What the Carnivore Misses, and What It Minds

What the Carnivore Misses, and What It Minds. An alternative protein may fail not because it lacks protein, but because the food lacks something the animal seeks, contains something it avoids, or both. Two different problems, and reaching for a single palatant answers neither on its own. On why the alternative-protein problem was never a protein problem.

Movement I

THE TWO PROBLEMS

Why would an animal refuse a food that is, on paper, everything it needs? Set down a bowl built on a new protein, a pea, a fungus, a single cell grown in a tank, complete by every nutritional measure and correct in every amino acid, and watch a cat walk away from it. Nothing a label can read is wrong. And yet the animal may be registering two complaints at once, and they are not the same complaint. Something it was listening for is not there. Or something it would rather not meet is. Or both.

Those are the two palatability problems of an alternative protein, and the whole of this Conversation turns on keeping them apart. The first is a problem of subtraction. Meat tends to speak to a carnivore in a particular chemical language, the free amino acids, the nucleotides, the fat and what heat makes of it, and a protein that did not come from an animal often arrives quieter in that language. The food is not offensive. It is simply saying less where the animal expected to be spoken to. The second is a problem of addition. The new ingredient can bring notes of its own that meat did not carry, a bitterness, a green or earthy smell, a texture the mouth was not expecting, and some of these a carnivore may be inclined to refuse. Here the food is not quiet at all. It may be saying the wrong thing.

A food can fail by lacking what the animal seeks, by carrying what it avoids, or both.

The trade, for the most part, treats these as one problem with one solution. Make it more palatable. Add the palatant, lift the score, move on. But a single dial set to one number cannot answer two different complaints, and, as a later movement will show, the motion that addresses one can disturb the other. Subtraction and addition are not two names for the same deficit. They are different problems that happen to arrive in the same bowl.

The cat and the dog do not necessarily hear the same message, and we should resist a tidy law that ranks one as forever the harder judge. Their sensory systems differ, their feeding histories differ, and so does the individual animal at the bowl, shaped by its own experience. It is plausible, and worth testing rather than assuming, that a carnivore with a long evolutionary commitment to meat misses the meat signal more and forgives the off-note less, and that a dog, nearer an omnivore in its history, meets more of it halfway. But that is a hypothesis to put to the bowl, not a settled ranking of the species, and a formulation that succeeds with one animal can still fail the next. The useful rule is narrower: the two problems are real for both species, and which one bites, and how hard, is a question to be asked of each food and each animal, not assumed in advance.

So the question for the rest of this Conversation is not the one the industry usually asks, which of these proteins tastes best. It is a pair of questions, held apart on purpose. What did the animal stop hearing when the meat left? And what did it start hearing that it would rather not? Answer only the first and you may have a worthy, quiet food the animal passes over. Answer only the second and you may have a clean, inoffensive food the animal still has no reason to want. The alternative-protein problem can be either, or both at once, and telling which is where the work begins.

---

Movement II

THE MISSING SIGNAL

What, exactly, is the animal listening for? Not protein, as the formulation sheet means it. An animal does not perceive protein, the nutritional category we write down; it perceives what that protein has broken into and what travels with it. Meat, by the time it reaches a mouth, is often rich in free amino acids, the loose savoury fragments that autolysis and ageing have cut from the chains. It can carry nucleotides, the small molecules released as the tissue's own machinery winds down. It carries animal fat and the volatiles that come off it, and the raw material that heat turns into the smell of cooking. Meat is not one flavour. It is a chord, several notes sounded together, and a carnivore has spent its evolution learning some of those notes and moving toward them.

An alternative protein often plays fewer of those notes, though not always, and not all of them. Much of a pea or a legume arrives as intact protein, with less of it cut loose into free amino acids. But the picture is not uniform. A yeast or other microbial biomass can be genuinely rich in nucleotides and free amino acids and other flavour-active compounds, and fermentation and downstream processing can raise or lower them further. Meat, for its part, is not automatically generous with every precursor, and no two animal sources carry quite the same signature. So the honest claim is not that the alternative protein is silent and meat is loud. It is that they are differently composed, and the composition an alternative source arrives with is frequently thinner in exactly the notes a carnivore is most tuned to. The food is often not offensive. It is, often, saying less, and the animal is not recoiling so much as failing to be called.

Meat is a chord, and an alternative protein often plays fewer of its notes.

And the cat reads that chord through an umami system worth describing precisely, because the precise version corrects a common one. In the cat, the receptor that reports savour responds most strongly not to amino acids but to certain nucleotides, the purine nucleotides acting as the agonists that switch it on. Some amino acids then enhance that response, but in the cat they enhance rather than initiate it, and the familiar human favourites, glutamate and aspartate, are not the ones doing the work. It is a particular molecular partnership, nucleotide first and select amino acids second, not a generic appetite for protein. Replace the meat without preserving the balance of nucleotide and amino acid the partnership needs, and you may weaken a specific collaboration the feline umami system is equipped to detect, wherever that balance is or is not supplied.

There is a tempting way to answer all this, and it is a trap the next movements will spring. If the problem is too little free amino acid, break the protein down and free some. If the problem is too little reaction flavour, drive the cooking, or supply the precursors and let the heat work. Each of those moves can genuinely restore part of the missing signal. Each of them, we will see, can create the second problem while solving the first, though whether it does depends on the substrate and the method. For now it is enough to name the quiet precisely, because a deficit named wrongly is answered wrongly.

That is the shape of the subtraction problem, and on its own it is often the more tractable of the two. A worthy, honest, quieter food, passed over not because it offends but because it does not say enough. If that were the only difficulty, the alternative-protein project would be an exercise in putting the missing notes back. It is harder than that, because the same ingredient that went quiet in one register may be speaking up in another.

---

Movement III

THE INTRUDING NOTE

What is the alternative protein saying that the animal would rather not hear? Begin with the irony Movement II walked us up to. One way to answer the subtraction problem, to free the savoury amino acids a carnivore reads as meat, is to break the protein down. And breaking the protein down, on a plant or a single-cell substrate, can free more than the savoury fragments. It can expose bitter ones. The short, hydrophobic peptides that hydrolysis lifts out of some proteins taste bitter, at least to a human panel, and the same cut that yields savour can yield them alongside it. The dial we reached for to fill the quiet may, in the same turn, have begun to make a noise.

Bitterness, where an animal does perceive it, is not a flavour to be coaxed away like a preference. It reads closer to a warning, an evolved caution against the alkaloids and toxins that often announce themselves as bitter. That caution is not absolute; concentration, context and learning can all move it. But it is not a mere preference to be coaxed away either. But two honest qualifications belong here. The first is that bitterness does not rise in lockstep with the degree of hydrolysis; where one compound rises another can fall, and some processing continues to cut the protein while lowering the bitterness, which was the whole burden of an earlier Conversation, No. 7, on where to stop the cut. The relationship was charted years ago, largely in dairy proteins, some of that work my own, and it remains only partly charted in the plant, fungal and single-cell substrates the industry is now reaching for. The second qualification is sharper: a peptide that a human panel calls bitter is not thereby proven aversive to a dog or a cat at the concentration it reaches in the finished food. The chemistry of the off-note and the animal's response to it are two separate questions, and this series has no business collapsing them.

A compound a human panel calls bitter is not yet proven aversive to the animal in the bowl.

Bitterness is not the only intruding note, but the others ask for the same care. A legume can carry enzymes that, meeting the plant's own fats, generate green, grassy, beany smells, most notoriously in soybean, though the intensity depends heavily on the raw material and how it was handled. Meat does not smell of cut grass, and an animal may notice the difference even where we have stopped noticing it. Some novel ingredients also arrive wrapped in chitin, the tough material of an insect cuticle or a fungal wall, which can in principle change how a piece breaks and feels in the mouth. But how much chitin, presented in what form, and whether a given cat or dog actually registers it in a finished kibble, are open questions rather than settled facts, and the honest move is to raise them, not assert them.

And there is a distinction worth drawing, provided we draw it as a question and not a verdict. When an animal refuses a new food, is it refusing because the food is genuinely aversive to it, or only because the food is unfamiliar? The two can look identical at the first meeting and diverge by the tenth, because an unfamiliar but harmless food is often accepted over repeated, unpressured exposures, while a genuinely aversive one may not be. Dogs have been seen to hesitate and eat more slowly when first met with a novel diet; cats' preferences are known to be shaped by experience. But the line is not clean. An aversion can soften with exposure, and an unfamiliarity can prove stubborn, so the question is not which fixed category a refusal belongs to but which way it is moving, and whether anyone watched it long enough to know. An instrument that scores a single first encounter, and much of the industry's testing does exactly that, cannot see which way the refusal was heading.

