A Litre of Beer Is Worth USD 1.09 and a Litre of Champagne USD 30.87

Twenty-three markets, one segment, one ladder — what the Beverages cluster says when you stop ranking it by growth rate.

The same 750 millilitres, 28 times apart

Beer sold in 2025 was worth USD 1.09 a litre. Champagne sold in 2025 was worth USD 30.87 a litre. That is 28.3 times as much money for the same quantity of liquid, in the same segment, measured in the same year.

Both are fermented from sugar. Both are roughly nine-tenths water. Both are carbonated, both sit between 5 and 12 per cent ethanol, and no laboratory would tell you the champagne has the better molecule. The difference is everything wrapped around the litre: eight years in a chalk cellar instead of eight days in a tank, a delimited place instead of a recipe, glass that holds six atmospheres instead of one, a duty treatment, a distribution system that cannot be short-circuited, and a name that a grower is legally forbidden to put on a bottle they did not make.

Strip the chain out of both and the ratio narrows. It does not close. Between brewery gate and retail shelf sit four different systems that each add cost in a different currency — time, place, glass and law — and no beverage market in this cluster can be read without knowing which of them is doing the work.

That is the argument of this segment. Not that beverages grow, but that every market in it prices a litre from a different rung of the chain, and the rung explains more about the number than the forecast does.

Read down the column and the segment stops looking like a category and starts looking like a price ladder. Gin is a bargain at 14.27 because it is the cheapest liquid in premium-plus spirits — that cheapness is the growth engine, not a problem. Wine at 25.40 costs more per litre than sparkling wine at 21.38, which is only possible because still wine carries the ageing and the appellation while sparkling carries the pressure. Champagne carries both and lands at 30.87. And matcha, which is not a drink at all until somebody adds water, sits at the top of the table per kilogram of the thing that is actually traded — the only honest unit for a product sold as powder.

The ladder is not a ranking. It is a map of what each market is actually charging for.

Why every growth rate in this cluster is two numbers

A single percentage hides the only question a beverage investor asks: did it sell more, or did it charge more? Where a market states both a value and a physical volume, that question has an arithmetic answer — and across this segment the answers split the cluster in two.

The unit-value column is the residual: value growth divided through by volume growth. It captures price and mix together, which is why it is labelled unit value rather than price — a trade-up and a price rise look identical in this column, and telling them apart takes the customer-level work described below.

The fermented and distilled half of this segment is a shrinkage business that grows. Beer loses half a per cent of its litres a year and adds 3.18 per cent of value; wine loses 1.19 per cent of its litres and adds 3.81. Neither is selling more drink. Both are selling a more expensive drink to the same or fewer people, and the reports say so with unusual directness — the wine market grows because the bottle got dearer, and rosé grows 4.8 per cent a year while its volume stands still.

The functional half is the opposite business entirely. No-alcohol beer grows 7 per cent of volume for 1.29 per cent of unit value. Beverage-grade ethanol grows 2.49 per cent of volume for a 1.58 per cent price pass-through that is mostly duty. These markets are buying distribution, not pricing power, and their value growth is largely a count of new doors.

Two markets at nearly the same headline rate — no-alcohol beer at 8.40, matcha at 8.35 — sit on opposite sides of that line. No-alcohol beer raises what a litre fetches while it adds litres. Matcha adds kilos faster than it adds money, which is another way of saying the marginal tonne is cheaper than the average one: the report attributes it to Chinese supply entering a Japanese-defined category, and the three numbers behind the 170 per cent green-tea price move is the same mechanism seen from the other end.

The segment’s reports already refuse to let that distinction slide. Brussels has stopped calling it a market and the EU created a wine category nobody can size are both about definitions that determine what counts as volume in the first place. Ten countries make almost all the world’s rosé is about how thin the supply base under a price is. Every one of them is a volume argument wearing a value headline.

The one number that decides the largest forecast in the segment

Wine is the clearest case in the cluster, because it is the one where the arithmetic and the mechanism are visible at the same time.

USD 528.4 billion in 2025 on 208.0 million hectolitres. USD 712.6 billion in 2033 on 189.0 million hectolitres. Price per litre goes from USD 25.40 to USD 37.70, a 5.06 per cent annual increase, and it has to hold for eight consecutive years while the litre count falls 1.19 per cent a year. The forecast is not a demand forecast. It is a forecast that the trade-up never stops, in a market where fourteen per cent of the base is growing 1.4 times as fast and the rest is flat or declining.

Champagne shows what happens when the trade-up breaks. It shipped 266.1 million bottles for EUR 5.7 billion in 2025 — revenue down 12.31 per cent against a 1.95 per cent fall in volume, so revenue per bottle fell from EUR 23.95 to EUR 21.42, down 10.56 per cent, in the most protected appellation in the world. The segment published that the day it happened (the most protected wine appellation in the world just cut its price 10.6 per cent), and then published the consequence: the 2033 forecast now rests on recovering the price it just lost (Champagne’s 2033 forecast rests on recovering the price it just lost). At 1.57 per cent a year, USD 30.87 becomes USD 34.95 and no more — the forecast is a repair job, not a growth story.

