A potato is two different foods depending on whether it has been cooked, and the difference is not about texture. Raw potato carries glycoalkaloids that are genuinely toxic to dogs. Cooked potato carries almost none of them, and is a starch with a moderate calorie load and nothing much else.
That split is the whole article. Everything else β the chips, the mash, the skin, the green patches, the question of whether potato is a vegetable or a carbohydrate β follows from which of the two you are holding.
The reason this matters more than it does for most foods on this site is that potatoes are the one staple where the dangerous form is the form sitting in a sack in the cupboard, unprocessed, and where a dropped peeling is a realistic exposure in a way that a dropped pear is not. Dogs scavenge kitchen floors. Potato peelings are what ends up on kitchen floors.
Potatoes belong to the nightshade family, and like tomatoes and aubergines they produce glycoalkaloids as a defence against pests. The two that matter in potato are solanine and chaconine. Both disrupt nerve signalling, partly by inhibiting acetylcholinesterase, and both cause gastrointestinal irritation at lower doses.
The distribution inside the tuber is the part owners need, because it determines what to cut away. Roughly 30 to 80 per cent of a potato's total glycoalkaloid content sits in the skin and the thin layer immediately beneath it. The sprouts and the eyes concentrate them further. The flesh of a sound, well-stored tuber contains far less, which is why peeling is not a cosmetic step.
The absolute numbers are worth having, with the caveat that published figures vary by variety, storage and measurement method. A normal raw potato runs at roughly 2 to 15 mg of glycoalkaloid per 100 g of fresh weight. A green-skinned potato can reach 25 to 100 mg per 100 g. Sprouts and eyes can exceed 200 mg per 100 g. The regulatory threshold for human consumption in most jurisdictions is around 20 mg per 100 g, which tells you that a green potato is not sold as food for people either.
The toxicity figures for dogs are less settled, and the sources disagree by orders of magnitude. The estimate that appears most consistently in veterinary and toxicology writing is roughly 2 to 5 mg of solanine per kg of body weight for gastrointestinal signs, and above that for neurological signs. A 5 kg dog would therefore be in the range of 10 to 25 mg of solanine, which is a small fraction of a green potato. That is the arithmetic behind the warning, and it is why a green potato is not something to leave on the floor.
Two visible signs tell you a potato has gone wrong, and neither is negotiable.
Greening is chlorophyll, and chlorophyll itself is harmless. It is a marker rather than a poison. What it signals is that the tuber has been exposed to light, and the same light exposure that produces chlorophyll also drives glycoalkaloid synthesis. A potato with a small green patch near the surface can be salvaged by cutting away the green and a margin around it, provided the flesh underneath is sound. A potato that is green across a large area, or green through the flesh, is not salvageable and should be discarded.
Sprouting is the stronger signal. A sprout is the tuber trying to become a plant, and the growing points are where glycoalkaloids concentrate. A potato with small, freshly formed eyes can be used if the eyes are cut out deeply and the rest of the tuber is firm and normal-coloured. A potato with substantial sprouts, wrinkled skin or soft spots is compost.
The third sign is less visible: a bitter taste. Glycoalkaloids are bitter, and bitterness in a cooked potato is a reason to stop eating it, for people as much as dogs. Bitterness does not survive cooking, which is one of the reasons cooking is not a reliable way to make a bad potato safe.
The rule that covers all of this: green, sprouted, soft or bitter potatoes go in the bin, not in the pot and not on the floor. Peelings from any potato go in a covered bin rather than an open bowl, because a kitchen floor with peelings on it is a dog exposure.
Cooking reduces glycoalkaloid content substantially but does not eliminate it. The published reduction for boiling or baking is in the range of 30 to 70 per cent, depending on method, duration and the potato itself. Frying is less effective at reducing the compounds and adds a fat load of its own.
That partial reduction is the reason the practical advice is not "cook it and it is fine". It is "start with a sound potato, peel it, cook it thoroughly, and serve a small amount". Cooking rescues a normal potato from being mildly risky; it does not rescue a green one.
The method matters less than the starting material. Boiling leaches some glycoalkaloids into the water, which is then discarded β a small advantage. Baking concentrates everything as water evaporates, so a baked potato has a higher concentration per gram than a boiled one, though the total is similar. Neither method destroys the compounds at ordinary cooking temperatures; the destruction temperature commonly cited is well above what a kitchen oven reaches.
What cooking does reliably is make the starch digestible. Raw potato starch is resistant to digestion, which is why raw potato causes gas and loose stool even in the absence of glycoalkaloids. Cooked starch is gelatinised and accessible. This is a real nutritional change, and it is the reason cooked potato is a food while raw potato is not.
The nutrient profile of a plain potato is widely misunderstood, and the misunderstanding is worth correcting because it changes how the food should be treated.