Getting this wrong is expensive in both directions. Read an unfamiliarity as an aversion and you abandon a food the animal would have grown into. Read an aversion as mere unfamiliarity and you keep presenting a food that was never going to win, waiting out a refusal that was not fading. So the addition problem is really a cluster of them, chemical and perceptual and temporal at once, and only some of it is yours to engineer. Which raises the question the next movement must face, carefully: when is the cure for one problem bound to the cause of the other, and when is it not?

---

Movement IV

WHEN ONE MOVE CHANGES THE OTHER

Why not simply do both, then, fill the quiet and remove the intrusion, and meet the animal in the middle? Sometimes you can. That is the correction this movement exists to make, because the tempting story, that every gain in signal buys an equal loss to noise, is too tidy to be true. The real situation is both better and more demanding. The two problems are sometimes coupled, so that the same change moves both, and sometimes they are not, so that a well-chosen change can move one and leave the other largely alone. The difficulty is not that improvement is impossible. It is that we cannot assume a change improves only the thing we aimed it at.

When are they coupled? When the signal and the off-note are made by the same event. The hydrolysis that frees a savoury fragment and the one that exposes a bitter peptide can be, at that moment, the same cut, and asking it for more of the first will sometimes bring more of the second. The reaction that builds a cooked, meaty aroma and the one that builds a scorched or bitter note are the same chemistry run to different extents. Where the same step generates both the wanted and the unwanted, asking for more of one may move the other, and whether that trade-off can be dodged depends on the reaction pathway and the control you have over it. Assume it never can, and you resign yourself to worse food than the chemistry required.

There is more than one dial, and they do not all move signal and noise together.

But coupling is not the law of the whole field, only of certain moves within it. Change the substrate and the balance of savoury to bitter fragments changes with it. Stop the cut at a different point, as No. 7 argued, and the same protein yields a different mixture. The point is not to list the levers here, and I will not, but to correct the fatalism: there is more than one dial, the dials do not all move signal and noise together, and part of the craft is knowing which are coupled on a given substrate and which can be worked apart. That knowledge is specific, hard-won and, frankly, the asset, so this essay will say plainly that the decoupling exists without saying how it is done.

Masking deserves its own caution, because it is the move the trade reaches for most and the one least likely to decouple anything. To lay a louder flavour over an off-note is not to remove the off-note, nor to restore the quiet true signal underneath; it is to add a third thing on top of both. Sometimes it works well enough to sell. But a precise feeder, and some animals are more precise than others, may read the whole stack as wrong rather than be won by it, and the underlying problem, signal too faint or noise too present, is still there beneath the cover. Masking can buy acceptance. It rarely buys a solution.

So the honest shape of the hard problem is not one cruel dial but a question that has to be answered before any dial is touched: for this food, on this substrate, are the two problems coupled or not, and which of them is actually present? That question has no single answer, and it is narrower and more useful than the industry's reflex, which is to turn up the palatant and hope. It returns us, for the last movement, to the only instrument that finally rules, the animal, and to what its verdict can and cannot tell us about which problem we were ever solving.

---

Movement V

WHAT THE ANIMAL IS ACTUALLY JUDGING

What is the animal actually judging, when it lowers its head to the bowl or turns away from it? Not the protein, in the sense the formulation sheet means. It does not judge a nutritional category; it judges a food, and it renders on each food a single verdict that folds everything together, the signal and the noise, the savoury and the bitter, the familiar and the strange, into one decision with no partial credit. It does not eat the half we got right and leave the half we got wrong. It eats, or it does not.

That is the quiet reproach buried in this whole subject. We filed the alternative protein under protein, in the protein plans and the protein targets and the protein press, and all the while the animal was not judging the protein. It was judging the taste, the smell and the feel of a food, which is to say it was judging palatability, and palatability does not sit in the protein department. The problem we named for its most expensive ingredient was, in good part, a problem of the senses. We put it in the wrong room, and then wondered why the tools in that room did not fix it.

An animal never judged a protein. It only ever judged a food.

Seen from the bowl, the use of the two-problem distinction is finally plain, and it is diagnostic, not merely descriptive. A low palatability score tells us only that the animal responded differently. It does not tell us why. Several quite different faults produce the same low number: too little of the signal the animal seeks, too much of the note it avoids, a food merely unfamiliar, or some interaction among these. A two-bowl preference ratio cannot tell them apart, and No. 9 asked what lies beyond that bowl for exactly this reason. Yet the remedy for each is different, and some remedies make the others worse. So the first discipline this Conversation asks for is cheap and almost never practised: before you reach for more palatant, find out which problem you actually have.

A low palatability score tells us the animal responded differently. It does not tell us why.

Each of these faults deserves its own return, and some we have visited before, the fat of No. 4, the meat chemistry of No. 2 and No. 3, the cut of No. 7. The Fridays ahead go back to them not as unopened subjects but as open diagnostic questions: not only what the missing signal is, but how to know when it is what is missing; not only whether a cut can turn savour bitter, but when it must and when it need not. The unfamiliar and the aversive most of all deserve to be told apart properly, because an industry that cannot distinguish them is measuring the wrong thing. And a food grown in a tank is already posing the whole question in miniature, signal and noise and novelty at once.

For now, the first and largest thing is only to hold the problem the right way round. We kept calling it a protein problem and reaching for protein's tools. The animal kept answering the question it was actually asked, which was never about the protein as a category at all. It was about the food. It is always, only, about the food.

References

1.  Bradshaw, J.W.S. (2006). The evolutionary basis for the feeding behavior of domestic dogs (Canis familiaris) and cats (Felis catus). Journal of Nutrition 136(7):1927S-1931S. doi:10.1093/jn/136.7.1927S

2.  McGrane, S.J., Gibbs, M., Hernangomez de Alvaro, C., Dunlop, N., Winnig, M., Klebansky, B. & Waller, D. (2023). Umami taste perception and preferences of the domestic cat (Felis catus), an obligate carnivore. Chemical Senses 48:bjad026. doi:10.1093/chemse/bjad026

3.  Zoran, D.L. (2002). The carnivore connection to nutrition in cats. Journal of the American Veterinary Medical Association 221(11):1559-1567. doi:10.2460/javma.2002.221.1559

4.  Saha, B.C. & Hayashi, K. (2001). Debittering of protein hydrolyzates. Biotechnology Advances 19(5):355-370. doi:10.1016/S0734-9750(01)00070-2

5.  Rackis, J.J., Honig, D.H., Sessa, D.J. & Steggerda, F.R. (1970). Flavor and flatulence factors in soybean protein products. Journal of Agricultural and Food Chemistry 18(6):977-982. doi:10.1021/jf60172a026

6.  Callon, M.C., Cargo-Froom, C., DeVries, T.J. & Shoveller, A.K. (2017). Canine food preference assessment of animal and vegetable ingredient-based diets using single-pan tests and behavioral observation. Frontiers in Veterinary Science 4:154. doi:10.3389/fvets.2017.00154

---

About the Author

Dr. rer. nat. habil. Dr. Seronei Chelulei Cheison is the founder and Chief Executive Officer of Sinonin Biotech GmbH, where he works with petfood companies on palatant sourcing strategies, the innovation of palatability enhancers, and the design of palatability assessment. A trained food enzymologist, he led alternative protein and palatant development at Mars Petcare before founding Sinonin, and his research on protein chemistry spans two decades, a doctorate from Jiangnan University, and a habilitation from the Technical University of Munich.

---

Sinonin Biotech GmbH is a partner in the ZEST project (Grant Agreement No. 101157382) and the PROSCALE project (Grant Agreement No. 101288362), both funded by the Circular Bio-based Europe Joint Undertaking (CBE JU) under the European Union's Horizon Europe research and innovation programme. Views and opinions expressed are those of the author only and do not necessarily reflect those of the European Union or CBE JU. Neither the European Union nor the granting authority can be held responsible for them.

Why Is Africa Still Importing the Protein It Could Make?

Essays & Op-Eds · Protein · Africa

Africa is not short of protein resources. It is short of protein industrialisation.

By Dr. rer. nat. habil. Dr. Seronei Chelulei Cheison3 November 2026

Why does a continent this rich in biomass import its protein?

Why does a Kenyan feed miller buy soybean meal from Zambia, while market refuse, abattoir offal and crop residues rot down the road? Why does a continent with the youngest workforce on earth import fishmeal it could replace with insects? And why does the global debate on sustainable protein take place in Brussels, Washington and Singapore, when the decisive battle will be fought in Nairobi, Lagos and Addis Ababa?

Here is my answer, and I offer it as a provocation.

Africa is not short of protein resources. It is short of protein industrialisation.

