti & Gardner 2019.
The toddler whose parents replaced cow milk with almond or oat for sustainability reasons grows up a little shorter than they would have — about half a centimeter per daily cup, tracked in dose-response in the Morency cohort Morency et al. 2017. The growth chart slips below the line it was on. The parents won't know which decisions slowed it; the milk is just one of many things.
The vegan whose B12 source was supposed to be the fortified oat milk that turned out not to be fortified develops tingling in their feet at forty-two. Their doctor takes a year to think of testing B12. The deficiency had been silent for the previous five years; some of what came back to normal will, and some of it won't Allen 2009.
None of these is dramatic. All of them are real and all of them are preventable by reading the side of the carton.
What changes when you read the label
Almost nothing dramatic. This is a small daily decision, not a transformation. What it buys you is the calm of knowing what you bought.
Within a week, you stop overpaying for almond milk thinking it's doing nutritional work, or you keep buying it knowing you like the taste and getting your protein elsewhere. Either way the small uneasy feeling at the dairy aisle — too many SKUs, conflicting headlines, vague guilt about whichever one you picked — goes away. You look at the protein number, the B12 line, the added-sugar line, and the call takes fifteen seconds.
Within a year, if you swapped to fortified soy because that's the actual nutritional swap, you're getting cow milk's protein and calcium plus the small isoflavone bonus on cardiovascular markers and, in postmenopausal women, on bone density Messina 2016, Zhang et al. 2007. If you stayed on cow milk because you tried fortified soy and didn't like it, you didn't lose anything by trying. If you switched to almond for taste and source your protein from yogurt, eggs, and beans, your nutritional bottom line hasn't changed and the coffee tastes how you want it to. None of these are wrong answers — they're the same person making the same call with their eyes open.
Over years, the parent who picked fortified soy for their toddler keeps the kid on the same growth curve they would have had on cow milk. The vegan who confirmed the B12 fortification stays neurologically fine through their thirties and forties. The older adult who replaced the protein lost from the milk swap keeps the muscle. None of these are visible payoffs — they are decades of nothing-going-wrong, which is the payoff this entry actually offers.
If lactose is the reason you're switching, the underlying question is about lactose intolerance — its own topic, and one where lactase-treated cow milk preserves the full dairy nutrient profile and may be the right answer before you ever reach for a plant alternative. The broader debate about dairy, saturated fat, and cardiovascular risk sits in the dairy entry. The full glycemic story of oat milk — what the enzymatic-hydrolysis step does to your blood sugar at meal-scale doses — deserves its own treatment. And if you're switching mainly for the climate footprint, the comparison gets technical fast and varies a lot by which plant milk and which region.
Substance and claimed effects
Plant-based milks are liquid beverages designed to substitute for cow milk in drinking, cereal, coffee, cooking, and baking applications. The category covers legume-based (soy, pea), grain-based (oat, rice), nut-based (almond, cashew, hazelnut), seed-based (hemp, flax), and tuber/other (coconut, potato) products. Most are manufactured by milling or soaking the source ingredient, extracting a water slurry, separating solids, then adding stabilizers, emulsifiers, oils, sweeteners, salt, and a fortification premix. Marketing positions them as nutritionally equivalent dairy substitutes; the actual nutrient profile varies by a factor of roughly eight on protein and is contingent on fortification choices each manufacturer makes voluntarily Chalupa-Krebzdak et al. 2018, Drewnowski 2021. The entry covers consequences for protein adequacy, calcium, vitamin D, vitamin B12, riboflavin and iodine intake when fortified or not, glycemic response, added-sugar load, additive exposure, and the comparative gap to cow milk and to soy milk specifically — soy and pea being the only widely sold plant milks that match dairy's protein-quality bar Mathai et al. 2017, Mariotti & Gardner 2019.
Evidence by addressing question
mechanism
Each plant milk inherits the macronutrient and micronutrient profile of its source ingredient, diluted ~8-12× by water during extraction. Almond milk is typically 2% almonds by weight; the remainder is water, oil for mouthfeel, and stabilizers. The single almond's protein, fiber, and most micronutrients are diluted with it Vanga & Raghavan 2018. Oat milk uses enzymatic hydrolysis (alpha-amylase) to break starch into shorter chains (maltose, maltodextrin) — this produces the perceived sweetness without "added sugar" on the label and pushes glycemic index into the 60s Atkinson et al. 2008. Soy and pea isolates, by contrast, are protein-concentrated streams of the source: the manufacturing target is roughly 3% protein in the finished beverage, matching cow milk's ~3.3%. Fortification is added post-extraction as calcium carbonate or tricalcium phosphate (the same forms used in cow milk in fortified jurisdictions), vitamin D2 or D3, cyanocobalamin (B12), riboflavin, and sometimes iodine. Calcium settles to the bottom of the carton between shakes — the fortificant is not chemically bound to the matrix the way cow milk's calcium is bound to casein micelles, so an unshaken pour delivers only a fraction of the labeled dose Singhal et al. 2017.