Per 100 g of raw potato with skin, the USDA figures are: 77 kcal, 17.5 g carbohydrate, 0.82 g sugar, 2.1 g fibre, 2.05 g protein, 0.09 g fat, 425 mg potassium and 6 mg sodium.
The sugar figure is the surprise. Less than a gram per 100 g β lower than almost any fruit discussed in this series, lower than a carrot. A potato is not a sweet food. What it has is starch, and starch is a chain of glucose that the digestive system breaks down at a rate that depends heavily on how the potato was cooked.
That distinction explains why a potato can be low in sugar and still produce a sharp rise in blood glucose. Starch digests to glucose; the sugar number on the label tells you nothing about the glycaemic effect. For a dog with diabetes, the relevant question is not "how much sugar" but "how fast does this become glucose", and for potato the answer depends on preparation in a way it does not for most foods.
The potassium figure, 425 mg per 100 g, is high β similar to a banana, and much higher than a pear or an apple. For a dog on a renal diet that is a meaningful number rather than a footnote.
The fibre, 2.1 g, is modest. Potato is not a fibre source. The protein is negligible at the quantities a dog would eat.
So the nutritional summary is: a starch with moderate energy, high potassium, low fibre, and almost no sugar. It is not a vegetable in the nutritional sense, whatever the supermarket signage says.
There is a detail about potatoes that most dog-feeding articles miss entirely, and it is relevant to any owner managing a dog's blood sugar.
When a potato is cooked and then cooled, some of the gelatinised starch re-crystallises into what is called resistant starch β a form that resists digestion in the small intestine and passes to the colon, where it ferments. The practical effect is that a cold cooked potato has a lower glycaemic impact than the same potato eaten hot, and the effect survives reheating to some degree.
This is genuine food science rather than a wellness claim, and it has been studied in humans for decades. For a dog, the relevance is modest but real: a small piece of cold cooked potato produces a slower glucose rise than the same piece straight from the oven, and the resistant starch that forms is a substrate for gut bacteria rather than a sugar load.
That is not a licence to feed potato to a diabetic dog. It is a reason to prefer cold or reheated potato over hot where potato is being used at all, and a reason not to be alarmed by the glycaemic index tables that rank potato near the top β those tables are built on hot, freshly cooked potato.
The same mechanism sits behind the observation that cold potato can cause gas. Resistant starch fermenting in the colon produces gas, in the same way that beans do. A dog that gets wind from cold potato is experiencing the other side of the same effect that makes it gentler on blood sugar.
The plain cooked potato is a moderate food. Almost every potato product sold is a different and much worse one, and the mechanism is consistent β fat, salt, or both.
Chips β the thick-cut, fried kind β are potato that has absorbed oil. Fat content runs from around 10 to 20 per cent depending on the fry, which takes a 77 kcal food to somewhere between 250 and 350 kcal per 100 g. Add salt and the sodium load becomes significant. A dog that eats chips is eating fat and salt with a potato garnish, and for a dog whose pancreas has already struggled once, that is a real risk rather than a theoretical one.
Crisps are the same process taken further. Fat around 35 per cent, sodium in the hundreds of milligrams per 100 g, and a texture that shatters into sharp fragments. The sharpness is not a joke: a crisp fragment can irritate the mouth and oesophagus, and a dog that gulps a handful is at risk of a piece lodging.
Mash is potato plus whatever went into it. Butter, milk, cream, salt β each is a separate issue, and the combination is a fat and dairy load that a dog with a sensitive gut will react to. Plain mash with nothing added is fine in small amounts; restaurant mash is not the same food.
Roast potatoes are the highest-fat form in a normal kitchen, because they are parboiled and then cooked in a layer of oil or animal fat. They are also the most likely to be on a plate at floor height during a family meal.
The pattern across all of these is that the potato is the least problematic ingredient. What changes the food is what was done to it, which is the same conclusion that applies to fruit β the fruit is rarely the problem, the product is.
Sweet potato is routinely presented as the safer choice, and the comparison is worth doing properly because the reasons are not the ones usually given.
Per 100 g, sweet potato is about 86 kcal, 20.1 g carbohydrate, 3 g fibre, and it carries a very high beta-carotene content. White potato is 77 kcal, 17.5 g carbohydrate, 2.1 g fibre, and essentially no beta-carotene. So sweet potato is slightly more calorie-dense and slightly higher in carbohydrate, with more fibre and a genuine micronutrient advantage.
The safety difference is not about nutrition at all. Sweet potato belongs to the morning glory family rather than the nightshade family, and it does not produce solanine or chaconine. There is no green-skin toxicity, no sprout toxicity, and no equivalent of the raw-potato rule. That is the real reason sweet potato is the easier recommendation: it removes an entire category of risk rather than improving the nutrient profile.