I grew up a herdsboy in Nandi, walking cattle to water before school. Today I am a food scientist and biotech entrepreneur in Germany, and I still farm in Kenya. I have seen the protein question from the boma at dawn and from the fermentation hall in Europe. From both ends, the conclusion is the same: Africa is the least prepared arena in the most important food contest of the century.

Is this about hunger, or about industry?

Let me be precise, because the lament is easy and mostly wrong. The United Nations projects Africa's population to rise from about 1.5 billion in 2024 to almost 2.47 billion by 2050. Africa's average protein supply, about 69 grams a head a day in FAO-derived food balance figures, is the lowest of any region. That is supply, not measured intake, and the average conceals wide differences in access, quality and price. The challenge is not simply producing more protein; it is making nutritionally adequate protein available, affordable and accessible to the people who need it. Recorded meat supply is under ten kilograms a head a year, yet a World Bank, FAO, ILRI and AU-IBAR study projected Africa's total meat consumption rising from 10.5 million tonnes in 2005/07 to almost 35 million by 2050.

Where quality falls short, children pay. Africa is the only region where the number of stunted children has grown since 2012, from 61.7 million to 64.8 million, according to UNICEF, the WHO and the World Bank. Stunting has many causes, from infection to sanitation, and protein is only one of them. But the demand is coming either way.

So this is not a lament about protein hunger. It is a question about who will make the protein Africa is about to eat: African factories, or someone else's.

❦

Why does the money look away?

Ask any Kenyan poultry keeper what decides whether the year is profitable. Not the bird: the feed bag. Feed millers put soybean meal at 16 to 22% of a poultry ration, alongside fishmeal. Kenya grows little soya, sunflower or cotton, so it buys protein cake from Uganda, Malawi, Zambia and Tanzania. When Zambia suspended exports of soya and sunflower meal in 2021, Kenyan feed costs hit record highs and soya doubled in price. Is that food security, or a dependency we have chosen not to see?

Then look at the capital. Alternative-protein investment worldwide has fallen for four straight years. The Good Food Institute counts US$881 million raised in 2025, the first year below a billion since 2018. European companies took US$418 million and North American ones US$347 million. What did Africa get? Not enough to register in the headline figures.

Oxford's TABLE programme found that most European and North American funds regard sub-Saharan Africa as too risky for alternative protein. A handful of mission-driven investors are, in practice, the only door open to African founders. Sub-Saharan Africa received about 1% of global agricultural credit in 2019. Development finance still pours into conventional livestock, as if the West's twentieth-century protein model were the only one on offer. Who decided that?

And why must a Kenyan entrepreneur rearing black soldier fly larvae collect separate approvals from KEBS, NEMA and KEPHIS for each product? Each agency has a legitimate mandate: standards, environment, plant health. The complaint an ILRI workshop heard in 2025 was not that they exist, but that they act in sequence rather than together. Three agencies queuing at one larva is not a framework. It is an unnecessary tollgate.

❦

What does Africa have that Europe must buy?

A frontier is defined by what has not yet been built. By that measure, Africa holds several cards that industrialised competitors cannot easily replicate.

It has room. The African Development Bank reckons the continent holds about 65% of the world's remaining uncultivated arable land. Europe is arguing over how to shrink its livestock on land long since spoken for. The figure is contested, and much of that land is better left wild. But Africa can still choose how its land is used, instead of inheriting the choice.

It has feedstock, and it is lying in the open. FAO and the World Bank have put post-harvest grain losses in sub-Saharan Africa at about US$4 billion a year. Kenya's own Livestock Principal Secretary told feed manufacturers in 2024 that about 46% of the livestock feed Kenya produces, most of it fodder and forage, is lost after harvest. Add tea waste, brewery spent grain, bagasse, fruit pulp from export packhouses, abattoir by-products and the refuse of markets in Nairobi and every county town. What Europe calls a sidestream and pays to collect, Africa still calls rubbish and frowns at it. Insects, fungi and fermentation turn that rubbish into protein. My partners and I do exactly this in Europe, in the EU-funded ZEST and PROSCALE projects. Not every residue is idle, mind: dairy farmers already buy brewers' grain, and crop residues feed soils and stoves. A protein industry earns its place only where it adds value, not where it merely outbids the cow.

It has a head start in law. In 2017 Kenya adopted KS 2711, drafted in 2016, a national standard for insect products in compound feed, and in 2020 standards for rearing and edible insects. These were among the first such standards anywhere. ICIPE trained more than 200 small enterprises and 3,000 farmers in black soldier fly rearing. Over 85% of Kenyan feed producers it surveyed said they would use insect meal. In 2025 IFAD and Kenya's aquaculture programme announced more than 500 youth-led black soldier fly units. That is an industry being born from the bottom up, on waste, by the young.

And it has the right first market. A feed miller does not ask whether a protein is natural; he asks for its amino-acid profile, its digestibility and its price per tonne. Feed and petfood buy protein on specification, not sentiment. That makes them the fastest route to scale for novel protein. It is also exactly where Africa's import dependence hurts most.

❦

Is the sun worth more than a subsidy?

In Europe, the weather is a line item. From 2026 to 2028 Germany will compensate eligible energy- and trade-intensive firms for part of their electricity costs, under a scheme pegged to about five cents a kilowatt-hour; an insect or algae start-up is unlikely to qualify. Kenya's climate gives away part of what that money buys: warmth and light. It will not mill, aerate or dry for you. Two protein routes show how.

Parameter	Germany	Kenya	Consequence
Black soldier fly larvae: approximate optimum for growth	27–30 °C	27–30 °C	Same biology; only the cost of providing it differs
Ambient temperature	January means of roughly 0–3 °C; summer heatwaves above 30 °C	Mild in the highlands, with cool nights; warm and humid in the lowlands	Germany heats for months and cools in summer; Kenya trims, ventilates and dries
Spirulina: growth range	Optimum about 30–35 °C for common strains; little growth below 20 °C	Same	Heated greenhouses or closed reactors in Europe for much of the year
Sunlight	About 2.5–3.3 kWh/m² a day on annual average; about 0.5 in December against 6.4 in June	About 6 kWh/m² a day or more over most of the country, with little seasonal swing	Europe buys light in winter; Kenya receives it
Day length	About 8 hours in midwinter, 16 in midsummer	About 12 hours, all year	A stable photoperiod for algae growth and fly mating

Why does a German insect farm pay twice?

Black soldier fly larvae grow fastest at roughly 27 to 30 °C, and what matters is the temperature of the feeding substrate, not merely of the room. At 36 °C survival collapses; at 20 °C growth crawls. In a German winter, halls must be held near 28 °C while it freezes outside, sometimes down to minus 20 °C. In a German summer the bill comes again, because dense larval masses generate their own heat, enough in mass rearing to overheat and kill the larvae if it is not drawn off. The farmer pays to heat in January and to cool in July. Adult flies mate only under light, so breeding rooms are lit year-round. Energy and climate control are routinely named among the heaviest running costs of European insect farms.

The market has delivered its verdict. Enorm Biofactory in Denmark, once Northern Europe's largest insect farm, was declared bankrupt in October 2025, with energy and climate control among its burdens. Second-generation European ventures are moving to Andalusia for the sun. Why stop in Spain?

Kenya is no paradise, and I will not pretend otherwise. The Nandi Highlands are cool at night, so larvae there still need some heating after dark. Lowland heat demands ventilation. Humidity complicates drying, and the long rains make solar drying unreliable. Africa also has its failures: South Africa's AgriProtein collapsed in 2021 after raising more than US$100 million, and Inseco shut down in 2025, undone in part by power cuts. Climate is an advantage; it is not a business plan.

Why light a pond in Hamburg when the equator is free?

Spirulina is roughly 60% protein by dry weight. It grows best between 30 and 35 °C and needs 10 to 14 hours of good light a day. A commercial producer in sunny Portugal records winter output more than 60% below summer. A northern European producer who wants year-round output must heat the culture and light it with LEDs, or accept a lean winter. Dutch growers are now building fully indoor, daylight-free systems. Kenya sits on the equator. Arthrospira, the cyanobacterium sold as spirulina, blooms naturally in the alkaline soda lakes of Kenya's Rift Valley, such as Nakuru and Bogoria, where it feeds the flamingos. Yet Kenya imports spirulina from Asia. What irony!

Sunlight is free. Converting it into saleable protein is not. Too much sun can stall growth, open ponds evaporate scarce water, and harvesting and drying cost energy anywhere. Algae grown in the dark on sugars, in fermenters, gain nothing from Kenya's sunshine at all.

So the honest test is not Germany versus Kenya. It is a specific German plant against a specific Kenyan plant, run for a full year to the same specification, measured by total annual energy cost per kilogram of saleable, specification-compliant protein. On that measure, I am willing to bet on Kenya. Is anyone willing to fund the test?

