Final Wholesale Inventories +1.3% (vs +1.3% Est) — Two Sectors Made 59% of the Build

Fundamentals · 2026-09-10

Wholesale inventories $958.9 billion at the end of July, up 1.3% (±0.2), exactly matching consensus and unrevised from the advance estimate, with the annual rate at +5.7% (±1.2); sales rebounded +0.8% (±0.4) to $801.3 billion and are up 13.0% (±0.5) on the year; the inventories-to-sales ratio is 1.20 against 1.28 a year ago, but it bottomed at 1.15 in May — the leanest since April 2012 — and has now risen two months; June's inventory change was revised up from +0.2% to +0.4%; petroleum and electrical goods were 11.6% of the stock a year ago and produced 59.3% of the entire annual build, with petroleum inventories +51.7% and electrical +22.7%; strip both out and inventories rose just 2.6%; every headline change in this release is statistically significant.

What Is This?

• What it is: The Census Bureau's Monthly Wholesale Trade Survey, released at 10:00am ET, measuring sales, end-of-month inventories and the inventories-to-sales ratio for merchant wholesalers — the distributors who buy goods and resell them, sitting between the factory and the shop. This is the final estimate; a partial advance figure appeared in the Advance Economic Indicators Report on 27 August. It is a sample survey of roughly 4,200 employer firms, and for this period 54.8% of surveyed companies reported, giving a total quantity response rate of 61.5% for sales and 62.1% for inventories.
• Why it matters: Inventory change is a direct line in GDP, so a build adds to the quarter it happens in and subtracts from the next when it unwinds. More usefully, wholesale is the layer where a demand slowdown shows up before it reaches the shelf — goods pile up one step back before anyone discounts them at the till. This one also lands the same morning as producer prices and existing home sales, which turns out to matter, because the same two forces run through all three.
• How to read it: Read the ratio, not the level, and then read what is inside it. Inventories-to-sales says how many months of selling the stock represents, so a rising ratio means goods are accumulating faster than they move. Then hold one caveat throughout: Census adjusts these figures for seasonality but explicitly not for price changes. Every number here is dollars, so in a month when producer prices for energy are running 24.4% over the year, a petroleum "inventory build" is partly just the same barrels costing more. And check the confidence intervals — Census publishes them, and unlike some of its other releases, this one clears them.

Want the full explainer? Economic News Events, Explained breaks down this release and every other one we cover, in plain English.

Summary

Wholesale inventories stood at $958.9 billion at the end of July, up 1.3% (±0.2) from a revised June and up 5.7% (±1.2) on the year. Consensus was +1.3%, and the monthly change was unrevised from the advance estimate published on 27 August. On the headline, this is a release that told us nothing we did not already know two weeks ago.

Sales did the more interesting thing. Merchant wholesalers sold $801.3 billion in July, up 0.8% (±0.4) on the month and up 13.0% (±0.5) on the year. That annual rate is the number to sit with — 13.0% nominal sales growth is not what a slowing economy looks like, and it is running more than twice the pace of the inventory build.

Which is why the inventories-to-sales ratio is 1.20 against 1.28 a year ago. Wholesalers are carrying eight fewer days of stock per dollar of sales than they were last July. Against 2024, when the ratio spent the year between 1.30 and 1.35, the improvement is larger still. The overhang framing that has attached itself to this series does not survive the ratio.

But the ratio has turned. It fell all year to 1.15 in May — the leanest reading since April 2012, and inside the bottom quarter of 415 monthly observations back to January 1992 — then rose to 1.19 in June and 1.20 in July. Two consecutive increases. The level is still historically lean; the direction is no longer improving. That is the first thing in this release that is genuinely new.

The revisions point the same way. June's inventory change was revised up from +0.2% to +0.4%, doubling the previously published build, while June's sales decline was revised only marginally, from −3.0% to −2.9%. More stock, no more selling.

Now the finding, which is a concentration problem. The annual inventory build was $51.9 billion. Petroleum accounted for $12.3 billion of it, and electrical goods for $18.5 billion — together $30.8 billion, or 59.3% of the entire national build. Those two lines were 11.6% of the wholesale inventory stock a year ago. One-ninth of the shelf produced three-fifths of the growth.

It is the same story in sales. Petroleum sales are up 34.0% over twelve months and electrical up 30.8%; between them they generated 56.0% of the $92.4 billion annual sales increase from 22.6% of the base. On the month, petroleum, electrical goods and professional equipment together produced 51.3% of the $12.2 billion build.

