The Pantry / Grocery Perspectives

Volume didn't
come back. The
reason to buy did.

Americans are not drinking more milk. The category sells the same 4.8 billion units it sold a year ago. But dollars are up, one seven-percent slice is growing five to ten times faster than the rest, and Chobani is spending $1.2 billion to re-enter a business it quit in 2022 after eleven weeks.

A Chobani Ultra-Filtered Milk carton beside the figures 2022 pulled out, 2026 back in, and a $1.2 billion planned dairy investment
$17.7BUS fluid milk, +3% dollars
FlatUnits, at 4.8 billion
+14.8%Ultrafiltered dollars, on 7% of the category
96%Of that segment is one brand

The short version

Fluid milk is a $17.7 billion category growing 3% in dollars on flat volume. That is a mix story, not a demand story. Almost all of the movement sits in one sub-segment: ultrafiltered, high-protein milk, growing 14.8% in dollars and 5.5% in units off roughly 7% of category dollars.

Two things make that interesting rather than merely tidy. The first is that ultrafiltered milk is the only part of the milk case with real volume growth — conventional is flat to declining, plant-based is falling in both dollars and units. The second is that the premium shoppers are paying for it is a per-ounce premium, not a per-gram one. Measured on cost per gram of protein, ultrafiltered dairy is one of the best values in the entire dairy case, and oat and almond milks run five to eighteen times more expensive.

Seven charts below. Then Chobani, which is the clearest test of the thesis anyone is running: the same company, the same product, the same shelf, four years apart — and a $1.2 billion bet that what changed was not the milk but the reason to reach for it.

Chart one / where the growth is

One segment is carrying the whole category.

Plotting dollar growth against unit growth separates real demand from price. Anything above the horizontal line is selling more dollars; only what sits right of the vertical line is selling more milk. Almost nothing does.

Dollar growth versus unit growth, by milk segment

Circana readings, most recently the 52 weeks ending 30 November 2025. The protein-claim measure spans all dairy products, not just milk, and is plotted for scale.

↔ Swipe to see the full chart

Selling less milk Selling more milk 0% +4% +8% +12% +16% 0% −5% +5% +8% All protein-claim dairy +13.7% · +7.5% Ultrafiltered milk +14.8% · +5.5% Lactose-free milk +7.3% · +4.1% Total fluid milk +3% · flat Conventional milk +1.5% · −0.4% Plant-based milk −3% · −5% Horizontal: unit growth · Vertical: dollar growth
Category and brand readings from Circana via eDairy News; segment growth rates from DairyReporter and DairyReporter on protein-claim dairy. Circana's own read is that milk unit sales "stabilised in 2024 and 2025 after a decade long decline" — the category stopped shrinking, it did not start growing.

Chart two / the base

Seven percent of the dollars, most of the momentum.

The growth rate is only half the picture. Ultrafiltered milk compounds at 14.8% off a base that is roughly a fourteenth of the category. The other ninety-three percent grows at 1.5 to 2%. That is the shape of a segment early in its run, not one approaching a ceiling.

Share of US fluid milk dollars, and the growth rate attached to each part

Circana. The source publishes the 7% share but no dollar figure for the segment; against the $17.7 billion category total that implies roughly $1.2–1.24 billion.

↔ Swipe to see the full chart

Ultrafiltered, 7% growing +14.8% Everything else, 93% of dollars growing +1.5% to +2% ≈ $1.2B ≈ $16.5B Share of $17.7 billion in US fluid milk dollars
Share and growth from DairyReporter, citing Circana. The dollar figures either side of the bar are our arithmetic against the $17.7 billion category, not published values — treat them as scale, not as a source.

The concentration problem

Fairlife accounts for 96% of ultrafiltered milk dollars. A segment growing 14.8% that is effectively one brand is not a segment yet — it is a brand with a category-shaped hole next to it. The two fastest movers in the last year were the challengers: private-label ultrafiltered grew 20%, and Darigold FIT more than doubled.

Chart three / brands

Premium brands grew units. Private label did not.

Inside refrigerated white milk, the brands attached to a reason — protein, organic, A2 protein type — grew dollars and volume together. The $9.7 billion private-label block, which is most of the category, grew 1% in dollars while losing volume.

Refrigerated white milk brands: dollar growth and unit growth

52 weeks ending 30 November 2025, Circana. Fairlife's flavored line is shown separately because it is measured separately.

↔ Swipe to see the full chart

a2 Milk $104M Fairlife white $980M Organic Valley $326M Fairlife flavored $276M Private label $9.7B +30% +25% +25% +14% +21% +14% +20% +10% +1% −1% 0% +10% +20% +30%
Dollar growth Unit growth
All figures Circana via eDairy News, 52 weeks ending 30 November 2025. Fairlife is now the number three brand in refrigerated white milk. Note the gap between the paired bars on every premium brand: dollars ran ahead of units everywhere, so even the winners took price.

