Nuts
How the almond took the blame for California's water — while a thirstier crop drank in peace, a bigger one escaped notice entirely, and in the end it wasn't guilt that moved the water at all
Second in a series on how debates about scarce resources get told as stories about personal guilt.
You have met this number too, and it is older than the bottle: It takes about a gallon of water to grow a single almond.
For a few years in the middle of the last decade, as California ground through drought, that number was everywhere, and it did the same work the bottle would later do for the chatbot: it took something small and repeatable and made you feel the hidden cost of it in your hand. A public-radio piece at the time called it fighting words, and it was — the almond became the face of a state's thirst, the crop you could point at, the thing on the shelf you could put back.
Why invoke the almond now? The argument that formed around it a decade ago was full dress rehearsal of the water panic we are having about artificial intelligence today. Every part of the mechanism the first essay found in that AI-and-water article — the memorable number that its own carriers half-disown, the strong argument buried under the weak one, the accounting that shifts with the question — was already assembled, in public, around a nut, ten years before anyone worried about the water cost of an email. If you want to know how the current panic ends, it helps to read the one that already ran.
Let's consider our specimen: a 2015 public-radio segment on how almonds became a scapegoat. Nearly the whole machine is in it.
Watch the crop's most committed critic, a water-policy advocate, build the case: Almonds, he notes, drink more than all of California's indoor water use; more than all its residential landscaping. It is a real indictment, and then, in almost the same breath, he concedes the thing that dissolves it: almonds are not the state's largest agricultural water user. Alfalfa is. He knows this. He says it. He goes on aiming at the almond anyway. That is the first and deepest tell of the whole genre — the target is chosen by how visible it is, not by how much it drinks. The biggest user is not the one that gets named. The most nameable one is.
The rest of the parts are there too. The industry answers the drought criticism with an efficiency number: California growers, the Almond Board reports, cut the water used per pound of almonds by roughly a third between the 1990s and the 2010s. True enough, and beside the point the critics immediately make in return — that water saved per pound does nothing for the total when you keep planting orchards. Efficiency per unit fell; total water climbed, because acreage climbed faster. Take a moment to consider this: the efficiency did not cause the expansion. Prices did, and export demand, and land values, and the returns on a perennial crop. But the pattern — per-unit thrift swallowed whole by growth — is the one this series will meet again in a data center, and it is already fully formed here in the dirt. And there is the reassurance machine, funded and named: the Almond Board, a marketing organization, calling the criticism misinformation and bad facts. It is the same move a technology company will later make when it publishes its own flattering water study, only wearing overalls.
The line that compresses all of it comes, as it often does, from the one participant with nothing to sell — a water scientist at UC Davis, who observed that "waste" is always water used by other people. Everyone's villain is someone else's allocation. And he added the detail that explains the timing: without the drought, nobody would have been talking about almonds at all. The panic did not go looking for a villain until a crisis handed it a reason — exactly as the water-and-AI panic waited for the boom. The machine idles until an event switches it on, and then it aims at the most legible thing in range.
But of all the crops the machine could have reached for, why this one? The answer starts with a fact that should have ended the whole argument and almost never got mentioned: the almond is not even the nut Americans eat most.
By the U.S. Department of Agriculture's own consumption figures, Americans eat roughly 7.5 pounds of peanuts a head each year and about 2.36 pounds of almonds. The peanut outweighs the almond about three to one — and outweighs every tree nut combined. When headlines around 2012 crowned the almond America's number-one nut, they meant its number-one tree nut, on a per-capita ranking it had only just seized on a wave of superfood marketing; the peanut remained, quietly, the far larger staple. The crown was a product of attention, not of appetite. (It is worth admitting, in an essay called Nuts!, that the almond is botanically the seed of a drupe and the peanut is a legume, and neither is a nut at all. The category we are arguing about is a folk fiction — a fair early warning about how much the label was ever tracking.)
Now set the two crops side by side and the puzzle sharpens instead of resolving. The peanut is grown mostly rain-fed in the Southeast. By the National Peanut Board's accounting it carries the smallest water and carbon footprint of any nut, a fraction of what the tree nuts require. It is cheaper. It is the single most-requested item at American food banks, and peanut paste is the frontline treatment for severe childhood malnutrition. So here is a crop that is bigger by volume, greener by the drop, and more defensible by almost any humane measure you choose — and it drew no fire whatsoever. Not one viral number. Not one guilt headline.