Strip petroleum and electrical out and the picture changes character completely. Inventories across everything else rose 2.6% over twelve months and sales rose 7.4%, with the ratio moving from 1.46 to 1.40. So the leanness is real and survives the exclusion — about three-quarters of the improvement in the headline ratio is still there — but the *growth* in this release is two sectors, and they are two very different stories.

Petroleum is substantially price, and the release says so. Census adjusts for seasonality and trading days, "but not for price changes." Wholesale petroleum inventories are up 51.7% in dollars over twelve months. Producer prices for final demand energy, in this morning's release, are up 24.4% over the same period. Divide one by the other and roughly half of that 51.7% is the same fuel costing more. Petroleum's inventories-to-sales ratio is 0.36, the lowest of any line in the table and up only from 0.32 a year ago, which is what you would expect if the dollars are inflating rather than the barrels accumulating.

Electrical is not price, and that is the more interesting half. Electrical goods sales are up 30.8% over twelve months with inventories up 22.7% — inventories growing fast, sales growing faster. Underneath it, computer equipment sales are up 22.1% on the year and 5.0% on the month, with its own inventories up 17.2% and 5.9%. Professional equipment, the parent category, is up 17.4% on sales. This is the AI build-out arriving in the distribution layer, and it is the fourth independent place this feed has found it: capital goods imports up 46.9%, electronic components on the ISM short-supply list for 18 months, printed circuit assemblies appearing at three separate stages of this morning's producer price report, and now a fifth of the national wholesale inventory build.

Metals belongs alongside them. Sales up 28.2% over twelve months and 5.7% on the month, the strongest monthly sales gain in the durable table, while inventories rose only 2.0% on the year — the metals ratio has collapsed from 2.16 to 1.72. Something is consuming metal faster than distributors can restock it.

The consumer-facing lines say the opposite thing, and they say it consistently. Against 5.4% producer price inflation and 3.7% headline consumer inflation, wholesale sales growth of: alcohol −4.2%, miscellaneous nondurables −4.2%, paper −1.6%, drugs +2.3%, groceries +2.3%, apparel +2.8%, automotive +2.5%. Every one of those is a real decline. Seven categories covering most of what a household actually buys are shipping fewer goods than a year ago once you take the price rise out. Automotive was also the only durable line with falling sales on the month, −0.8%.

Two lines connect straight to this morning's housing release. Lumber inventories are down 1.8% over twelve months and furniture inventories down 3.1% — the two deepest annual inventory declines in the durable table. Existing home sales just posted their first sub-4.0 million print since June 2025 on a third consecutive monthly decline, and new home months' supply sat at 9.6 in July. Distributors of the things that go into houses have already adjusted. Furniture's ratio has fallen from 2.01 to 1.59 in a year, the largest improvement of any line, and it has done it by cutting stock rather than by selling more.

Machinery carries the heaviest ratio in the report at 2.66, though down from 2.91 a year ago, with sales up 12.1% and inventories up only 2.6%.

One methodological note worth making because it cuts the other way from a release this feed covered recently. Every headline change here is statistically significant. Inventories +1.3% (±0.2), inventories +5.7% (±1.2), sales +0.8% (±0.4), sales +13.0% (±0.5) — all four confidence intervals exclude zero, and Census states that "estimated changes shown in the text are statistically significant unless otherwise noted." July's new home sales release, by contrast, had a headline decline of 10.5% with an interval of ±14.0. When Census gives you a number you can rely on, that is worth saying as clearly as when it does not.

Finally, a date to put in the diary. Census has announced an intention to revise: revised estimates incorporating historical corrections and the results of the 2023 and 2024 Annual Integrated Economic Survey are tentatively scheduled for 26 October 2026. That is the second benchmark revision this feed has had to flag in a fortnight, after the BLS preliminary payroll benchmark on 28 August. Two of the statistical series this desk reads weekly are about to be restated against fuller data.

