Grocery Prices in July 2026: Our Index Rose 0.72% While the CPI Said They Fell

The Bureau of Labor Statistics reported this morning that grocery prices fell 0.1% in July 2026. We measured them rising 0.72%.

Both numbers are defensible. Neither is a mistake. And the gap between them is the most interesting thing in this month's data.

Here is what we saw, where we saw it, and why two honest measurements of the same month can point in opposite directions.

The one-minute answer

• GroceryChop Price Index, July 2026: +0.72% month over month, built from 26,772 distinct price observations drawn from 81,957 shelf-price readings across 1,381 stores and 27 grocery banners.
• The BLS food-at-home index: −0.1% month over month, seasonally adjusted, and +2.7% over the past twelve months.
• The increase was not broad. Beverages (+4.08%) and meat (+3.36%) carried it. Pantry staples fell 3.88%.
• Why the disagreement: the BLS seasonally adjusts and we do not, and they price a designed sample while we observe a much larger non-random one.
• Free to cite. Everything is CC BY 4.0 with a full methodology page and a CSV download.

What we measured in July

GroceryChop Price Index, July 2026. Month-over-month changes compare the same items at the same stores across both months. Not seasonally adjusted.

Two notes on reading that table honestly.

Fruits and vegetables cleared our publication threshold on 888 readings — the thinnest cell we published this month. Treat it as directional. We print the sample size next to every figure precisely so you can make that call yourself instead of trusting us.

A "price reading" is one scrape of one product at one store. Because a single product is returned by many different search terms, the same unchanged price often gets read repeatedly. Collapsing those repeats gives 26,772 distinct observations behind the headline. We report both numbers because the honest answer to "how much data is behind this?" is the smaller one.

Why our number disagrees with the CPI

If you only read one section, read this one.

The BLS reported food-at-home down 0.1% for July on a seasonally adjusted basis. We reported up 0.72%, unadjusted. That is a gap of roughly eight tenths of a percentage point, which is large for a single month.

Two structural differences explain most of it:

Seasonal adjustment. The BLS smooths out predictable seasonal patterns — summer produce gluts, holiday demand, back-to-school cycles — so their number answers "how did prices move relative to a typical July?" Ours answers "what did a shopper actually pay in July versus June?" Both are useful. They are not the same question, and in a month with a real seasonal pattern they can easily point different directions.

Sample design. The BLS prices a carefully designed statistical sample built to represent the country. We observe whatever is on the shelves of the stores our users shop — vastly more prices, from a set of stores that was never randomized. That is a genuine limitation of our data and we say so on the methodology page. It also means we can see things a smaller sample cannot, like a 4% move in a single category.

We do not think the useful question is which number is "right." A monthly survey and 26,772 observed shelf prices are two different instruments pointed at the same reality, and when they disagree, the disagreement is itself worth reporting.

What actually moved

The headline hides the story. July was not a month where everything got slightly more expensive — it was a month where two aisles got noticeably more expensive and one aisle got noticeably cheaper.

Beverages, up 4.08%. The largest single move of the month in either direction, and on a healthy 13,896 readings. If your July grocery bill felt higher than June's, the drinks aisle is the most likely reason.

Meat, poultry, fish and eggs, up 3.36%. The second-largest move, on our second-deepest sample at 22,189 readings. Protein has been the most volatile part of the basket since we started publishing in April.

Other food at home, down 3.88%. This is the pantry-and-snacks group — the middle of the store. It fell nearly as hard as beverages rose.

Dairy, up 0.44%. Nearly flat, on our single deepest category at 29,958 readings. When the biggest sample barely moves, that is a real signal about a stable category rather than a data artifact.

The practical version: a shopper buying mostly shelf-stable pantry items had a cheaper July. A shopper buying mostly meat and drinks had a more expensive one. A single national inflation number cannot tell you which of those you are, which is exactly why we publish the split.

If you want the longer view on what has been driving these moves, we wrote up the structural causes in why grocery prices keep going up.

Where it hit hardest

Grocery inflation is not evenly distributed across the country. Three metros had enough store coverage in July for us to publish a month-over-month figure:

San Diego is our deepest panel and our home market, which makes the contrast useful: grocery prices there rose at roughly half the pace we measured nationally, while Dallas shoppers saw more than four times it. Same month, same basket definition, wildly different experience at the register.