Chart four / the product

Eight grams is the number the whole category is priced against.

Conventional milk delivers 8 grams of protein per cup and always has. Ultrafiltration concentrates it to 11–15, and the tier that actually carries the segment sits at 13–15 — between 63% and 88% more protein. That difference is the entire physical basis for a segment selling at 2.7 to 4.2 times the price per ounce.

Protein per serving, by product

Grams per one-cup serving, read from each retailer's nutrition panel in September 2026.

↔ Swipe to see the full chart

Alta Dena Protein Darigold FIT ALDI Friendly Farms fairlife OV Protein Plus Silk Protein (soy) Kroger CARBmaster Good & Gather 2% Conventional milk Ripple (pea) Ripple Organic Oatly Elmhurst oat Silk almond 15g 14g 14g 13g 13g 13g 11g 10g 8g 8g 5g 3g 3g 1g the 8g baseline 0g 4g 8g 12g 16g Protein grams per one-cup serving · lime is ultrafiltered dairy, lavender is plant-based
Read from retailer nutrition panels in September 2026; the full 22-product table with a source link for every value is further down this page. "Conventional milk" is the shared reading for Kroger whole, Horizon Organic, Organic Valley whole and Lactaid, all 8g. Silk Protein is the one plant milk that reaches dairy's concentrated tier.

Chart five / the inversion

The expensive milk is the cheap protein.

This is the chart nobody in the category is putting on a shelf tag. Ultrafiltered milk carries a 2.7 to 4.2 times per-ounce premium over a conventional gallon. Measured on what shoppers say they are buying it for — grams of protein — it is among the best value in the case, and every plant milk except Silk Protein is dramatically worse.

Cost per gram of protein

Package price divided by total package protein grams. Prices as displayed on retailer pages on 11 September 2026; grocery pricing is store- and zip-specific.

↔ Swipe to see the full chart

Good & Gather 2% gal Kroger whole gal Alta Dena Protein ALDI Friendly Farms Darigold FIT fairlife 2% Silk Protein (soy) Ripple (pea) Oatly Silk almond Elmhurst oat 1.9¢ 2.7¢ 3.7¢ 4.2¢ 4.8¢ 5.9¢ 6.8¢ 11.4¢ 22.9¢ 36.6¢ 66.6¢ 20¢ 40¢ 60¢ Cents per gram of protein · lime is ultrafiltered dairy, lavender is plant-based, grey is conventional
Computed from package price and package protein grams, each sourced in the table below. Elmhurst oat milk costs eighteen times what Alta Dena's high-protein dairy costs per gram of protein. The cheapest protein in the case is still a plain gallon — but a shopper reading label faces has no way to know that, which is the point of chart seven.

Chart six / two ways to price

Switch the denominator and the ranking flips.

Every product below is expensive per ounce relative to a conventional gallon. Only some of them are expensive per gram of protein. Where the two dots sit close together, the premium buys protein. Where they fly apart, the premium buys something else — and for oat and almond, that something else is not protein.

Price premium versus a conventional gallon, measured two ways

Multiples of Kroger whole milk by the gallon, which runs $0.0273 per fluid ounce and $0.0273 per gram of protein.

↔ Swipe to see the full chart

Darigold FIT fairlife 2% Silk Protein (soy) Oatly Silk almond Elmhurst oat 3.1× per oz → 1.8× per gram 3.5× → 2.2× 4.0× → 2.5× 3.1× → 8.4× 1.7× → 13.4× 9.1× → 24.4× 10× 25× Multiple of a conventional Kroger gallon, log scale · hollow dot is per ounce, filled dot is per gram of protein
The three dairy and soy products move left when you switch the denominator to protein: the premium is buying the thing it claims to. The three nut and oat products move sharply right. Almond milk is the cleanest case — barely a premium per ounce, thirteen times the cost per gram, because there is almost no protein in it.

Chart seven / the shelf signal

Front-of-pack protein tracks format almost perfectly.

We read the actual pack-front photography for all 22 products rather than inferring from marketing copy. The split is close to total. All six ultrafiltered cartons whose fronts could be read put the gram count on the label face. All five conventional and lactose-free-only jugs keep protein on the back panel, leading instead with fat level, organic, or lactose-free.

Where the protein claim sits, by format

Each mark is one product whose pack front could be read in September 2026. Filled means the protein gram count appears on the label face.

Ultrafiltered dairy 6 of 6 on the front Conventional & lactose-free 0 of 5 on the front Plant-based 3 of 6 — Silk Protein, both Ripples Filled: protein grams on the label face · hollow: nutrition panel only
Placement confirmed by reading Kroger's front-of-pack image endpoint and Walmart primary imagery for each SKU, not from listing copy. The commercial consequence is in chart five: a shopper scanning label faces cannot tell that the conventional gallon delivers protein at a lower cost per gram than most of the products shouting about it.