The Internals

The headline against consensus and against the advance estimate:

Measure · July 2026 · June 2026, revised · Consensus · Result

Inventories, monthly change · Up 1.3%, plus or minus 0.2 · Up 0.4% · Up 1.3% · In line, and unrevised from the advance

Inventories, level · $958.9 billion · $946.7 billion · Not forecast · Up $12.2 billion on the month

Inventories, 12-month · Up 5.7%, plus or minus 1.2 · Not applicable · Not forecast · Statistically significant

Sales, monthly change · Up 0.8%, plus or minus 0.4 · Down 2.9% · Not forecast · A partial rebound only

Sales, level · $801.3 billion · $794.9 billion · Not forecast · Up $6.4 billion on the month

Sales, 12-month · Up 13.0%, plus or minus 0.5 · Not applicable · Not forecast · More than double the inventory rate

Inventories-to-sales ratio · 1.20 · 1.19 · Not forecast · 1.28 in July 2025

What was revised since the 6 August release:

Series · As published 6 August · Revised, 10 September · Direction

June inventories, monthly · Up 0.2% · Up 0.4% · The build was twice as large

June sales, monthly · Down 3.0% · Down 2.9% · Marginally less bad

July inventories, monthly, from the 27 August advance · Up 1.3% · Up 1.3% · Unrevised

July inventories, aggregate levels · Advance estimate · Revised · Levels revised, percentage unchanged

The ratio in context. It bottomed in May and has risen twice since:

Month · Inventories-to-sales ratio · Note

January 2026 · 1.25 · The year began here

March 2026 · 1.21 · Falling steadily

April 2026 · 1.19 · Still falling

May 2026 · 1.15 · The low, and the leanest since April 2012

June 2026 · 1.19 · First increase

July 2026 · 1.20 · Second increase

July 2025 · 1.28 · A year ago, for comparison

Where the annual build came from. This is the section the headline hides:

Line · Share of the stock, July 2025 · Inventories, 12-month · Share of the total build

Petroleum · 2.6% · Up 51.7% · 23.6%

Electrical goods · 9.0% · Up 22.7% · 35.7%

Petroleum and electrical combined · 11.6% · Not applicable · 59.3%

Everything else · 88.4% · Up 2.6% · 40.7%

Total · 100% · Up 5.7% · 100%

The same exercise on sales, and on the month:

Contribution · Petroleum · Electrical goods · The two combined

Share of sales, July 2025 · 10.5% · 12.0% · 22.6%

Sales, 12-month change · Up 34.0% · Up 30.8% · Not applicable

Share of the annual sales increase · 27.5% · 28.4% · 56.0%

Share of the monthly inventory build · 18.0% · 18.2% · 36.2%

With professional equipment added · Not applicable · Not applicable · 51.3%

Durable against nondurable:

Grouping · Sales, monthly · Sales, 12-month · Inventories, monthly · Inventories, 12-month · Ratio now · Ratio a year ago

Durable goods · Up 1.1% · Up 16.9% · Up 1.1% · Up 5.3% · 1.48 · 1.64

Nondurable goods · Up 0.5% · Up 9.4% · Up 1.6% · Up 6.5% · 0.91 · 0.93

Total · Up 0.8% · Up 13.0% · Up 1.3% · Up 5.7% · 1.20 · 1.28

Where The Build Actually Is

Every durable line, sorted by annual sales growth:

Kind of business · Sales, monthly · Sales, 12-month · Inventories, monthly · Inventories, 12-month · Ratio now · Ratio a year ago

Electrical goods · Up 0.1% · Up 30.8% · Up 2.3% · Up 22.7% · 0.90 · 0.96

Metals · Up 5.7% · Up 28.2% · Up 1.7% · Up 2.0% · 1.72 · 2.16

Furniture · Down 1.7% · Up 23.0% · Down 1.0% · Down 3.1% · 1.59 · 2.01

Computer equipment · Up 5.0% · Up 22.1% · Up 5.9% · Up 17.2% · 0.80 · 0.84

Miscellaneous durable · Up 1.8% · Up 18.1% · Up 0.9% · Up 2.2% · 1.20 · 1.38

Professional equipment · Up 2.4% · Up 17.4% · Up 2.7% · Up 7.4% · 1.05 · 1.14

Machinery · Up 2.0% · Up 12.1% · Up 0.6% · Up 2.6% · 2.66 · 2.91

Hardware · Up 1.5% · Up 4.2% · Up 0.4% · Up 2.8% · 2.06 · 2.09

Lumber · Up 0.3% · Up 3.9% · Down 0.1% · Down 1.8% · 1.67 · 1.76

Automotive · Down 0.8% · Up 2.5% · Up 0.3% · Up 0.7% · 1.67 · 1.70

Every nondurable line, same sort:

Kind of business · Sales, monthly · Sales, 12-month · Inventories, monthly · Inventories, 12-month · Ratio now · Ratio a year ago

Petroleum · Up 3.1% · Up 34.0% · