Segment 1 — The Cold Open

A half-liter bottle of water at a gas-station cooler commonly sells for something like a dollar and a half, sometimes more. Wave a magic wand, arrange the raw atoms — municipal water, a dozen grams of PET, a polypropylene cap — and the materials inside that bottle are worth a couple of cents. The liquid is not scarce. The polymer is not scarce. What you are buying is a sealed, branded, refrigerated object that has been driven across a landscape already plumbed with safe water, and the gap between those two cents and that dollar is the entire industry.

Segment 2 — Why It Costs What It Costs Today

Bottled water did not start as a way to move municipal supply in disposable plastic. The nineteenth- and twentieth-century European springs — Evian, Perrier, San Pellegrino — sold mineral content, a place, and a status object, often in glass, to people who already had water at home. That luxury template was still visible when the American market exploded in the 1970s and 1980s. Then the category flipped. In the 1990s the big beverage companies walked in with a different product: city water, run through carbon filters, reverse osmosis, and ultraviolet or ozone, poured into PET, and named. Aquafina and Dasani are the famous cases, and they are not unique; a large share of the “purified” water on North American shelves is municipal supply with a treatment step and a label. The industry learned that you do not need a glacier. You need a plant, a bottle, a truck, and a story about purity.

Once that template existed, every incentive in consumer packaged goods piled onto it. Beverage companies already had filling lines, cold-drink distribution, and supermarket relationships; water was a volume SKU that did not compete with their own soda as harshly as a rival cola would, and it rode the same trucks. Retailers learned that water is an impulse item with reliable turns, especially at convenience and in the checkout cooler. Brands learned that mountain photography and words like “alpine,” “artesian,” and “pure” sell even when the intake pipe is in the same town as the customer. Advertising, slotting fees, and trade spend accreted on top of manufacturing the way they do on every other branded grocery product. None of this required a conspiracy. It required the ordinary machinery of CPG, applied to a liquid that costs almost nothing.

The price then feels normal because the object in your hand is no longer “water.” It is a cold, sealed, portable serving with a trusted name, sitting where you already stopped for fuel. Tap water requires a glass, a fountain that works, and a belief that the pipe is safe. In cities that have had lead, boil-water notices, or simply old buildings with ugly fountains, that belief is not free. Airports, stadiums, and offices stripped out fountains or let them decay, then rented coolers and sold cases. Parents packed bottles into lunch bags until the next generation treated the single-serve PET as the default unit of hydration. By the time Americans were drinking more bottled water than soda — tens of gallons per person per year, in the commonly reported tallies — the category no longer had to justify itself against the tap. It only had to justify itself against the other bottles on the same shelf.

Regulation quietly helped the feeling of specialness. In the United States, tap water is an EPA matter and bottled water is an FDA food. The two regimes test for overlapping but not identical things, on different schedules, with different public reporting. That split is often sold, implicitly, as “the bottle is more tightly controlled.” It is not a reliable safety ranking; it is a jurisdictional accident that marketing was happy to wear. Add the fact that chlorine in city water is noticeable and reverse-osmosis filtrate tastes “clean,” and you have a sensory alibi for a thousand-percent markup. Inside the industry, a dollar for a half liter is not a scandal. It is the working retail price of a convenience beverage, and convenience beverages are supposed to be expensive.

Segment 3 — The Magic Wand Number & The Idiot Index

Start with the actual atoms in the most common object: a half-liter single-serve bottle. The water is 500 milliliters of something a city already treated and piped. Residential volumetric water charges in much of North America, once you stop mixing them with sewer and fixed meter fees, often land around a few dollars per thousand gallons. A thousand gallons is 3,785 liters, so a few dollars spread over that volume is a small fraction of a cent per liter — commonly in the neighborhood of a tenth of a cent, sometimes less, sometimes a few tenths depending on the utility. Half a liter is then a few hundredths of a cent. It is fair to call the water, as a commodity, rounding error.

Purifying it again does not rescue the price. Reverse osmosis at industrial scale is an energy-and-membrane problem. Efficient systems often use something like one to a few kilowatt-hours per cubic meter, which is a thousandth of a kilowatt-hour per liter or a little more. At ordinary industrial electricity prices that is another tiny fraction of a cent. Carbon, ultraviolet lamps, and ozone at the throughput of a bottling plant are not free, but they do not add cents per liter either; they add fractions. The magic-wand water-plus-treatment number, the cost of arranging H2O into a drinkable, low-TDS liquid, is still a fraction of a cent per half-liter serving. Anyone who has run a home undersink RO system and divided the filter-and-electricity bill by the liters produced has seen a version of the same arithmetic.