Why the almond and not the peanut, when the peanut is larger, and thirstier only in the imagination? Because consumer guilt does not track magnitude, and it does not track harm. It tracks prior virtue. You can only be billed for the pride you were first sold. The almond, in its milk-and-butter form, was a virtue product before it was a guilt object — plant-based, dairy-free, wellness-coded, the very drink people had switched to in order to feel better about what they consumed. That earlier pride is the entire mechanism: the charge lands because it turns the mirror back on the buyer's own signaling. The peanut was never a virtue signal — it is a cheap, unfashionable, working-class staple — so there was nothing to reverse, no pride to convert into shame. The guilt reached the wellness aisle and never the food bank, because the food bank had never been invited to feel superior in the first place.
That points at something the acre-feet cannot explain on their own, and which a later essay will have to take up properly: the almond was doing identity work. A guilt totem has to be legible, shareable, and refusable — a badge others can watch you set down — and its value to a group lies in sorting the members and ranking them, not in any effect on a watershed. This is not a habit peculiar to one political tribe, and it would be dishonest to pretend it is. The viral almond takedown did originate in the progressive press — it was launched in Mother Jones, aimed by name at almond milk and the people who drank it, in-group speaking to in-group — but the machinery is about groups, not ideologies. The other side of the aisle runs the identical play with different totems: buy the domestic brand, refuse the foreign one, prove your loyalty at the register. The crop is a coincidence. The sorting is the point. For now it is enough to notice that a debate ostensibly about water was, in part, a debate about who belonged — and to carry that question forward.
Return to the water, because the crop's own numbers hold a second surprise, and it is one the guilt articles tend to print against themselves. A sustainability blog runs the almond-milk question in the plainest possible form — can we drink it without guilt? — and then, a few paragraphs down, prints the figures that answer it. Draw on the same widely-cited source it does, a Pacific Institute analysis built on 2010 California water data in 2014 dollars, and measure crops the way an economist would, by value produced per unit of water. Almonds and pistachios together returned roughly $1,150 for every acre-foot of applied water. Alfalfa — the state's single largest agricultural water user — returned about $175. On the exact denominator that asks how much value a drop of water buys, the almond is among the most productive things in the valley, and the crop nobody was guilting is nearly the least. The article leads with the guilt and files the exoneration in its own statistics, and never notices the contradiction, because the denominator that would have surfaced it was not the one anyone reached for.
The consumer product that actually carried the shame was almond milk, and the fight was staged as almond versus oat. On the narrow axis that fight chose — water per liter of finished drink — oat honestly wins; published lifecycle figures put oat milk well below almond, which sits well below dairy. So "oat is the greener carton" was not wrong on its own terms, and it would be dishonest to pretend otherwise. But look at what the almond-versus-oat frame quietly hid: the gap between any plant milk and cow's milk dwarfs the gap between two plant milks. Anyone who had already left dairy had captured almost all of the water savings available; the further switch from almond to oat was the last thin sliver, sold as though it were the moral hinge. And the "buy organic" counsel that so often rode alongside is quietly incoherent with the stated worry, since organic almonds use as much water or more. The measurement was not doing the persuading. The choice of denominator and the choice of boundary were doing the persuading, and the number was the costume they wore.
So follow the water where the crops' own figures point — past the almond, to the thing at the top of the table.
Alfalfa, in that same 2010 accounting, took about 5.2 million acre-feet of applied water, the largest single crop category in the state; almonds and pistachios together took roughly 3.8 million. Alfalfa is forage — hay. No human eats it. It is grown to be converted: into milk, into beef, and from there into a patty, and from there into a branded meal sold under a lit sign for a price no raw field crop will ever command. Each of those conversions stacks another markup on the same original gallons, and — this is the part that matters — abstracts the water one more step from view. By the time the Colorado River reaches you it is invisible, dissolved inside a six-dollar combo that looks like it never touched a river in its life.
Which yields the rule this essay is really about. Guilt is inversely proportional to laundering. The almond wears its water on its sleeve — one nut, one gallon, one bag you can decline at the store — and so it absorbs the shame. The burger hides its water behind four transformations and a continental supply chain, and so it walks free, even though the alfalfa-to-beef path is both the larger draw and, by that value-per-drop measure, among the least efficient uses in the state. Legibility is the tax. Opacity is the shield. The single most water-intensive thing on the tray is the one nobody ever handed you a number for.
And here is the image this series will keep coming back to, because it is the whole argument in one frame.
In that decade-old segment, an almond grower stands beside his drought-stressed trees while federal water runs past him in a canal he is legally forbidden to touch — it belongs, by senior right, to users hundreds of miles to the south. He can see the water. It is not his.
Move ten years forward and one state east, and the same picture returns with harder edges. In rural La Paz County, Arizona, where groundwater goes essentially unregulated, residents watch the Central Arizona Project canal — Colorado River water — run past while their own household wells drop and fail, because the canal's water is spoken for and theirs is being drawn down from beneath them. Some of that pumping feeds a Saudi dairy company's subsidiary, growing alfalfa on state-leased land for a reported twenty-five dollars an acre with essentially unlimited access to the aquifer, much of the hay shipped home to feed cattle in a country that restricted its own forage farming to protect its own water. The state has since moved to cancel the leases; the attorney general has sued. You can see the water. It is not yours.