evidence
The category's central nutritional comparison is Vanga & Raghavan's 2018 review in Journal of Food Science and Technology, which placed soy milk closest to cow milk on protein quantity and quality (~3.3 g/100 mL, DIAAS ~0.91), with almond, coconut, rice, and oat falling far behind on protein (often <0.5 g/100 mL for almond and rice) Vanga & Raghavan 2018. Chalupa-Krebzdak et al. (2018) ran nutrient-density scoring on US-market products and concluded that without fortification only soy milk approached cow milk's nutritional adequacy index; almond, coconut, and rice scored substantially lower even when fortified Chalupa-Krebzdak et al. 2018. Drewnowski's 2021 analysis of the USDA Branded Food Products Database found that the majority of 219 plant milks did not meet the National Dairy Council's proposed nutrient standards for calcium, protein, vitamin D, and vitamin A simultaneously — fortification is partial and inconsistent at the brand level Drewnowski 2021. For calcium bioavailability, Heaney et al. (2000) and Zhao et al. (2005) showed that calcium carbonate-fortified soy milk, when shaken thoroughly, achieves absorption fractions statistically indistinguishable from cow milk (~24-30%) Heaney et al. 2000, Zhao et al. 2005. For protein synthesis, Tang et al. (2009) showed soy protein isolate stimulated mixed muscle protein synthesis at rest and post-exercise — less efficiently than whey but comparably to casein — while the muscle-anabolic literature generally finds soy and pea protein isolates around 70-80% of dairy's per-gram efficacy and easily compensated by ~20% larger doses Tang et al. 2009, Reidy & Rasmussen 2016.
The clinical-outcome evidence is thinner. Morency et al. (2017) found Canadian children aged 24-72 months consuming non-cow milk had a dose-response shorter stature: 0.4 cm shorter per daily cup of non-cow milk consumed, attributed primarily to lower protein and fat density rather than calcium Morency et al. 2017. This effect is the basis for the AAP / AHA / AND joint statement (2019) recommending against plant milks other than fortified soy as the primary milk for children under 5 Lott et al. 2019. For adults, the USDA 2020-2025 Dietary Guidelines recognize only fortified soy milk as a nutritionally equivalent substitute in the Dairy food group; other plant milks are explicitly not categorized as dairy equivalents USDA & HHS 2020.
protocol
The intake target is dictated by what cow milk would have delivered: in a typical Western diet that drinks ~250-500 mL of milk daily for coffee, cereal, and beverages, cow milk supplies ~8-16 g protein, ~300-600 mg calcium, ~120-240 IU vitamin D (where fortified), and a meaningful fraction of B12, riboflavin, and iodine. A direct swap to almond, coconut, oat, or rice milk preserves roughly the calcium and D if fortification is present and the carton is shaken, but loses most of the protein. Soy or pea milk preserves both. The decision rule is therefore: (1) read the protein figure on the nutrition panel — <3 g per cup means functionally non-protein milk; (2) read the calcium and B12 fortification — store-brand and "natural" oat milks frequently omit B12 and sometimes calcium; (3) confirm "unsweetened" on the label — sweetened versions carry 5-15 g added sugar per cup, comparable to soda; (4) shake the carton before each pour to deliver labeled calcium. For adults relying on plant milk as a primary milk substitute, fortified soy or pea is the like-for-like swap; other types are flavored fortified water and need protein replaced elsewhere in the day USDA & HHS 2020, Singhal et al. 2017.
contraindications
Rice milk carries elevated inorganic arsenic from the rice grain and the FDA explicitly advises against it as a primary milk for infants and young children FDA 2016. Plant milks other than fortified soy are nutritionally inadequate as a primary milk for children under 5 per joint guidance from the American Academy of Pediatrics, the Academy of Nutrition and Dietetics, the Academy of American Dentistry, and the American Heart Association Lott et al. 2019. Tree-nut allergies preclude almond, cashew, and hazelnut milks. Soy allergy precludes soy milk; soy isoflavones at typical dietary intake do not affect endogenous testosterone, thyroid function, or breast cancer risk in adult men or women per Messina's 2016 comprehensive review, so the popular "feminization" concern does not survive the evidence Messina 2016. Patients with phenylketonuria need to count pea milk's phenylalanine. Older adults at risk of sarcopenia who replace milk with a low-protein plant milk should be aware they have removed an easy ~8 g/cup protein source from their day.