The bottle is the first material that actually shows up. Bottle-grade PET resin is a commodity that moves with oil and paraxylene; across recent years it has typically traded in a band around one to two dollars per kilogram, sometimes spiking outside that, often sitting inside it. Years of lightweighting have brought many half-liter bottles down to roughly eight to twelve grams for the body. Ten grams is a hundredth of a kilogram. At a dollar fifty a kilo, that body is about a cent and a half of polymer. Use twelve grams and two dollars a kilo and you are still around two and a half cents of PET. The cap is a couple of grams of polypropylene or HDPE — another few tenths of a cent of resin. The label is a sliver of film and ink. Glue, ink, and the dust of process chemicals do not change the column. Add it up out loud: water and re-treatment, negligible; PET body, about one and a half to two and a half cents; cap, a few tenths; label, rounding error. A rough magic-wand estimate for the finished half-liter object is two to four cents of raw material, and it is honest to lean toward the low end of that range for a modern lightweight bottle at ordinary resin prices. A one-liter bottle uses more polymer — often something like twenty grams, give or take, after lightweighting — so the floor rises, but it rises into the low single cents, not into the dollars.

Now the Idiot Index, kept visible as arithmetic rather than as a fact of nature. Take two and a half cents as a generous material floor for the half-liter. A warehouse-club case can put that same bottle on a pantry shelf for roughly fifteen to twenty-five cents; divide twenty cents by two and a half cents and you get a ratio around eight. That is already a loud number for a product whose physics are this simple, but it is the cheap end. Grocery singles sit higher. A convenience-store or airport cooler commonly asks a dollar and a half to two and a half dollars for the identical 500 milliliters. A dollar fifty divided by two and a half cents is sixty. Two dollars divided by two cents is a hundred. Premium spring and imported still waters — the Fiji and Evian shelf — can sit at two to four dollars a liter, which is a still larger multiple, though those at least sometimes contain dissolved minerals you cannot get from the kitchen tap. None of these ratios is a laboratory measurement of anyone’s cost of goods. They are a back-of-the-envelope comparison of public retail prices to public commodity prices, and they are enough to show the shape of the opportunity: even the “cheap” case of bottled water is several times the atoms, and the bottle you actually grab when you are thirsty is tens of times the atoms, sometimes a hundred.

Where does a factor of sixty live, if it does not live in the PET? Walk the chain. Pellets become preforms become stretch-blown bottles. That conversion is one of the most optimized pieces of plastics processing on earth; high-speed blow molders do not turn two cents of resin into a fifty-cent empty bottle. They turn it into an empty bottle that still costs cents — energy, machine time, scrap, labor, the mold — a small multiple of the resin, not two orders of magnitude. Filling, capping, and labeling on a rotary line that runs hundreds of bottles a minute likewise cannot explain a dollar. Capital equipment is expensive, but it is amortized over enormous counts; the labor and utilities per unit at a modern plant are small. Secondary packaging — the shrink wrap, the corrugated tray, the pallet — adds more cents. You can spend a generous ten or fifteen cents of conversion, fill, pack, and plant overhead and still not have accounted for the convenience-store price. The materials-plus-plant story tops out well before the register.

The rest is movement, rent, and narrative. Water weighs one kilogram per liter. A trailer that can legally haul about twenty tonnes is hauling about twenty thousand liters if it is full of filled bottles, and a non-trivial share of that mass is PET and corrugated rather than water. Diesel, driver, tractor, insurance, and the empty backhaul are being spent to relocate a liquid that the destination city already has in its mains. Regional distribution might add cents per bottle rather than dollars — exact freight allocations are not public and vary wildly with distance — but the direction is insane in principle: the industry’s core logistics task is to move water to places that have water. Then a distributor takes a margin. Then a retailer takes a larger one. Convenience retail in particular prices to the thirsty person standing in front of a cooler, not to the cost stack; the facing in the vault is expensive floor space, the refrigeration is a real electricity bill, and the category can bear a high markup because the comparison set is other cold drinks, not the tap in the bathroom. Brand advertising and trade spend sit on top of that. A national water brand is buying the feeling that this particular municipal filtrate is the safe one, and that feeling is not a polymer. For the premium tier, add the photography of islands and Alps, the glass-bottle cousin on the restaurant list, and the permission to charge a beverage price for a water price.