Metros with thinner coverage — Chicago, Seattle, Atlanta and others — show their underlying figures on the metro pages but do not publish a month-over-month change yet. A change computed on too few stores is noise wearing the costume of a number, and we would rather publish nothing than publish that.

How we measure it

The short version, with the full detail on the methodology page:

A fixed 19-item basket. Whole milk, large eggs, boneless chicken breast, ground beef, white bread, russet potatoes, ground coffee and a dozen more everyday staples. The basket does not change month to month, so the index measures prices rather than shifting composition.

Strict product matching. Product title, package size and unit all have to agree, so a 3-lb value pack and a per-pound listing land on the same comparable series rather than being treated as different products.

Matched-sample comparison. Every month-over-month figure compares the same item at the same store across both months. A store that appears in July but not June contributes nothing to the change, which is what keeps panel growth from masquerading as inflation.

BLS category weights. Relative importance comes from the official CPI tables, not from how much data we happen to collect. Otherwise a category we scrape heavily would dominate the index for no economic reason.

Aggregates only. We publish no store-versus-store comparisons and no individual product price lists. Every figure clears minimum thresholds on observations, distinct stores and distinct banners; anything below the floor is withheld rather than estimated.

What we will not claim

This is the part most price indexes leave out, so we will be explicit.

No year-over-year figure. Our series starts in April 2026. Annual comparisons begin in spring 2027 and not a month before. Anyone quoting a GroceryChop year-over-year number for 2026 is quoting something that does not exist.

Do not read April through June as a trend. Our store panel grew from 67 stores in April to 1,381 in July. The early months come from a much smaller matched sample and are genuinely noisy. July is the first month we would call stable.

Basket costs are indicative, not precise. We publish a weighted basket cost by city, but which items clear our thresholds differs between markets, so part of any gap between two cities is coverage rather than price.

Published months are frozen. Once a month is published we lock it. The figure you cite today is the figure that will be there next year. If we ever need to revise one it will be a deliberate, dated, logged revision — never a silent change.

Frequently asked questions

Did grocery prices go up in July 2026?

By our measure, yes — 0.72% month over month across 26,772 distinct shelf-price observations. By the BLS measure, no — food at home fell 0.1% seasonally adjusted. The increase we observed was concentrated in beverages and meat, while pantry staples fell.

Why is the GroceryChop number different from the CPI?

Seasonal adjustment and sample design. The BLS smooths seasonal patterns and we do not, so ours is the raw change a shopper experienced. The BLS prices a designed national sample; we observe a much larger but non-random set of stores. Both are legitimate, and neither is a correction of the other.

Which grocery categories went up the most in July 2026?

Beverages at +4.08%, fruits and vegetables at +3.65% on a thin 888-reading sample, and meat at +3.36%. Cereals and bakery fell 0.89% and pantry staples fell 3.88%.

How does GroceryChop measure grocery inflation?

A fixed 19-item basket priced across every chain we track, compared as the same item at the same store across consecutive months, weighted using BLS relative-importance tables, with minimum sample thresholds on every published cell.

Is the data free to use?

Yes, CC BY 4.0. Quote it, chart it, republish it commercially — just credit GroceryChop. There is a CSV export, a data dictionary, and an RSS feed for new releases.

Can I see grocery inflation for my city?

For metros where our panel is deep enough, yes — see the metro directory. Thinner markets show their figures but withhold the month-over-month change until coverage supports it.

The takeaway

July 2026 was a month where the national headline and the shopping experience diverged, and where the shopping experience diverged again depending on what was in your cart and which city you bought it in. Beverages and meat up meaningfully, pantry staples down, San Diego at half the national rate and Dallas at four times it.

That is the case for publishing the split rather than a single number. "Groceries rose 0.72%" is true and nearly useless on its own. "Beverages rose 4% and pantry staples fell 4%, and your metro may look nothing like the national figure" is something you can actually shop against.

If you want to know what that means for your basket specifically, that is what the rest of the site is for: compare a product across every nearby store, build your list and find the cheapest store for it, or see what is actually on sale near you.

And if you are a reporter, researcher or analyst: the full July release, category tables, metro detail and raw CSV are at grocerychop.com/price-index, free to cite. If you need a cut we do not publish, ask.

See the full July 2026 release →