Beer’s 3.70 per cent unit-value climb looks less fragile for a structural reason rather than a demand one: two brewers make 37.9 per cent of the world’s litres. Price leadership in a market that concentrated does not depend on the consumer feeling generous; it depends on nobody breaking ranks. And whisky shows the same mechanism priced by geography rather than by firm — the United States pays 1.97 times the world average per bottle and France pays 0.67, for the same traded category.

That last figure is worth pausing on, because no dataset contains it. It comes from matching customs unit values against realised price per bottle collected from the trade. It is the kind of number this cluster is built on.

How the cluster is put together

Twenty-three published reports and forty-nine published articles sit in this segment. They are not twenty-three parallel forecasts. They are four spines, and each spine answers a different question.

The spines cross, and that is where the useful work happens. Spirits passes wine in 2031 on a value basis (the report’s own call) while wine sells three times the litres. RTD alcoholic grows twice as fast as wine on a twelfth of the revenue (RTD grows twice as fast as wine on a twelfth of the revenue) — and gin’s real competitor is not vodka but the 5 per cent band (gin’s real competitor is the 5 per cent band), which is an RTD argument landing inside a spirits report.

Three cross-spine findings are worth travelling for. Packaging is a beverage cost before it is a beverage decision: wine buys EUR 5.1 billion of glass a year from an industry that just started shrinking (wine buys EUR 5.1 billion of glass a year), the heavy bottle is worth about eight cents, and above 693 grams a bottle starts consuming container space instead of wine. Tax is a route-to-market decision: when the tax is on the bottle, the bottle stops travelling, and two-thirds of America’s wine bottles cross the Pacific before the wine does. Supply is a lagging storyteller: blue agave fell 66 per cent and the boom didn’t notice, because the 2032 crop is being decided today at MXN 7.62 (the 2032 agave crop is being decided at MXN 7.62).

How these numbers were built

Every figure in this cluster has two parents, and neither is a forecast.

Secondary work sets the perimeter. Customs unit values on the shipment lines that actually carry the trade — beer, wine, spirits, tea — give a price per unit of the good that crosses a border. Excise and duty schedules give the tax wedge, which for ethanol and for spirits is larger than any margin in the chain. Appellation and GI rules set what may be called what, and therefore which litres are allowed into which series. Official production and shipment statistics — harvest declarations, denominación-registry volumes, trade-association shipment counts — set the volume line. Listed-company filings give realised revenue per case for the firms large enough to publish one, and the two-brewer concentration figure is a filing number, not an estimate.

Primary work explains why the price sits where it does. Three instruments, run per market, every cycle:

• Structured interviews with producers, importers, distributors, on-trade buyers and the suppliers who shape cost — glassmakers, coopers, can makers, concentrate suppliers. Unstructured enough to let somebody explain a mechanism, structured enough that the same questions are asked in the same order in every market, which is what makes a Rioja answer comparable to a Marlborough one.
• A closed-ended questionnaire put to a standing panel of trade respondents, where every question asks for one answer from a fixed set of bands — a price band per case, a volume band per SKU, a share band per channel. Nobody is asked for a growth rate. Closed bands are the whole point: they aggregate, they can be re-asked unchanged next cycle so movement is measurable rather than reported, and they cannot be talked up.
• A price audit — realised price per case, captured at a stated layer of the chain rather than asked for in the abstract. This is where USD 14.27 a litre for gin comes from, and why that figure is a realised price rather than a list price nobody pays.

The two halves are then set against each other. If a respondent’s realised price per case and the customs unit value for the same product disagree by more than one band, the number is reopened rather than averaged. The whisky multiple — 1.97 times the world average in the United States, 0.67 in France — survives precisely because it was produced by that comparison and by two independent instruments agreeing on it.

Two limits are stated in every report rather than hidden. A closed-ended instrument cannot surface a category that nobody on the panel sells, which is why new formats enter each cycle through the interviews first. And where a product is traded by weight — tea, matcha, anything sold as leaf or powder — the kilogram is the unit and a litre is a property of somebody’s kettle, so the litre column above stops where the trade stops.

What the ladder is for

The point of building twenty-three markets on one unit is that it makes the segment falsifiable.

Three calls are visible in the numbers above. Wine at USD 37.70 a litre in 2033 fails if the trade-up pauses for two years — that is 5.06 per cent of unit value every year for eight years on a shrinking litre base, and the 2025 champagne price cut shows how fast an appellation can lose 10 per cent of its price when demand turns. Gin’s 2.08 per cent is the smallest unit-value ask in the cluster and the one most exposed to the 5 per cent band taking its drinker. And no-alcohol beer at 7 per cent volume is the only market here whose growth is a distribution count, which makes it the one where a single large brewer’s listing decision moves the series.

Matcha is the warning. It grows faster than anything in the cluster except the two largest functional categories, it carries the highest price per traded unit in the table, and its unit value falls every year of the forecast. Growth and value are not the same variable, and in one market out of eleven they point in opposite directions.

That is why this segment is worth reading as a cluster rather than as twenty-three files. The rung a litre sits on explains its price. The volume line explains whether the growth is real. Everything else — the brand, the appellation, the glass, the tax — is the distance between those two numbers.

Pheonix Research publishes twenty-three market reports and forty-nine companion articles across the Beverages segment, all built on the same demand model, the same price-layer discipline and the same primary panel.