❦

Too theoretical? Come to Chebole, Bomet and Kipkenda, Nandi.

In 2022 Hyvene Kirui, a trained secondary-school teacher, turned to farming on less than an acre at Chebole in Bomet County. In 2023 her farm became the first project I supported through WIDU.africa, the German diaspora grant programme. Today it carries poultry, including ornamental birds, organic vegetables, Dorper sheep, pigs and dairy. Eight 550-watt panels, 4.4 kWp of peak capacity, charge a 5 kWh battery that runs all her lighting and water pumping. A biogas digester supplies backup energy. Black soldier fly larvae and Azolla, a nitrogen-fixing water fern, supplement the rations of her poultry and sheep. Frass and digester slurry go back to the soil as organic fertiliser. On less than an acre, she employs herself and five others.

My own farm, Kipkenda Poultry in Nandi County, makes the same point at a larger scale. Sixteen 500-watt panels, 8 kWp of peak capacity, charge two 10 kWh lithium batteries. They run everything: brooder heating, lighting, security cameras, an incubator that hatches 1,056 eggs at a time, and the machinery that processes our own poultry feed. Beside the poultry houses, 100 square metres of Azolla grow a green protein supplement for the flock.

Heating chicks, milling feed and running an incubator day and night are the very loads a European farmer pays dearly for each winter. In Bomet and Nandi, the sun pays. What Europe models in strategy papers, a Bomet schoolteacher is already doing. So why are there not ten thousand of her?

❦

If the grid is green, why is the power so dear?

Kenya draws about four-fifths of its electricity from renewables, and in the year to June 2026 geothermal from the Rift Valley alone supplied about 41%. Germany managed 55% in 2025. So why does a Kenyan factory pay more for power than many of its competitors?

Because renewable does not mean cheap. The Kenya Association of Manufacturers puts industrial power at 18 to 23 US cents a kilowatt-hour, and KIPPRA, the government's own think-tank, finds Kenyan industry paying about twice what China's pays. Uganda has cut rates for large manufacturers to about a third of Kenya's. Monthly fuel and foreign-exchange surcharges add volatility on top. A tariff built on cheap geothermal steam and then loaded with charges is a policy choice, not a law of physics.

The opening lies in heat. Geothermal steam is not free: wells, maintenance, reinjection and pipework must all be paid for. But it has no fuel bill, and fermentation, drying and rendering need heat as much as electricity. A protein plant at Olkaria could tap geothermal process heat at costs and carbon intensities that many European competitors would struggle to match. That is worth proving plant by plant, because a geothermal field does not guarantee cheap heat until someone signs a contract for it.

Factor	Kenya	Germany
Industrial electricity	US$0.18–0.23 per kWh (KAM, 2026)	about €0.16 per kWh without relief, €0.10 with; compensation pegged to about €0.05 for eligible energy- and trade-intensive firms, 2026–2028
Renewable share of electricity	about 80% (EPRA, 2025)	55.1% (UBA, 2025)
Statutory minimum wage, monthly	KSh 18,047, about US$139, general labourer in the five cities (2026)	€2,409 gross (2026)
Average bank lending rate	14.3% (CBK, July 2026)	far lower
Median age	19.3 years (Africa)	42.8 years (Europe)

Is cheap labour a strategy, or an insult?

A Nairobi labourer's legal minimum is about US$139 a month; Germany's is €2,409. Kenyan technical colleges enrolled 825,484 students in 2025, and universities hundreds of thousands more. It is tempting to sell Africa on cheap, plentiful hands.

Resist it. A biotech plant runs on fermentation technicians, process engineers, microbiologists, analysts and maintenance crews. The measure that matters is not the wage but total labour cost per kilogram of specification-compliant product. A German plant with high salaries, automation and steady uptime can beat a Kenyan plant with more staff, lower yields and frequent breakdowns. A Kenyan plant with skilled, well-paid technicians and high utilisation can beat them both.

That should be the ambition: decent, sustainable jobs, not cheap jobs. Kenya created about 822,000 jobs in 2025, and nearly nine in ten were informal. A protein industry that offered formal contracts, safe plants, training and careers would do more for the young than any wage advantage. An industry that depends on low pay cannot call itself sustainable. Otherwise it is a sweatshop with a sustainability report.

❦

What does a 15% loan do to a good idea?

It can kill it. Take a hypothetical €10 million protein plant. Financed at 6%, it pays €600,000 in first-year interest. At 15%, it pays €1.5 million. The €900,000 gap can wipe out every saving that sunshine and wages provide. These are illustrative rates, not quotations, but Kenya's commercial lending averaged 14.3% in July 2026, with the central bank rate at 8.75%. Some argue that heavy government borrowing at home crowds out private lending. I leave that debate to the economists.

Why is money so dear? Public debt stands near 68% of GDP, and debt service is consuming about 72% of ordinary revenue in 2025/26, with nearly 80% projected for 2026/27. The Treasury hunts for revenue through each year's Finance Act, and investors never know which tax will move next. The World Bank's November 2025 Kenya Economic Update found Kenya's regulation more restrictive to competition, deterring private investment. The US Trade Representative's 2026 trade-barriers report flags bribery in public procurement. Add imported equipment, freight, currency swings and spare parts that take months to arrive. The true cost of a Kenyan protein plant is far above its interest rate.

Development finance, concessional loans, equipment leasing and public–private partnerships can close part of that gap. But let me be equally blunt with my own side: cheap money must not be used to disguise an uncompetitive plant. The only number that settles the argument is the landed cost:

Landed cost = feedstock + energy + labour + capital recovery + maintenance + quality assurance + compliance + packaging + logistics + risk allowance.

Does shipping kill the carbon case?

Not necessarily. A kilogram of protein made on a geothermal grid, from local waste, for a local feed mill, may well carry lower emissions than soybean meal grown on cleared land in South America and shipped to Mombasa. That is a hypothesis worth testing, not a result I can yet claim. For a shelf-stable ingredient sent to Hamburg by sea, freight may be a modest share of the total footprint.

But "should" and "may" are not evidence. EPRA reported that Kenya's grid emission factor rose about 18% in the second half of 2025 as thermal generation climbed, and imports now supply around 12% of electricity. The proper measure is kilograms of CO₂-equivalent per kilogram of specification-compliant, digestible protein delivered to the customer, assessed to ISO 14040 and 14044 against the German alternative. Europe demands that proof. Kenya should welcome the test, and invest in the conditions required to pass it.

❦

Where should Kenya build what?

Why does every county want a special economic zone, and almost none a specialisation? Kenya should not chase every protein technology everywhere. It should put each where the resources, utilities and buyers already are.

Location	What belongs there	What must be proven
Nairobi–Athi River	Fermentation, hydrolysates, ingredient formulation, analytical services; De Heus opened a 200,000-tonne feed mill here in 2026	Land and utility costs; maintaining sophisticated plant
Naivasha (Olkaria)	Geothermal-linked fermentation, drying and integrated biorefineries in the special economic zone	Delivered heat and power prices; water; contracts, not maps
Eldoret–Nandi	Insect protein, feed ingredients, farmer-linked circular enterprises; tea and maize sidestreams	Feedstock aggregation; cool highland nights in BSF models
Mombasa	Export consolidation, storage and packing at the port	Port handling, clearance and inland freight in the landed cost

The export doors are already open. Since 1 July 2024 the EU–Kenya Economic Partnership Agreement has given every Kenyan export, arms excepted, duty-free and quota-free entry to the European Union. AGOA, reauthorised in February 2026, was extended in September 2026 to December 2028. Mombasa is the gateway to Uganda, Rwanda, South Sudan and eastern Congo.

But is an open door a market? Duty-free is not the same as access: feed and petfood bound for Europe must still clear EU rules on animal by-products, feed hygiene and approved establishments. And Mombasa's flagship Dongo Kundu zone, conceived more than twenty years ago on some 3,000 acres, had drawn only two major private investors by August 2026, held back by missing roads, water and power. Kenya does not lack plans. It lacks finished infrastructure.

❦

Is the law an asset or a tollgate?

Biology does not confer a licence. A larva that can eat something does not make that thing a legal feed. The European Union allowed insect protein in aquafeed in 2017 (Regulation 2017/893) and in pig and poultry feed in 2021 (Regulation 2021/1372). It still forbids feeding insects on catering waste, manure or meat-and-bone meal. A Kenyan plant that wants to sell to Europe must, therefore, give up the cheapest substrates, meet EU feed-hygiene and animal-by-product rules, and qualify as an approved establishment. A product that is lawful in Nairobi is not automatically lawful in Hamburg.

Algae have their own maze. Spirulina was on the European market before 1997 and needs no novel-food approval; other species and processes may. Classify the product before you build the plant, not after.