The caveat is that sweet potato is still a starch, still high in carbohydrate, and still something to portion. It is also, in large amounts, a cause of gas and soft stool. The advantage is safety, not licence.
For a dog that likes potato and tolerates it, plain boiled white potato is fine. For an owner who wants the simpler rule, sweet potato avoids the glycoalkaloid question entirely β provided it is cooked, since raw sweet potato is hard to digest and a choking risk in chunks.
Any article about potato and dogs has to address the thing that made potato controversial, because owners bring it up and the coverage has been confusing.
In 2018 the US Food and Drug Administration began investigating a possible link between certain diets and dilated cardiomyopathy, a heart condition in which the heart muscle becomes thin and weak. The diets in question were grain-free formulations that used potatoes, legumes or both as substitutes for grains β peas, lentils, chickpeas, and potato in various forms. The agency published a series of updates between 2018 and 2022, and the final position was that the data were insufficient to establish a causal relationship.
That conclusion is worth stating precisely, because both the alarmist and the dismissive readings are wrong. The FDA did not find that potato causes heart disease. It also did not clear the diets. What it found was a signal that could not be resolved with the data available, in a population where DCM has other well-established causes β genetics in predisposed breeds, taurine deficiency, and other nutritional factors.
Several mechanisms have been proposed. One is that legume-heavy diets may interfere with taurine availability, since taurine is an amino acid the heart depends on and dogs can synthesise it from precursors. Another is that the total protein and amino acid profile of some grain-free formulations was different from what the animals needed. A third is that the association is confounded β dogs fed grain-free diets may differ from the general population in ways that matter.
For an owner reading this, three practical points follow. First, potato itself is not implicated; the diets that attracted attention were multi-ingredient formulations and the common factor was the absence of grain rather than the presence of potato. Second, a complete diet from a manufacturer with a feeding trial behind it is a different proposition from a diet formulated on paper, and the difference is not visible on the ingredient list. Third, a dog on a grain-free diet that develops exercise intolerance, coughing or fainting needs a cardiac workup rather than a change of food, because DCM is treatable when caught early and the diet question can be settled afterwards.
The one place where the potato discussion intersects with safety directly is home-cooked diets. A home-made diet that uses potato as a staple needs the same formulation rigour as any other, and potato is not a substitute for a protein source or a calcium source. It is a starch, and in a balanced recipe that is exactly the role it should play.
Diabetes. Potato is the food in this series where "no added sugar" is least reassuring, because the starch load matters more than the sugar. A boiled potato eaten hot produces a fast glucose rise; the same potato eaten cold produces a slower one. For a dog on insulin, potato is not forbidden but it needs to be counted, and the timing relative to the dose matters. Small amounts, cold or reheated rather than hot, and let whoever manages the diabetes set the number.
Kidney disease. At 425 mg of potassium per 100 g, potato is one of the higher-potassium staples. A dog on a renal diet is often being managed partly by limiting potassium, which makes potato a food to weigh rather than estimate. The good news is that boiling leaches potassium into the water, so a boiled potato is lower in potassium than a baked one β for a renal dog, that is the preparation to use if potato is being used at all.
Pancreatitis. Plain boiled potato is low in fat at 0.09 g per 100 g, so the tuber itself is not a trigger. The risk is entirely in the preparation: chips, roast potatoes, mash with butter, any of the fried forms. For a dog with a history, plain boiled potato is acceptable and everything else in the potato family is not.
Overweight dogs. Potato is a starch with a moderate energy density and no satiating fibre to speak of. As a treat it is less useful than a carrot or a green bean, which deliver bulk for fewer calories. It is not harmful; it is simply not the best use of a small treat budget.
Puppies. No specific toxicity beyond the general raw-potato rule, but a puppy's gut handles starch and fat less well than an adult's. Plain boiled, small amount, and no skins.
Start with the calorie arithmetic. A 6 kg dog on 300 kcal a day has 30 kcal for treats, which is about 39 g of plain potato. A 12 kg dog on 580 kcal has 58, which is 75 g. A 25 kg dog on 1,100 kcal has 110, which is 143 g.
Those are ceilings. The practical portions that most vets would suggest are smaller, and expressed in spoons rather than grams: one to two teaspoons for a small dog, one to two tablespoons for a medium one, two to four tablespoons for a large one, once or twice a week at most.
The preparation rules are short and each addresses a specific risk.
Peel it. The skin holds most of the glycoalkaloids, and peeling removes the majority of the risk along with it. This is the one step that is not optional.
Cut away anything green, and cut a margin beyond the green. If the greening goes deep, discard the potato.
Cut out the eyes and any sprout, deeply. A potato that has grown substantial sprouts is compost, not dinner.
Boil or bake with nothing added. No salt, no butter, no oil, no stock. A dog does not need seasoning and the seasoning is the part that causes problems.
Cool it before serving, or serve it cold. This reduces the glycaemic impact and removes the burn risk from a hot piece.
Cube it to a size appropriate to the dog. Cooked potato is soft and moulds to the throat, which makes it more of a choking risk than its size suggests β a piece that squashes can seal an airway in a way a hard piece would not.
One point about leftovers. A boiled potato that has been in the fridge for three days is fine for a dog if it has been stored properly and shows no mould or off smell. A potato salad that has been sitting out at a barbecue is not, and the mayonnaise is the reason.
Most potato exposures are not a dog raiding a sack. They are a peeling that fell, a bag left open at floor height, or a compost bin that was never shut.
Three changes to kitchen routine cover almost all of it, and none of them is onerous. Peel into a container with a lid rather than onto newspaper on the floor, and empty it straight into the bin rather than leaving it out. Keep the raw potato sack in a cupboard rather than a low basket, and keep the compost caddy closed if a dog can reach it. None of that requires vigilance, only placement.
The other exposure route is guests. A visitor peeling potatoes for dinner has no reason to know that the peels are the risky part, and a dog sitting hopefully at their feet is hard to refuse. The useful thing to say is not "don't feed the dog", which invites negotiation, but "nothing raw from the potato". It is a short sentence that covers the actual risk.
Children deserve a mention because the pattern is different. A child who has been told potatoes are food will share them, and a child is also the person most likely to drop a piece and least likely to notice a dog taking it. The rule for a household with both is that the potato preparation happens on a counter, not on the floor, and that the bin is out of reach.
None of this is about living in fear of a vegetable. It is about the fact that the dangerous form of this particular food is an ordinary kitchen waste product, and that the fix is a change of habit rather than a change of diet.
This is the scenario the article exists for, and the response depends on what was eaten and how much.
The first task is identification. A piece of raw peeled flesh is the lowest-risk version. A raw peeling is higher, because the skin holds most of the glycoalkaloids. A green potato, a sprouted potato, or any part of the plant above ground β leaves, stems β is the highest-risk version and warrants a call regardless of quantity.
The second task is quantity, relative to the dog's weight. A 5 kg dog that ate a quarter of a green potato is in a different category from a 30 kg dog that ate a single peeling. The toxic threshold is usually expressed per kilogram, which means the same piece of potato is a different dose in different animals.
The third is timing. Signs typically appear within a few hours and up to a day. The window to watch most closely is the first twelve hours.
What to do: never make the dog vomit on your own initiative β that is a decision for the vet. Call a vet or a poisons service with the dog's weight, what was eaten, roughly how much, and when. If the dog is already vomiting, that is useful information and also a reason to call rather than wait.
What to watch for: vomiting, diarrhoea, abdominal pain, drooling, lethargy. The neurological signs β tremors, twitching, weakness, dilated pupils, a slow heart rate β indicate a more serious exposure and are an emergency rather than an observation.
What usually happens with a small exposure in a healthy dog is nothing, or a day of soft stool. What happens with a large exposure to a green potato can be serious, which is why the assessment should be made on the specifics rather than on a general sense that potatoes are food.
Best not. The skin carries most of the glycoalkaloids, and cooking reduces them without removing them. A baked potato skin is a concentrated version of the part you were trying to avoid. Peel the potato and discard the skin, especially for a small dog.
No. A single green patch on an otherwise sound potato is a small exposure, and cutting it away with a margin makes the rest usable. The rule that matters is about potatoes that are green through, or sprouted, or soft β those go in the bin rather than being trimmed.
They are the worst form on the list and the answer is essentially no. Fat, salt and a fried texture, in a food that was already a starch. A single chip is unlikely to harm a healthy dog; a habit of chips is a fat load with a pancreatitis risk attached.
Plain mash with nothing added, in a small amount, yes. Mash with butter, milk, salt or gravy is a different food, and gravy in particular often contains onion, which is toxic to dogs. Check what went into the mash before deciding.
It reduces the glycoalkaloids by roughly 30 to 70 per cent and makes the starch digestible, which is why cooked potato is a food and raw potato is not. It does not eliminate the compounds, which is why the starting potato has to be sound rather than green.
Nutritionally it behaves like a grain β starch, moderate energy, little fibre. Dogs do not need vegetables at all, and potato is not a substitute for a balanced diet. Treat it as a starchy extra rather than as a serving of vegetables.
PawAtlas treats the potato as the clearest case in this series where the same ingredient is safe and unsafe depending on handling. The dog does not need to understand solanine; the owner needs to know that a green potato goes in the bin, that a peeling on the floor is an exposure, and that everything fried is a different food.