On the search trend, honestly

Google Trends returned an HTTP 429 during this research, so we have no first-hand index values for "ultra-filtered milk" or "fairlife" and will not invent any. What published, Google-derived analysis does establish: search interest in "high protein" hit a five-year high in 2023 and stayed there, and "protein" set new all-time search records in the US and globally in early 2026. Cross-platform, Spate puts "protein" up 86.1% year over year. The demand curve under this category is still accelerating, not plateauing — but the specific term-level numbers are marked unconfirmed in our data, and should stay that way until someone can pull them.

The test case

Chobani is spending $1.2 billion to re-enter a business it quit in eleven weeks.

In April 2022, less than three months after launching ultra-filtered milk, Chobani pulled it — saying it did not make sense to be in the dairy milk business at that time. In 2026 it agreed to acquire Keurig Dr Pepper's 1.5 million square foot campus near Allentown, Pennsylvania and turn it into a dairy manufacturing hub.

$1.2BInvested over five years
2027Production begins
900+Full-time jobs
3B+ lbsOf PA milk a year at capacity

That is roughly 30% of Pennsylvania's current milk output, and the state calls it the largest private agriculture investment in its history. The question worth asking is not whether the milk is better this time. It is what changed between the two attempts.

What changed

Not volume. The category still moves the same 4.8 billion units. What changed is that a reason to buy arrived, and it is the one Chobani has spent a decade installing in shoppers' heads in a different aisle.

Every other processor chasing this premium has to convince shoppers that its milk is a protein product. Chobani does not. The 2022 launch failed because nobody was shopping the dairy case for protein yet. They are now — and the brand that already owns the word walks in with the association pre-built.

That is the strategic read, and it is worth being precise about why it is not simply "protein is having a moment." Moments are available to everyone. Permission is not. Fairlife has 96% of the ultrafiltered segment because it got there first and spent accordingly. Chobani's asset is that it does not have to buy the association at all — it transfers, at zero marginal cost, from yogurt.

The risk sits on the other side of the same coin. Transferred permission gets you tried. It does not get you kept, and the shelf Chobani is walking back onto has a private-label tier growing 20%, a challenger brand that doubled, and an incumbent with 96% share and Coca-Cola's distribution behind it. Brand permission wins the first purchase. Cost per gram wins the fourth.

What follows from this

Five things the data actually tells you to do.

Implication 01

Put the gram count on the front, or forfeit the segment

Front-of-pack protein is the category's defining merchandising signal, and it tracks format so closely that it functions as the segment boundary. A conventional milk that quietly delivers 8g and says so only on the back panel is not competing in this segment, whatever its nutrition panel says.

Evidence: 6 of 6 ultrafiltered packs front-of-pack · 0 of 5 conventional and lactose-free · the three plant milks that do it are the three positioned on protein

Implication 02

Sell cost per gram, not price per ounce

The per-ounce premium is the objection. The per-gram value is the answer, and almost nobody is making it. Ultrafiltered dairy beats every plant milk except Silk Protein on cost per gram of protein — by five to eighteen times against oat and almond — and no shelf tag, pack or ad in the category currently says so.

Evidence: Alta Dena 3.7¢/g, ALDI 4.2¢/g, Darigold FIT 4.8¢/g against Oatly 22.9¢/g and almond 36.6¢/g

Implication 03

The challenger economics already work

ALDI's private label at $4.39 per 59oz and Alta Dena at $4.49 per 64oz undercut Fairlife per gram of protein by roughly a third. That is the mechanism behind private-label ultrafiltered growing 20% and Darigold FIT doubling. A 96% share held on brand and distribution rather than on cost per gram is a share that can be taken.

Evidence: private-label ultrafiltered +20% · Darigold FIT sales more than doubled · Fairlife carries a reported 190% per-ounce premium over regular whole milk at Walmart

Implication 04

Stop treating plant-based as the competitive set

Plant-based milk is $2.4 billion falling 3% in dollars and 5% in units. It is not where the share is going. The competitive set for a high-protein dairy launch is other high-protein dairy, and the adjacent premium tiers moving on the same driver — lactose-free at +7.3% value, organic and A2 both growing units double digits.

Evidence: plant-based −3% dollars / −5% units · a2 Milk +30% / +25% · Organic Valley +21% / +14% · lactose-free milk +7.3% / +4.1%

Implication 05

Do not build a dairy-plant blend

The hybrid format is not an untapped opportunity, it is a tested and failed one. Live Real Farms' Dairy Plus, Kerry's Smug line, Arla's dairy-oat hybrid and Triballat Noyal's 50/50 range have all been discontinued — attributed to a lack of sales, botched marketing and wrong product choices. The only live hybrid activity is European private label sold at price parity with dairy. If the blend comes up in a planning session, this is the slide.