And let no one mistake weak regulation for a friendly business climate. Lax safety rules destroy consumer trust, close export markets and create liabilities that outlive the founders. Kenya's opportunity is to offer regulation that is science-based, fast and compatible with its export markets. Predictability is an investment incentive that costs the Treasury nothing.

❦

Is it all rosy? An honest SWOT

No. A reader who has been sold Africa before deserves the weaknesses beside the strengths. Note where they cluster: not in biology, but in the systems needed to industrialise it.

SStrengthsInternal · helpful

A young, increasingly technical workforce

Abundant, largely unused sidestreams

A largely renewable grid with geothermal heat

A climate suited to insects and phototrophic algae

A strong research base (icipe, ILRI, universities) and agricultural export experience

Legal National insect standards among the world's first (KS 2711, 2017; KS 2921 and KS 2922, 2020)

WWeaknessesInternal · harmful

Industrial power that is green but not cheap

Expensive, short-tenor capital and imported equipment

Inconsistent feedstock quality and aggregation

Few pilot and scale-up facilities

Legal Approvals split across KEBS, NEMA and KEPHIS; no clear pathway yet for fermentation-derived or algal feed protein; standards not harmonised across the EAC

OOpportunitiesExternal · helpful

The fastest-growing protein demand on earth

Import substitution for soybean meal and fishmeal

Feed and petfood as specification-driven entry markets

Higher-value exports: hydrolysates, palatants, specialty proteins

EU buyers seeking low-carbon, deforestation-free protein

Legal EU–Kenya EPA; EU approval of insect protein in fish feed (2017) and pig and poultry feed (2021); spirulina's established EU status

TThreatsExternal · harmful

A four-year retreat of global alt-protein capital

Currency swings, drought and water stress

Established foreign competitors and cheap imports

A safety failure that taints the whole sector

Dependence on foreign technology and IP

Legal EU bans on catering waste and manure as insect feed; EU establishment rules; AGOA's annual review; abrupt tax and court reversals

SWOT: Kenya's sustainable-protein industry. Strengths and weaknesses are internal; opportunities and threats are external.

❦

What does an investor actually see? A PESTEL scan

The SWOT looks inward, at what Africa and Kenya bring. The PESTEL looks at the weather outside the factory gate. Read it row by row and one pattern stands out: several of the most consequential constraints are institutional, financial and regulatory, and therefore open to deliberate reform. Water, distance and biology are not; finance, permits and tariffs are.

PPolitical

Tax protests in 2024, fuel-price unrest in 2026, a general election due in August 2027; AGOA under annual US review

So Plan for disruption; never rely on one trade window

EEconomic

Debt near 68% of GDP; debt service above 70% of revenue; lending at 14.3%; no IMF programme since early 2025

So Expect tax pressure and dear credit; seek blended finance and hard-currency offtake

SSocial

Median age about 19; TVET enrolment above 800,000 in 2025; most new jobs informal

So A ready workforce and a hunger for formal careers; build training into the business

TTechnological

Strong research base (icipe, ILRI, universities); almost no shared pilot fermentation capacity

So The lab-to-plant gap is the bottleneck; a shared pilot plant would unlock a cluster

EEnvironmental

About 80% renewable power but a rising thermal share; drought; water stress; large post-harvest losses

So Low-carbon claims are possible but must be measured; secure diverse feedstocks

LLegal

Early insect standards; fragmented permits; EU export rules; volatile Finance Acts

So Treat regulation as part of industrial strategy, not an afterthought

The verdict. Kenya has the biology, the talent and the research base to build a competitive sustainable-protein industry. Whether it succeeds depends on finance, infrastructure, regulatory predictability and measured performance.

PESTEL: the operating environment for a Kenyan protein plant.

❦

Are the SDGs a balance sheet or a brochure?

How many glossy reports have you read that tick five Sustainable Development Goals and measure none of them? Sidestream protein can serve several goals at once. But a claim counts only if it comes with a number, so here are the numbers to be held to.

Goal	Kenyan baseline	Indicator to report
SDG 2: Zero hunger	64.8 million stunted children in Africa, the only region where the number is rising	Cost per kilogram of digestible feed protein; price and household consumption of eggs, fish, milk and meat
SDG 8: Decent work	Nearly nine in ten new jobs in 2025 were informal	Formal jobs created, wage adequacy, safety record, women and young people employed
SDG 12: Responsible production	About 46% of the feed Kenya produces is lost after harvest	Share of incoming feedstock converted into saleable products, from a full material balance
SDG 14: Life below water	An estimated 60–80% of Lake Victoria's omena catch goes to animal feed	Tonnes of fishmeal displaced, with catch and stock indicators for the source fisheries
SDG 7 and SDG 13: Energy and climate	About 80% renewable power; a national target of a 35% emissions cut by 2035	kg CO₂e and kWh per kg of saleable product; litres of fresh water per kg
SDG 9 and SDG 17: Industry and partnership	Almost no shared pilot capacity; strong research partners in Europe	Local value added, technologies commercialised, intellectual property owned in Kenya

Is a job enough?

SDG 8 does not ask for jobs. It asks for decent work: safe, formal, fairly paid and skilled. Counting heads without counting quality is how development reports flatter themselves. A protein plant will not employ millions, but it can create the technician and laboratory careers a TVET graduate can build a life on, far from Nairobi. It can lift the incomes of the women who dry omena and ochonga and the smallholders who sell crop residues by giving them a buyer with a contract and a quality premium. Hyvene Kirui's farm shows the effect: six livelihoods on less than an acre, sustained by sunlight, insects and waste, sorry, sidestreams.

Can anything be zero waste?

Zero waste is a legitimate engineering ambition, not an automatic property of circular production. A black soldier fly unit fed on brewery grain can yield protein meal, insect oil, frass for treated soil use and, where it pays, chitin; every output needs processing, a safety assessment and a buyer. A spirulina pond can absorb CO₂ from fermentation. Very little in such a cascade need be discarded, though some residues and treatment losses always remain. But a zero-waste claim requires a material balance that shows what enters, what is sold, what is recovered and what remains. It also requires safety: circularity is no excuse for feeding contaminated substrate into the food chain. The honest goal is to maximise safe, profitable recovery, not to process waste at any cost.

❦

Is a kilogram of protein a kilogram of protein?

No, and this is where most of the alternative-protein debate goes wrong. A laboratory report that says "55% crude protein" tells a buyer almost nothing about what he is paying for. Crude protein is nitrogen multiplied by a factor. Value lies elsewhere.

It lies in the amino-acid profile: does the meal supply the lysine and methionine a broiler cannot make for itself? It lies in digestibility: how much of that nitrogen does the animal actually absorb? It lies in anti-nutritional factors, consistency from batch to batch, and safety. In petfood, it also lies in palatability, because a dog or cat that refuses the bowl renders every nutritional claim irrelevant. And in human food it lies in functionality: solubility, foaming, emulsification and taste.

Two meals with the same crude protein can differ widely in price, because they differ in what they do. That is why I spend my working life on petfood ingredients and palatants rather than on bulk meal. A hydrolysate tailored to a customer's specification commands a premium that an undifferentiated tonne never will.

The lesson for Africa is blunt. Sell specifications, not tonnes. A continent that exports crude protein will be paid like a supplier of raw materials. A continent that sells digestibility, amino-acid balance and palatability will be paid like an industry.

❦

Where should Kenya begin?

Not everywhere at once. The surest way to waste a frontier is to chase every technology on it.

The first phase is the home market. Insect protein, selected plant proteins and by-product processing should supply Kenya's own poultry, fish and pig farms. Kenya needs about 55 million tonnes of livestock feed a year by dry matter and produces only about 40% of it; roughly 400 millers make 2.5 million tonnes of compound feed. In July 2026 the government announced a ten-year National Feed Strategy valued at about KSh 465 billion, still being prepared rather than funded. Domestic demand is large enough to build an industry on before anyone ships a tonne abroad.

The second phase is shared regional hubs: pilot plants, analytical laboratories, quality assurance and training, sited where the feedstock, utilities and buyers are. No start-up should have to build its own pilot plant.

The third phase is export of higher-value ingredients: hydrolysates, functional proteins, fermentation products and petfood palatants. Kenya's own petfood market is small, about 2,568 tonnes in 2022, but sheltered by a 35% tariff and growing with urban pet ownership. Margins abroad are better than in bulk meal.

Keep the two markets apart. Replacing imported feed ingredients at home is about cheaper eggs, fish, milk and meat for Kenyans; that is where the nutrition dividend lies. Exporting hydrolysates and palatants is about margins and skilled jobs; its effect on the Kenyan plate is indirect. Both matter, but they are different businesses and need different policies.