But watch the machine run one level down, because this is where fairness demands attention. The foreign farm became the villain — outrageous, photogenic, a foreign straw in American ground — and the water experts on the scene kept trying to turn the camera. The problem, one of them said, is not who is doing it but that it is allowed at all. The largest groundwater pumpers in the state are domestic — a Minnesota-based mega-dairy, a home-grown nut operation — and in at least one county the biggest of them stood to be grandfathered in under the very pumping limits the residents had voted for. Nationally, export is a minority channel to begin with: the United States shipped about 2.1 million metric tons of alfalfa hay in 2024, on the order of four percent of production, and most western alfalfa still feeds American cows. So the export farm is the elected villain, and the larger domestic thirst beside it — closer, blander, unbranded — drinks in relative peace. The almond's fate, visited on the export bogeyman, on behalf of everyone pumping quietly next door.
Two temptations have to be refused right here, because surrendering to either is just the almond's mistake pointed in a new direction.
The first is to crown the burger — or beef itself — as the real monster. But beef's water, like every quantity in this series, is a question of where and which, never a bare total. A great deal of a cow's water is green: rain falling on grazing land that grows nothing else and irrigates nothing, arguably being used rather than wasted. The blue, consumptive, stressed-basin fraction — irrigated desert alfalfa pulled from an over-allocated river — is the defensible target; a cow eating rain on unfarmable range is close to a non-issue. "Beef is the hidden villain" is the almond move in a fresh costume, and it fails for the same reason.
The second is to sneer at the cheap meal. Run the honest denominators and that drive-thru combo is, for a family eating on a fixed budget, a rational purchase: close to unbeatable on calories per dollar, engineered to be unbeatable on pleasure per dollar. Alfalfa's dismal value-per-drop looks pathetic beside the almond's only if market price is your measure of worth — but underwrite affordable protein and calories for millions of people and "low value" begins to look like an artifact of the denominator you happened to choose. Whether the better use of a scarce gallon is the luxury export nut or the cheap-calorie forage is not, at bottom, an efficiency question. It is a distributional one — and the guilt framing exists, in part, to keep that question from being asked out loud.
So, in the end, who moved California's almond water?
Not the guilt. A decade of can we drink it without guilt, of one-gallon-per-almond, of switch-to-oat — and the thing that actually brought the state's almond water use down was none of it. It was a law and a price. Under the Sustainable Groundwater Management Act, overdrafted basins were pushed, district by district, toward balance; the cost of water in some places climbed past a thousand dollars an acre, the cost of capital roughly doubled, and the arithmetic that had made orchards the king of the valley quietly stopped working. California's almond acreage peaked at around 1.65 million acres in 2021 and has fallen every year since, to roughly 1.5 million, with new plantings at a multi-year low and something like 210,000 acres pulled out of the ground between 2022 and 2024. The orchards that came out came out because the water got allocated and priced — in the statute and at the headgate, not at the grocery shelf.
Hold the two side by side, because the comparison is the whole point. The discourse announced its purpose plainly: to save California's water. Run it for ten years and measure what it saved, and the answer is nothing you can find in an acre-foot. The law ran for three and the acreage bent. If you judge the discourse not by what it said it was for but by what it reliably produced — a decade of aimed consumer guilt, a viral number, a sorting of who drank what, and the steady absence of the water district from the conversation — you are left with an uncomfortable conclusion: saving water was the announced purpose, and those other things were the working ones. Not because anyone schemed to make it so. Because that is what the machine, left running, does. A system's purpose is better read from its output than from its mission statement, and the output here was guilt, membership, and a decade of looking the wrong way.
And the quietest party kept its silence the whole time. The aquifer beneath that Arizona county holds water twenty thousand years old — fossil water that no rain will refill on any human schedule; press the sponge hard enough and it stays pressed. It has no marketing budget. The family whose well ran dry paid to drill a deeper one and appears in no one's per-almond figure. They held the largest stake in the argument and the smallest voice in it — no vote in it, no price carrying their need, no seat at the table, nothing that would have registered if it had changed — and the number printed on the bag was never, at any point, about them. The debate would have run exactly the same way with them in the room or out of it, which tells you, more plainly than any figure, which it was built for.
The trail from here leaves the soil. The next thing we are being taught to feel guilty about does not grow — it computes. The same machinery that made a scapegoat of the almond and let the alfalfa drink in peace is running right now around the data center, with the orchard swapped for the server and the water bottle swapped for the kilowatt-hour, and not one of its moving parts has changed. Follow the water up out of the ground, and it runs straight into the grid.