misconceptions
The largest misconception is that plant milks are interchangeable nutrient substitutes for cow milk. They are interchangeable as a coffee whitener; they are not interchangeable as a nutrient delivery system except for fortified soy and pea. The 8 g protein per cup that cow milk provides is structurally not present in almond, coconut, oat, or rice milk and cannot be recovered by fortification (manufacturers do not add free amino acids). The Drewnowski 2021 audit of 219 USDA-database plant milks found that even with fortification, the majority failed a four-nutrient adequacy check matching cow milk's standard delivery Drewnowski 2021. The second misconception, going the other direction, is the panic that plant-milk additives (gellan gum, locust bean gum, dipotassium phosphate, sunflower lecithin, vegetable oils) are damaging. The mechanistic concern traces to Chassaing et al. (2015), which showed two specific emulsifiers (carboxymethylcellulose and polysorbate-80) altered mouse gut microbiota and promoted colitis at high dose — but the emulsifiers commonly used in plant milks (gellan, guar, lecithin) were not the tested compounds, the model is mouse not human, and the doses are not necessarily translatable Chassaing et al. 2015. The third misconception is that vitamin D2 in plant milks is equivalent to D3: D3 raises serum 25-hydroxyvitamin D roughly twice as efficiently as D2 at equivalent IU doses per Tripkovic et al. (2017) Tripkovic et al. 2017. Plant milks fortified with D3 exist (some from lichen-derived sources for vegan-acceptable D3); a reader specifically depending on plant milk for vitamin D should look for D3, not D2.
alternatives
Within the plant-milk category, soy milk is the protein-quality leader (DIAAS ~0.91, ~7-9 g protein per cup, isoflavone bonus on cardiovascular markers and bone density per Zhang et al. 2007's cohort showing soy-food consumption associated with lower fracture risk in postmenopausal women) Zhang et al. 2007, Mathai et al. 2017. Pea milk matches on protein quantity (~8 g/cup) with comparable digestibility (DIAAS ~0.7-0.8) and no isoflavones for readers avoiding them. Oat milk has marginal beta-glucan content (~0.4-1 g per cup, less than half the 3 g threshold for the FDA's qualified cholesterol-lowering health claim), so the beta-glucan benefit demonstrated in oat flour trials does not transfer at typical milk doses Whitehead et al. 2014, Brown et al. 1999. Outside the plant-milk category: lactose-free cow milk (lactase-treated) preserves the full dairy nutrient profile for the ~65% of adults with some lactose maldigestion; protein powder added to almond milk recovers the protein gap; fortified dairy alternatives (some Greek yogurts, fortified juices) cover calcium and D independently.
failure-modes
The dominant failure mode is the casual swap: a reader switches from cow milk to almond or oat milk for taste or sustainability reasons, doesn't change anything else in their diet, and silently drops 8-16 g protein and (when the carton isn't shaken or isn't fortified) 200-500 mg calcium per day. Over months, this affects protein adequacy in older adults and adolescents and bone-mineral accrual in growing children Morency et al. 2017, Singhal et al. 2017. The second failure mode is buying sweetened "Original" or "Vanilla" variants without checking the label, adding 5-15 g of added sugar per cup — a habitual coffee drinker putting 100 mL three times daily ingests roughly the sugar load of a can of soda from milk alone. The third is the B12 gap in store-brand or "minimal-ingredients" plant milks marketed as "just oats and water" — vegans who switched to plant milk specifically for the B12 fortification and bought an unfortified brand have removed their B12 source without knowing Allen 2009. The fourth is calcium settling: Singhal et al. (2017) note that calcium fortificants are typically insoluble particles dispersed in the liquid, and an unshaken pour from the top of the carton may deliver as little as 30% of the labeled calcium content Singhal et al. 2017.