There is a genuine remainder that is not waste in the insulting sense. Sealed, portable, cold, and trustworthy is a product. A traveler at a gate, a crew on a job site without potable hose, a household under a boil-water notice, a parent who does not want to wash a bottle at a park — those are real jobs to be done. The Idiot Index does not say those jobs are imaginary. It says that for the majority of liters sold into kitchens, offices, and gym bags in cities with functioning utilities, the jobs are being done with a cost structure built for a luxury import, and the physics of the liquid do not require it.

Segment 4 — The First-Principles Opportunity

A from-scratch redesign does not begin with a slightly lighter bottle. It begins by refusing to long-haul a ubiquitous liquid. The first move is to purify at the point of use or the point of sale and to sell the treatment and the vessel, not the kilometers. Home pitcher filters and undersink reverse-osmosis units already do a version of this; amortized over their rated liters they produce drinking water at a tiny multiple of municipal rates, with the Idiot Index crushed because the trucking, the brand, and the single-use conversion have been deleted. What has to be true for that to take the rest of the market is boring and physical: filters that are cheap and easy to replace, membranes that last, and a public record of tap quality that people can believe without a mountain on the label. Utilities that publish live, readable test data, and buildings that do not leach lead between the meter and the kitchen, are not marketing. They are the enabling infrastructure.

The second move is refill as the default public interface. That means fountains that are cold, fast, and obviously clean — not the warm dribble that taught a generation to walk to the cooler — and it means bottle-filling stations in the same places the industry currently owns: gyms, schools, airports, offices, parks, c-store forecourts. Some of those stations already exist; the first-principles version treats them as the primary product, not as a sustainability afterthought next to a wall of PET. For this to work at scale the empty vessel has to already be in the bag, which is a design problem (a bottle people actually carry) and a habit problem, not a chemistry problem. Deposit-return and refill pricing that makes the packaged liter look as expensive as it is would help, and several jurisdictions have already started down that road with plastic taxes and deposits. The saving is almost the entire gap: you keep the water and the treatment, you discard the intercity freight, the single-use conversion, and most of the brand.

The third move is what remains of true single-serve. Emergency supply, disaster staging, true mobile work, and the last mile of convenience will still want a sealed unit. First principles would fill that unit locally from the same municipal source, in the lightest competent package, with no pretense of a distant spring, and would price it near conversion-plus-logistics instead of near soda. A grocery chain already running a pharmacy and a kitchen could run an RO skid and a small blow-fill line in the back of the store, or — simpler — a staffed or automated refill of a standardized, deposit-bearing bottle. What has to be true is local code that allows it, a quality system as boring as a dairy’s, and a willingness to sell water as water. The dollars that fall out are the national trucking network and the advertising budget.

The hard parts are not the membranes. Trust is hard in cities that have earned distrust, and no amount of clever packaging should paper over lead services or failed plants; those are first-principles water problems of a different kind, and bottled PET is a costly, regressive patch on them. Convenience at 2 a.m. in a place with no fountain is hard, and pretending it is not a product is how refill schemes fail. Taste is hard in the small sense that dechlorinated RO water and chlorinated tap are different, and people are allowed to prefer the former — which is an argument for cheap local treatment, not for a brand from another watershed. The incumbent cooler footprint is hard because it is already paid for and it prints money. Premium mineral water is a different object again; if you want the ions from a specific spring, you are not in the municipal-filtrate business, and that slice of the shelf can stay expensive without confusing the rest. And the environmental accounting is not free even after the Idiot Index falls: PET, collection, and recycling have their own physics, which is why the better redesign is fewer single-use units, not merely cheaper ones.

Segment 5 — The Lesson

If a city already delivers the molecule to the building, a company that still makes its margin by putting that molecule on a truck is not selling hydration — it is selling distance, refrigeration, and reassurance, and those should be priced and designed as themselves. The second principle is older than PET: do not run a logistics network whose payload exists at both ends of the route. The ice trade collapsed when we learned to make cold locally; bottled municipal water is the same cargo, still waiting on the same realization. Tomorrow, another example or another gap. The question for this one is practical: who prices a liter of purified city water against the cents it contains — a utility, a grocer with an RO skid, a refill network — and what is the first SKU that admits, on the label, that the spring is the mains?

💬 Reply to this email — Patrick reads every one.

Share: X · LinkedIn · WhatsApp

Forwarded this email? Subscribe here — it's free.