The fourth phase is ownership. Kenyan firms must take part in process development, equipment adaptation, formulation, intellectual property and international marketing. Otherwise Africa will supply the waste while others capture the value, as it has with cocoa, coffee and tea.

What should be demanded, and of whom?

Of Nairobi: count protein the way you count maize, with a national protein balance sheet that includes livestock and pets. Write novel and sidestream protein into the National Feed Strategy, with targets for replacing imported soybean meal and fishmeal. Replace three regulators with one licence for insect, fungal, algal and fermentation protein, recognised across the East African Community. And make the Naivasha zone a protein cluster, with a guaranteed industrial tariff, direct geothermal heat and a shared pilot plant open to start-ups and universities.

Of the development financiers: stop treating alternative protein as a luxury for rich consumers. It is import substitution, waste management, youth employment and child nutrition in a single investment line. Shift a modest share of the capital now committed to conventional livestock into sidestream-protein plants beside breweries, tea factories, abattoirs and markets.

Of Brussels and Berlin: treat Africa as a partner in this transition, not a market for it. This is not an argument for relocating European plants to lower-wage countries. It is an argument for joint development: research in both places, pilot plants that test the same process under both climates, manufacturing sited where it is demonstrably cheapest and cleanest, and African ownership of a share of the technology. Research programmes on sidestream protein should include African partners and pilot sites by design, not as a line in the dissemination plan.

❦

So, can Kenya land a kilogram in Hamburg?

Africa skipped the landline and went straight to mobile money; the world now studies that leap. Must it now copy the North's path of imported soya, confined livestock and the long, expensive retreat that follows? Or can it build a protein economy that runs on its own sunlight and its own sidestreams from the start?

The advantages I have described are real, and so are the traps. A climate advantage disappears in an inefficient plant. A wage advantage disappears with low productivity. A carbon advantage disappears with poor design or careless logistics. Africa does not need another flattering narrative about its potential. It needs an industrial strategy that survives an audit.

So here is the test I put to every minister, financier and fellow scientist who reads this:

Can Kenya manufacture a kilogram of sustainable protein, deliver it to Hamburg at a competitive price, prove a lower footprint, and create decent jobs doing it?

If the answer is yes for a given technology, it is an opportunity worth funding now. If the answer is no, the analysis must tell us why, and what investment, technology or policy would change it. That is how a frontier becomes an economy.

When I drove cattle as a boy, the elders measured a man's wealth by his herd. The next generation may measure a county's wealth by something less visible: how much protein it can make from what it once threw away. The frontier is open. Will Africans settle it, or will we once again import the finished product of someone else's ingenuity?

Dr. rer. nat. habil. Dr. Seronei Chelulei Cheison is a food scientist and the founder and CEO of Sinonin Biotech GmbH, which develops alternative proteins for petfood. He farms in Nandi County, Kenya, where he runs Kipkenda Poultry. He leads Work Package 5 of the ZEST project, is a partner in PROSCALE, serves on the advisory board of WIDU.africa and acts as a consultant focal person for the programme. ZEST (Grant Agreement No. 101157382) and PROSCALE are funded by the Circular Bio-based Europe Joint Undertaking and its members under Horizon Europe. Views and opinions expressed are those of the author only and do not necessarily reflect those of the European Union or the CBE JU, neither of which can be held responsible for them.

References

Numbered in the order the claims appear.

1. Population 1.51 bn (2024) to ~2.47 bn (2050): UN World Population Prospects 2024, via INED.
2. Protein supply 69.1 g/capita/day, lowest region: Gebremedhin & Bekele 2021 (PLOS ONE), from FAO food balance data, summarised by Nestlé Nutrition Institute Africa; protein quality: Schönfeldt & Hall, Br J Nutr 2012.
3. Meat 10.5 Mt (2005/07) to 34.8 Mt (2050): World Bank, FAO, ILRI & AU-IBAR, Investing in African Livestock, 2013; per-capita supply under 10 kg: OECD-FAO data via Earthworm Express.
4. Stunted children 61.7 m (2012) to 64.8 m (2024): UNICEF/WHO/World Bank JME 2025, reported by TheCable.
5. Soybean meal at 16–22% of poultry feed: The Standard; Zambia's 2021 ban on soya and sunflower meal: Nation.
6. Alt-protein investment US$881 m in 2025; Europe US$418 m; North America US$347 m; fourth annual decline: GFI via Green Queen.
7. Investor risk perception: Brice, TABLE, Oxford; 1% of global agricultural credit: Alliance Manchester Business School.
8. Separate KEBS, NEMA and KEPHIS approvals, as reported by industry participants: ILRI PAFIE workshop, July 2025.
9. 65% of uncultivated arable land (AfDB claim): AfDB via MyJoyOnline; US$4 bn grain losses: FAO/World Bank, Missing Food, 2011.
10. 55 Mt feed requirement, 46.3% lost (PS Jonathan Mueke, 2024): The Star.
11. KS 2711:2017 (draft DKS 2711:2016, WTO TBT notification, Jan 2017; published standard cited by KEBS as KS 2711:2017 in DKS 2921:2020): GREEiNSECT brief; KS 2921:2020 and KS 2922: People Daily; INSFEED: IDRC; 500 planned BSF units: AllAfrica.
12. German industrial power compensation, 5 ct/kWh to 2028: Polyform; BNetzA prices: Ugebrev.
13. BSF survival at 27, 30 and 36 °C: Tomberlin et al. 2009; larval heat in mass rearing: Black soldier fly larval heat generation and management; European cost pressures: Ryba 2024.
14. Enorm bankruptcy, Oct 2025: AgFunderNews; AgriProtein: Wikipedia; Inseco: AgFunderNews; Andalusia: AgTechNavigator.
15. Spirulina winter output: Allmicroalgae, Portugal; lighting costs: OCL 2021; indoor systems and EU status: FreshPlaza; Kenya imports: WAS 2025; Arthrospira in East African soda lakes: FWF project P19911.
16. German irradiation by month: Grünes Haus; Kenya irradiation: Kariuki & Sato.
17. Kenya 80% renewable: AllAfrica/EPRA; geothermal 41%, imports 12%: Capital FM/EPRA; Germany 55.1%: UBA; industrial tariffs: KAM via The Star, KIPPRA via The Standard.
18. Minimum wages: Nation (Kenya), Connexion Emploi (Germany); TVET enrolment: Capital FM; jobs: The Standard.
19. Lending rate 14.39%: CBK via Khusoko; debt 67.6% of GDP: People Daily; debt service: Business Daily; World Bank: Zawya; USTR 2026: Capital FM.
20. Grid emission factor: The Star.
21. De Heus mill: AllAfrica; EU–Kenya EPA: European Commission; AGOA to 2028: BusinessDay, KNA; Dongo Kundu: The Standard.
22. EU insect feed rules: Regulation (EU) 2017/893, Regulation (EU) 2021/1372.
23. Omena to feed (estimate): Manyala; 35% emissions target: The Star.
24. National Feed Strategy, 400 millers, 2.5 Mt: The Star, 2026; petfood market: Petfood Industry/USDA.
25. Chebole and Kipkenda farm data: the author's first-hand operational records.

How Carbohydrate Shapes Petfood Palatability

The Carbohydrate Paradox. The largest ingredient in many a bowl is one no dog, cat or person tastes as itself. So what, exactly, does an animal prefer when it prefers the food? On the carbohydrate that may never be tasted as carbohydrate, and the food that would not be the same without it.

Movement I

THE CARBOHYDRATE NOBODY TASTED

What is an obligate carnivore really eating when it works through a bowl of dry food? Tip the measure in, and much of what holds that kibble together as a hard, dry, pourable object is starch, the carbohydrate that gives the piece its body and lets it leave the machine a solid thing. The cat may eat it readily. But does it taste it? Not as anything much. Purified starch carries little of the sweetness, sourness, bitterness or savour the tongue is tuned to, and it does not announce itself to the mouth the way a sugar does. We can build a food around it, an animal may come to prefer that food, and the starch itself goes all but unremarked by the palate that eats it.

That is the paradox worth opening on, and the obligate carnivore states it most sharply. Here is an animal whose evolutionary brief was meat, eating willingly a food whose very architecture may rest on the one macronutrient it has least instinct to seek, and finding little to taste in it. Starch is the bowl's silent partner, the biggest thing in it that the tongue barely registers. So how can it have anything to do with whether the animal wants to eat?

Starch is the bowl's silent partner, the biggest thing in it that the tongue barely registers.

The dog meets the same bowl as a different animal. It can register a sweetness the cat cannot, and it tends to handle cooked starch more readily, though how readily depends on how the starch was processed and on the dog itself. So its relationship with carbohydrate, in the mouth and after, is not the cat's. But the sweetness a dog can find lives in sugars, and the bulk carbohydrate in a dry food is not sugar; it is starch, which is no sweeter to a dog than to a cat. Whatever divides the two species, and a good deal does, it sits on top of something they share: the structural carbohydrate that builds the food gives the tongue little to work with, in either of them, or in us.