practicalities
Plant milks retail at roughly 2-4× the per-liter price of cow milk in most Western markets. Refrigerated shelf life is comparable (~7-10 days after opening); shelf-stable Tetra Pak versions keep unopened for ~12 months. Barista-formulated variants (higher oil and stabilizer content for foam stability and resistance to acid-induced curdling in coffee) are now standard in cafes. Home production is possible (soak, blend, strain) for almond and oat but yields an unfortified, lower-shelf-life product. Sustainability comparisons consistently favor most plant milks over dairy on greenhouse gas emissions (1.4-3.2 kg CO2e/L for plant milks vs. ~3.2 kg/L for cow milk), land use, and water use — except almond milk in California-water-stressed contexts. This category isn't a health entry, but it drives many readers' switch decisions and so colors how they read the nutritional tradeoffs.
stakes
For adults: a quiet protein-and-micronutrient gap. A typical Western adult consuming ~400 mL of cow milk per day in coffee, cereal, and cooking removes 10-13 g protein, ~480 mg calcium, ~190 IU vitamin D, ~1.2 µg B12, and ~0.5 mg riboflavin from their day by swapping to almond or coconut milk without compensating elsewhere. The protein loss is the largest practical consequence: ~10 g/day is ~10-15% of total intake for many adults, enough to cross under the EAR for some older women already eating at the low end Mariotti & Gardner 2019. For children: stunted growth, in dose-response, demonstrated in Morency 2017's Canadian cohort (0.4 cm shorter per daily cup of non-cow milk in 24-72 month-olds), driven by lower protein and fat density Morency et al. 2017. For vegans: the B12 failure is the silent one — neurological symptoms (paresthesia, gait instability, cognitive slowing) lag deficiency by years and can be irreversible if missed Allen 2009.
payoff
The honest payoff scales by which milk is chosen. Fortified soy or pea milk delivers cow milk's protein, calcium, D, and B12 while removing dairy's saturated fat and lactose — and, in soy's case, adding isoflavones with cohort-level associations with lower fracture risk in postmenopausal women and modest LDL reduction Zhang et al. 2007, Messina 2016. Switching from sweetened "Original" oat or almond milk to unsweetened removes 5-15 g of added sugar per cup, which over a heavy-coffee-drinker year is a roughly 4-7 lb sugar elimination. Switching from cow milk to fortified soy specifically is one of the few "all-upside" swaps the catalogue documents — equivalent macronutrient and micronutrient delivery with a small bonus on cardiovascular markers and bone density via isoflavones. Choosing almond or coconut milk for taste while keeping protein in eggs, yogurt, or beans is a defensible choice that simply uses milk for the things milk is good at (mouthfeel, coffee, cereal) and gets protein elsewhere.
out-of-scope
Lactose intolerance is upstream of this decision for many readers and warrants its own entry. Dairy-vs-saturated-fat-and-cardiovascular-risk debates belong with the dairy entry. The detailed glycemic response to commercial oat milk versus oats deserves its own treatment — the enzymatic hydrolysis step is non-obvious and reader-relevant. Carrageenan as an additive (mostly removed from US plant milks since 2016 reformulations) is a separate additive-class question.
The credibility range
Optimist case
Plant milks, properly fortified, are nutritionally equivalent or superior to cow milk: soy and pea milk match protein quality and quantity, fortification covers calcium and vitamin D at FDA-acceptable bioavailability, soy isoflavones add cardiovascular and bone-density benefits not present in dairy Zhang et al. 2007, Messina 2016. Plant milks remove dairy's saturated fat (the persistent question mark for cardiovascular risk in the dairy literature) and lactose (a real issue for ~65% of the global adult population). Even the lower-protein varieties (almond, coconut, oat) provide a vehicle for calcium and D fortification at bioavailability indistinguishable from cow milk when properly formulated and shaken Heaney et al. 2000, Zhao et al. 2005. The sustainability case is strong across most plant milks. The category is a clear net win for adults who choose fortified soy or pea, or who choose another plant milk while sourcing protein elsewhere.
Skeptic case
The category's marketing implies nutritional equivalence to cow milk that is true only for fortified soy and pea — and the average shopper does not know this. The Drewnowski 2021 audit found the majority of US-market plant milks fail to match cow milk on a simple four-nutrient adequacy check even with fortification Drewnowski 2021. Children swapped to non-soy plant milk are demonstrably shorter in dose-response Morency et al. 2017. Calcium settling means labeled values overstate delivered values Singhal et al. 2017. Vitamin D2 fortification, common in plant milks, is roughly half as effective at raising 25(OH)D as D3 Tripkovic et al. 2017. Sweetened variants are a stealth sugar source. Store-brand "natural" oat milks frequently lack B12 fortification, creating a B12 gap in vegan readers who thought they had it covered. Almond milk is mostly water with stabilizers and is sold at dairy prices. The category is being asked to do nutritional work most of its products are not built to do.