Evidence: Kerry confirms "the Smug hybrid dairy range is no longer in production" · the single US blend SKU with a verifiable label record delivers 5g per serving, below conventional milk's 8g

The receipts

Every number, with its source.

Category figures first, then the product-level readings the protein and pricing charts are built from. Package protein equals protein per serving times servings per package; price per gram is package price divided by package protein grams.

MeasureFigureBasisSource
Total dairy milk dollars$17.7B, +3%52 w/e 30 Nov 2025, CircanaeDairy News
Total dairy milk units4.8B, flat52 w/e 30 Nov 2025, CircanaeDairy News
Refrigerated white milk$15.8B, +3%; 4.2B units, flat52 w/e 30 Nov 2025, CircanaeDairy News
Ultrafiltered milk growth+14.8% $, +5.5% unitsLast year, CircanaDairyReporter
Ultrafiltered share of milk dollars7%CircanaDairyReporter
Fairlife share of ultrafiltered dollars96%CircanaDairyReporter
Private-label ultrafiltered growth+20%Last year, CircanaDairyReporter
Conventional milk+1.5% value, −0.4% volumeCircana, July 2026DairyReporter
All dairy carrying protein claims$10.4B, +13.7% value, +7.5% volumeCircanaDairyReporter
Lactose-free milk+7.3% value, +4.1% volumeCircanaDairyReporter
Plant-based milk$2.4B, −3% $; 591M units, −5%52 w/e 30 Nov 2025, CircanaeDairy News
Fairlife refrigerated white milk$980M, +25% $; 188M units, +14%52 w/e 30 Nov 2025, CircanaeDairy News
Fairlife flavored milk$276M, +20% $; 53M units, +10%52 w/e 30 Nov 2025, CircanaeDairy News
Organic Valley refrigerated white milk$326M, +21% $; +14% units52 w/e 30 Nov 2025, CircanaeDairy News
a2 Milk$104M, +30% $; +25% units52 w/e 30 Nov 2025, CircanaeDairy News
Private-label refrigerated white milk$9.7B, +1% $; 2.9B units, −1%52 w/e 30 Nov 2025, CircanaeDairy News
Fairlife per-ounce premium190% over regular whole milkAt Walmart, DairyReporter analysisDairyReporter
Alta Dena Protein15g/serving · $4.49/64oz · 3.7¢/gSRP, Jan 2026 launchDairy Farmers of America
Darigold FIT 2%14g · $4.99/59oz · 4.8¢/gKroger, 11 Sep 2026Kroger
ALDI Friendly Farms ultrafiltered14g · $4.39/59oz · 4.2¢/gALDI, 11 Sep 2026ALDI
fairlife 2% ultra-filtered13g · $4.97/52oz · 5.9¢/gWalmart, 11 Sep 2026fairlife
Kroger CARBmaster11g · $4.79/59oz · 5.9¢/gKroger, 11 Sep 2026Kroger
Kroger whole milk, gallon8g · $3.49/128oz · 2.7¢/gKroger, 11 Sep 2026Kroger
Good & Gather 2%, gallon10g · $2.99/128oz · 1.9¢/gTarget, 11 Sep 2026Target
Silk Protein soy milk13g · $5.29/48oz · 6.8¢/gTarget, 11 Sep 2026Target
Oatly Original oatmilk3g · $5.49/64oz · 22.9¢/gTarget, 11 Sep 2026Target
Silk unsweetened almond1g · $4.39/96oz · 36.6¢/gTarget, 11 Sep 2026Target
Elmhurst unsweetened milked oats3g · $7.99/32oz · 66.6¢/gBrand direct, 11 Sep 2026Elmhurst 1925
Hybrid dairy-plant formatDiscontinuedLive Real Farms, Kerry Smug, Arla, TriballatGreen Queen

Grocery prices are store- and zip-specific and were captured 11 September 2026. Servings per package are inconsistently reported — a 52oz bottle is variously "about 6" or "7" servings of 8 fl oz — so per-gram figures here use the 8oz-serving arithmetic and the full table in the source report shows both conventions. Four figures could not be confirmed and are excluded from this page: term-level Google Trends indices for "ultra-filtered milk" and "fairlife", the front-of-pack placement for four SKUs whose pack photography could not be read, and the current shelf price of the one remaining blend SKU.

Keep reading

Where this kind of work goes next.

A category map is the first artefact of an engagement, not the last. These three pages cover what happens after it, and how to judge the people offering to do it.

Weighing us against someone specific? Every comparison we publish, including the ones we lose, is on the comparison hub.

We build maps like this before we buy a single impression.

Schaefer works only with food and beverage brands. Category structure first, creative second, media third. If your category is growing dollars on flat volume, the question is which slice is moving, and that is where we start.

Work With Us