So what is this essay about, if not whether a dog or a cat needs carbohydrate, or which grain belongs in which bowl? Those are arguments for another room. The subject here is narrower and stranger, the sensory life of an ingredient that is, for the most part, not sensed as itself, and it turns on a proposition the trade's vocabulary resists. An ingredient need not be pleasurable on its own tongue to make a food more pleasurable to eat, and its influence need not run for the better: the same carbohydrate can help a food or hurt it, depending on the chemistry, the amount, the processing and everything else in the bowl. Palatable is not the same as preferred, preferred is not the same as digestible, and none of the three is the same as good for the animal. A food can be relished and wrong; a carbohydrate can be digested and unwanted; a carbohydrate can be barely tasted and still decide the meal.

How, then, does carbohydrate reach the animal's judgement at all? By three routes, and only the first runs through taste. A sugar can be tasted, by the animal equipped to taste it. The energy a digestible carbohydrate delivers can, after the meal, teach an animal something it carries back to the bowl. And the carbohydrate can change what the food is as a physical and chemical object, how it fractures, how it floods the mouth, what it smells of once heat has had its way, so that an animal may prefer the food without preferring the carbohydrate in it. The cat is shut out of the first door, the one marked sweet. The other two stand open to it, as far as we can tell, and through them the carbohydrate it cannot taste still helps build its bite and make its aroma. That is where we are going.

---

Movement II

SWEETNESS IS NOT THE WHOLE STORY

Why does a dog take an interest in something sweet while a cat looks straight past it? Offer each the same taste and you meet one of the most reliable indifferences in biology. The cat is not being fussy. It is missing a part. Somewhere in the ancestry of the cats the gene for one half of the sweet-taste receptor, the subunit called Tas1r2, decayed into a pseudogene, and the receptor that reports sweetness to the brain was never again built. The dog keeps it. The cat does not. A sugar that can please one of our two animals leaves the other largely unmoved, at least by way of sweetness. How much any particular dog likes any particular sugar is its own question, turning on the sugar, the amount and the individual; the point here is simpler and deeper, that one species can enter this part of the conversation at all and the other cannot.

But sweetness of what, exactly? In the sense the receptor means, it is mostly a report on sugars, the small soluble molecules, the glucose and sucrose and maltose and their short-chain kin, though a handful of non-sugar compounds can trip the same receptor too. What it was never a report on is carbohydrate as a whole. The starch that makes up the bulk of the carbohydrate in a bowl is not a sugar and does not read as sweet, and neither does dietary fibre. So the great species drama of sweet taste, real as it is, is a quarrel over the sugars, a thin and soluble fraction of the whole. It leaves the structural carbohydrate just where Movement I left it, giving the tongue little to go on.

The species drama of sweet taste is a quarrel over the sugars, a thin and soluble fraction of the whole.

Does the mouth not get a head start on the starch, at least? In the human mouth it does: salivary amylase begins cutting starch into maltose while we chew, which is why a plain cracker held long enough turns faintly sweet. It is tempting to carry that experience into the dog's mouth and the cat's. We should not lean on it. The dog and the cat are reported to carry little amylase in the saliva, leaving the heavy work on starch for the gut, so whatever their kibble's starch becomes later, it is not likely becoming sweetness on the way in. The human cracker illustrates the mechanism. It does not prove the carnivore's case, which needs evidence of its own.

All of which should warn us off a comfortable assumption, that an ingredient has to taste good in order to do good at the bowl. Does a carbohydrate have to be sweet, or even much tasted, to help decide whether an animal eats? It does not. There may be taste-active compounds that travel with carbohydrates, and gustatory effects beyond sweetness worth taking seriously, but the dog-and-cat evidence for most of them is thin, and the honest course is to say so rather than reach past it.

And we should keep tasting and preferring apart, as Movement I insisted. The sweet-preference work that does exist tends to rest on particular sugars and sweeteners, offered under particular conditions, to particular animals; a response to one sugar in one test is not a law of the species. That a dog can taste sweetness does not tell us how much a given food's appeal owes to it. That a cat cannot does not tell us the carbohydrate in its food is idle. The tongue has cast its one clear vote, for the sugars, in the animal built to count them. The rest of the carbohydrate is still waiting to be heard, after the meal and in the mouthfeel, long after sweetness has had its say.

---

Movement III

THE CALORIES AFTER THE TASTE

What does a meal keep telling an animal once it is swallowed? The taste test ends at the mouth. The food does not. It goes on reporting for hours, as energy arriving, a stomach filling, a body settling or unsettling. Three quite different things can grow from that reporting, and they are easy to mistake for one another: what an animal learns to expect from a food, how much of a nutrient it will take, and what it comes to prefer over a lifetime. They lean on one another. They are not the same, and a finding about one is not a finding about the rest.

Take the first, what an animal learns to expect by tying a food's smell and taste to the consequences that follow. Where does the evidence for it come from? Chiefly from work on rodents and on people, and it would be a mistake to lift that wholesale into the carnivore's bowl. The honest position is that the route plausibly operates in dogs and cats, and that the species-specific evidence is not yet strong enough to say how much of any real food's acceptance it explains. One caution follows that the trade does not always observe. Energy is not a flavour, and delivering it is not the same as being liked. A carbohydrate can pour digestible calories into a food without the animal preferring that food one bite more; the calories may buy acceptance slowly, over many meals, or not at all. And the quick first-bite test the industry leans on measures the front of the meal, which is not the preference an animal settles into over days of eating. The two can disagree.

Energy is not a flavour, and delivering it is not the same as being liked.

The second thing, the amount an animal chooses to take, is where the cat turns strange. Allowed to compose its own diet from foods of differing make-up, does it take carbohydrate without limit? It does not. It settles toward a balance weighted to protein and fat, and appears to meet a ceiling on how much of its energy it will draw from carbohydrate, stopping short even when more is on offer. Dogs, given a comparable choice, have been reported to settle at a different balance, taking a larger share of their energy from carbohydrate than cats do, though what any such study finds depends on the foods it offered and cannot be read as a fixed appetite for the nutrient. What is arresting about the cat's ceiling is what it cannot be. It cannot be a verdict of the tongue, which, as Movement II showed, has almost nothing to say about the carbohydrate. The body casts a vote the tongue cannot.

But cast how, and for what? The line has to be read for exactly what it claims, and no more. The ceiling tells us only that the cat's intake of carbohydrate is held in check by something other than canonical sweet taste, because there is no sweet taste to hold it. What that something is remains unresolved. It might be another property the animal senses in the mouth, or a signal that arrives long after, of energy, of protein sufficiency, or of some consequence learned or innate. The experiment that reveals the limit does not reveal its mechanism. And we should resist calling the ceiling a dislike: regulating how much of something an animal eats is a different act from being pleased or displeased by its taste, and different again from the preference it forms over a lifetime.

So the carbohydrate's dealings with palatability neither end when the sweetness is tallied nor fail to begin where there is no sweetness to tally. They continue out of sight, in the energy the food delivers and the limits that govern how much of it an animal will take. That is one reckoning made away from the tongue. The next movement is about the other, made in plain contact with it, in the bite and the smell of the food itself, and it is the one the carbohydrate shapes most directly of all.

---

Movement IV

THE STARCH THAT BUILT THE BITE

What makes a kibble crunch? Listen to a dog work a piece of dry food and you are hearing a whole structure fail under a tooth: starch, protein, fat, air and water, set by heat and shaped by the die, each with a hand in how the piece breaks. But in a great many dry foods the starch has done much of the building. Cooked in the presence of heat and water, it swells, gelatinises and then sets as the piece dries, giving the kibble a rigid, porous body that holds its shape on the shelf and shatters in a particular way in the mouth. We spent a whole Conversation, No. 8, on the physics of that first bite. Here is one of the ingredients that most shapes it.

The crunch is not the sound of starch alone. It is the sound of a structure in which starch has often done much of the building.

Where does the water go in all this? Into that same structure. The grip starch keeps on moisture helps set a kibble's mechanics, which is why a piece left open to the air softens and loses its snap: starch meeting water, the handshake No. 13 described, running the wrong way. And the body starch helps build is the stage the other sensory actors stand on. The fat and the palatant an animal first meets sit on and in a surface that starch has helped make rigid and porous enough to carry them. The carbohydrate the tongue barely reads is holding up much of what the tongue is there to enjoy.