Author's call
Both cases are partially right. The clean editorial position: fortified soy milk is a like-for-like dairy swap recognized by USDA dietary guidelines and the only plant milk with the protein quality and fortification profile to do that work alone USDA & HHS 2020. Pea milk is the same story without isoflavones for readers who prefer to avoid them. Almond, oat, coconut, rice, cashew, and hazelnut milks are flavored, fortified water, fine as a coffee whitener or cereal liquid as long as the reader sources protein elsewhere and reads the label for B12 and added sugar. Rice milk is a no for young children on arsenic grounds. The reader-actionable rule is: read the protein number on the panel before buying; <3 g/cup means it isn't doing milk's nutritional job. This lands the entry at evidence: 4 (nutritional comparisons are extensive and converge; clinical outcome data thinner but consistent) and controversy: 2 (genuine debate at the margins about additives and D2/D3, but the protein-quantity and calcium-fortification facts are not contested).
Stakeholder and incentive map
- Plant-milk manufacturers (Oatly, Califia, Silk, Ripple, Alpro): commercial incentive to market nutritional equivalence to dairy. Industry-funded comparison studies tend to emphasize parity on calcium and D and avoid foregrounding the protein gap.
- Dairy industry (National Dairy Council, lobbying bodies): commercial incentive to argue plant milks are nutritionally inferior and should not legally use the word "milk." Funded the labeling-rules petition the FDA addressed in its 2018 guidance. Sponsored research often emphasizes the protein and B12 gap.
- Vegan and animal-welfare advocacy: ethical incentive to encourage any plant-milk uptake, sometimes downplaying nutritional differences across plant-milk types.
- Pediatric and dietetics bodies (AAP, AND, AHA): incentive-neutral, conservative; their 2019 joint statement on plant milks for under-5s is the cleanest reference point for the children sub-question Lott et al. 2019.
- USDA / federal nutrition: explicitly designated fortified soy as the only plant-milk dairy-equivalent in the 2020-2025 Dietary Guidelines USDA & HHS 2020.
- Environmental advocacy: incentive to promote plant milks for sustainability, often without distinguishing the protein-adequacy spectrum across types.
Population variability
- Adults with adequate dietary protein from other sources (~1.2 g/kg): the protein gap from a plant-milk swap is functionally invisible. Almond, oat, or coconut milk is fine.
- Adults at the protein-intake margin (older women, elderly, those eating <1 g/kg): the ~10 g/day protein removed by swapping cow milk to almond crosses meaningful thresholds for sarcopenia prevention. Soy or pea or compensate.
- Children under 5: only fortified soy is recommended as an alternative to cow milk per joint AAP / AND / AHA statement Lott et al. 2019, Morency et al. 2017.
- Vegans: B12 source must be deliberate; fortified plant milk is one acceptable vehicle if the brand actually fortifies. Cyanocobalamin tolerance is universal in healthy adults.
- Postmenopausal women: soy milk's isoflavones associate with lower fracture risk in cohort data Zhang et al. 2007; this is a small bonus, not a replacement for resistance training and adequate calcium / D / protein.
- Lactose-intolerant adults: any plant milk works for the lactose-avoidance purpose; the nutritional choice is then the same as for lactose-tolerant adults.
- Patients with phenylketonuria: pea milk's phenylalanine load needs counting; soy milk needs counting; almond / oat / coconut / rice milks are low-phenylalanine alternatives.
Knowledge gaps
The clinical-outcome evidence in adults is thinner than the nutritional-comparison evidence. The Morency 2017 finding in children has not been replicated at scale in other populations, though the mechanism (protein and fat density) is well-grounded. There is no large RCT swapping plant milk for cow milk in older adults at sarcopenia risk and tracking lean mass; the evidence here is mechanistic and inferential. The long-term effect of the additive cocktail (gellan, guar, dipotassium phosphate, vegetable oils) on the human gut microbiome is not well-studied at habitual-consumption doses; the Chassaing 2015 mouse work is suggestive but does not establish human relevance for the emulsifiers actually used in plant milks Chassaing et al. 2015. The bioavailability of calcium from gellan-stabilized plant milks specifically (as opposed to the well-studied soy-with-tricalcium-phosphate formulation) is under-characterized. Whether oat milk's enzymatic-hydrolysis glycemic load matters at typical consumption (~30-50 mL in coffee) is plausible to dismiss but has not been rigorously tested.
The brief named protein adequacy, calcium / D / B12 fortification, glycemic response, added sugar, and additive profiles. The article covers all of these end to end except the deep glycemic-response story for oat milk specifically — the mechanism (alpha-amylase, maltose generation) is named in the mechanism section and the GI ballpark is given, but the meal-scale clinical-outcome literature is too thin to lean on and the typical reader uses oat milk in coffee-volume doses where the effect is plausibly negligible. Flagged as a separate-entry candidate below.
- Hard scoping call: framed as a label-reading entry, not a category survey. The temptation was to write a comprehensive comparison table across all seven plant-milk types. Instead the article foregrounds the one rule that does the work (the protein number on the panel) and trusts the type-by-type ranking section to carry the brand-shopping detail. The catalogue's voice is friend-with-a-rule, not registered-dietitian comparison chart.
- Energy and longevity scored 1, not 0. Both are honest small-positive calls. Energy because the B12 failure mode is real for vegans who bought an unfortified brand; longevity because fortified soy adds a modest isoflavone bonus on bone and cardio. Either could defensibly round to 0 in a stricter reading; I left them at 1 to keep the dimensions live in the article's coverage rather than silently dropping content.
- Soy "feminization" myth. Treated in misconceptions and again in contraindications. Decided to address it twice rather than once because it is one of the most common reasons readers avoid soy milk specifically — and avoiding soy collapses the only USDA-recognized like-for-like swap. Felt worth the redundancy.
- Additive question. Honest-uncertain rather than alarmist or dismissive. The Chassaing 2015 mouse study is the source of most public worry; it doesn't directly indict the emulsifiers actually used in plant milks. The article says "we don't know enough to be sure" rather than picking a side.
- Children-under-five guidance. Kept inside contraindications rather than spinning out into an audience block, because the parent buying for the toddler is one of the most common readers in this category and the warning callout earns a prominent place. The article's voice elsewhere is adult-to-adult; the contraindications section lets the pediatric rule sit cleanly.
- Action verb:
do, notdecide. Considereddecidebecause the act is a buying choice.dowon because the consumption is daily and ongoing, anddecidein this catalogue is reserved for clinician-input tradeoffs. - Separate-entry candidates: lactose intolerance (large, distinct topic, mentioned in out-of-scope); oat milk glycemic response at meal scale; dairy and cardiovascular risk (overlaps with a future dairy entry); environmental footprint of dairy vs plant milks. Once those entries exist, this entry should cross-link to them via
related. - Future-link candidates: a sarcopenia / protein-adequacy-in-older-adults entry; a B12 entry; a vitamin D entry; a label-reading skill entry.
- Citation density. The dossier carries 22 citations; the article uses 16, leaving Brown 1999, Reidy & Rasmussen 2016, and a handful of others as supporting context in the research layer. Within spec for the superset property.
Plant-Based Milks
One-off label-reading at the shelf (protein, fortification, unsweetened) and shaking the carton before each pour. No daily willpower cost once the brand is set.
Plant milks retail at roughly 2-4× the per-liter price of cow milk in Western markets; daily-cup consumption adds ~$50-150/year over cow milk, well inside the minor band.
Nutritional comparisons are extensive and converge (Vanga 2018, Chalupa-Krebzdak 2018, Drewnowski 2021). Clinical-outcome evidence is thinner: solid for children's growth (Morency et al. 2017) and calcium bioavailability (Heaney 2000; Zhao 2005); inferential for adult sarcopenia. USDA, AAP, AND, and AHA all align on fortified soy as the only dairy-equivalent plant milk.
A direct swap from cow milk to almond / coconut / oat / rice without compensating elsewhere removes ~8-16 g protein and ~200-500 mg calcium per day; for most adults at adequate baseline this is small, for those near the protein-intake margin it crosses thresholds (Mariotti & Gardner 2019). Fortified soy or pea milk is a wash on short-term wellness.
Soy milk's isoflavones associate with lower fracture risk in postmenopausal cohorts (Zhang et al. 2007) and modest LDL reductions per Messina's 2016 review; almond / coconut / oat / rice milk swaps are mortality-neutral. Small additive effect when soy is chosen, zero otherwise.
Energy effect is real only when the plant-milk swap creates a B12 or protein gap that wasn't filled elsewhere — a failure mode, not the default. Fortified soy or pea milk preserves B12 / protein delivery; unfortified store-brand oat milk does not (Allen 2009).