Change the format and does the carbohydrate change its nature? It changes its job, not its nature. In a semi-moist food there is no crunch to build; the work is to hold a soft, yielding, faintly tacky piece together, moist without being wet, and to keep it from spoiling at that moisture. Sugars do much of this, holding water in a form the food can keep and lowering its availability to spoilage without drying the piece to a biscuit, though they share the work with humectants that are not carbohydrates at all, glycerol and the like, a reminder that not every tool of water management is a sugar. What the animal meets is a texture as much as a taste, the give, the chew, the way the piece clings for a moment to the tooth. It is the water activity of No. 13 again, with the carbohydrate pressed into service as a tool of structure and stability, barely as a flavour at all.

And in a wet food? There the carbohydrate turns fluid mechanic. Starches and the hydrocolloids, the plant gums and their relatives, are what thicken a gravy, set a jelly and manage how water sits in the can. They are not one material doing one thing: starch gelatinises, gums hydrate and thicken, fibres bind water and add bulk, and they are different hands on the texture. What holds a chunk together is often not the gum at all but the setting of protein, with the carbohydrates working alongside it rather than in its place. A formulator who confuses these roles gets a sauce that breaks or a chunk that slumps. And the roster is still growing: novel plant fibres are being explored to reshape a wet food's body, binding water and, in some formulations, easing the stickiness of a chunk or a pâté, a sign that the carbohydrate's textural reach is still being extended, not exhausted.

Fibre deserves its own line, because it is the case that most tests the essay's logic. It is barely tasted, and we are tempted to call it barely nourishing, but that is only half right. The poorly fermentable fibres pass through largely for their bulk; the fermentable ones are broken down by the gut's microbes into short-chain fatty acids the animal can use, a quiet contribution of energy that varies with the fibre, the species and the microbes it meets. Set that aside, though, because energy is not where fibre earns its keep at the bowl. So how does it? Through bulk, through the way it holds water, through the body it gives and the things it dilutes. Here is a carbohydrate valued in the food for neither its sweetness nor its calories, and still it has a say in the texture, and the texture has a say in the meal.

And the smell? One correction keeps this honest, and it is the same one No. 14 turned on. The browning and the roasted notes that cooking builds come from reducing sugars meeting amino compounds, and starch, though it is built from glucose, behaves almost nothing like a free reducing sugar. A starch molecule carries just one reducing end, however many thousands of glucose units it strings together, which leaves very little reducing-end reactivity for its mass, far too little to drive much browning on its own. Starch can build the scaffold; the reducing sugars that help build the aroma are its smaller relatives and its breakdown products, not the intact starch itself. The jobs belong to the same chemical family. They do not all belong to the same molecule. That was the whole of No. 14 and need not be re-run here, beyond the reminder that the carbohydrate is at work in both the structure and the smell, by different members of its family.

This is the third route, and the most direct the carbohydrate has. It does not wait to be tasted or digested. It changes what the food is as an object, and an animal may prefer the food the carbohydrate helped build without preferring the carbohydrate in it. Which raises the last question, and the sharpest: what happens to all of this when we take the carbohydrate out?

---

Movement V

WHEN THE CARBOHYDRATE DISAPPEARS

What actually leaves the bowl when we decide the carbohydrate has to go? It is a fashionable decision, sometimes a sound one, and it is worth asking the question plainly, because the easy answer is wrong. The temptation is to picture subtraction: the same food, minus a nutrient someone judged unnecessary, revealed at last as what it should always have been. That is not what happens, and part of the reason is that there is no single thing called removing the carbohydrate. Take out a grain, a starch, the sugars, the fibre or the gums and you have performed five different operations, each with its own consequences. Which jobs vanish depends entirely on what was removed, and what, if anything, was put in its place.

Remove the wrong one and several things move at once. The structure that gave the kibble its crunch may need another agent to stand in, or the piece fractures differently. The water a starch was holding has to be managed by something else. If the removal also takes the reducing sugars that fed the browning, the smell that cooking makes is a different smell, built now from whatever remains to react. And the surface the coating sat on may have changed beneath it. The animal that meets the reformulated food is not tasting the absence of a nutrient. It is meeting a new object, and it will judge that object on its own terms, not against the ghost of the food it replaced.

Take the carbohydrate out and you have not found the true food underneath. You have made a different one.

Which brings us to the familiar grain-free food. Grain-free is not carbohydrate-free. Many grain-free dry foods simply replace the cereal with a pulse or a tuber, the potato, the pea, the lentil and their kind, so the carbohydrate stays and most of its palatability jobs stay with it; what changes is the source, and the new source brings its own baggage, a different flavour to account for, a different behaviour in the process, questions of its own that a later Conversation will take up. Removing the grain did not remove the job the carbohydrate was quietly doing. It handed the job to a different plant.

So we return, at the end, to the question the whole essay was shaped around. If an animal prefers a food made with carbohydrate, what has it preferred? The taste of the sugar, in the animal equipped to taste it? The energy that followed the meal, and the body's slow accounting of it? The texture the starch helped create, the crunch or the chew or the set of the gravy? Or the aroma that cooking, with the carbohydrate's smaller relatives, made possible? The least tidy answer is the true one. Several of them at once, in proportions that shift with the species, the format and the food, and we should distrust anyone who says confidently which.

The carbohydrate may never be tasted as carbohydrate. Yet take the right one away and the food may not taste, smell or feel the same. That is the paradox we opened with, and it is reason enough to retire a lazy word. To call an ingredient a filler tells us almost nothing about what it does in the food. Some carbohydrate is indeed there for bulk or for cost. Some of it is holding up the structure, managing the water, feeding the aroma and carrying the coat. The word cannot tell the two apart. The food can.

References

1.  Li, X., Li, W., Wang, H., Cao, J., Maehashi, K., Huang, L., Bachmanov, A.A., Reed, D.R., Legrand-Defretin, V., Beauchamp, G.K. & Brand, J.G. (2005). Pseudogenization of a sweet-receptor gene accounts for cats' indifference toward sugar. PLoS Genetics 1(1):e3. doi:10.1371/journal.pgen.0010003

2.  Hewson-Hughes, A.K., Hewson-Hughes, V.L., Miller, A.T., Hall, S.R., Simpson, S.J. & Raubenheimer, D. (2011). Geometric analysis of macronutrient selection in the adult domestic cat, Felis catus. Journal of Experimental Biology 214(6):1039-1051. doi:10.1242/jeb.049429

3.  Hewson-Hughes, A.K., Hewson-Hughes, V.L., Colyer, A., Miller, A.T., McGrane, S.J., Hall, S.R., Butterwick, R.F., Simpson, S.J. & Raubenheimer, D. (2013). Geometric analysis of macronutrient selection in breeds of the domestic dog, Canis lupus familiaris. Behavioral Ecology 24(1):293-304. doi:10.1093/beheco/ars168

4.  Tran, Q.D., Hendriks, W.H. & van der Poel, A.F.B. (2008). Effects of extrusion processing on nutrients in dry pet food. Journal of the Science of Food and Agriculture 88(9):1487-1493. doi:10.1002/jsfa.3247

5.  Slade, L. & Levine, H. (1991). Beyond water activity: recent advances based on an alternative approach to the assessment of food quality and safety. Critical Reviews in Food Science and Nutrition 30(2-3):115-360. doi:10.1080/10408399109527543

6.  Dainton, A.N., Doğan, H. & Aldrich, C.G. (2021). The effects of select hydrocolloids on the processing of pâté-style canned pet food. Foods 10(10):2506. doi:10.3390/foods10102506

7.  de Godoy, M.R.C., Kerr, K.R. & Fahey, G.C. Jr. (2013). Alternative dietary fiber sources in companion animal nutrition. Nutrients 5(8):3099-3117. doi:10.3390/nu5083099

8.  Guichard, E. (2002). Interactions between flavor compounds and food ingredients and their influence on flavor perception. Food Reviews International 18(1):49-70. doi:10.1081/FRI-120003417

---

About the Author

Dr. rer. nat. habil. Dr. Seronei Chelulei Cheison is the founder and Chief Executive Officer of Sinonin Biotech GmbH, where he works with petfood companies on palatant sourcing strategies, the innovation of palatability enhancers, and the design of palatability assessment. A trained food enzymologist, he led alternative protein and palatant development at Mars Petcare before founding Sinonin, and his research on protein chemistry spans two decades, a doctorate from Jiangnan University, and a habilitation from the Technical University of Munich.

---

Sinonin Biotech GmbH is a partner in the ZEST project (Grant Agreement No. 101157382) and the PROSCALE project (Grant Agreement No. 101288362), both funded by the Circular Bio-based Europe Joint Undertaking (CBE JU) under the European Union's Horizon Europe research and innovation programme. Views and opinions expressed are those of the author only and do not necessarily reflect those of the European Union or CBE JU. Neither the European Union nor the granting authority can be held responsible for them.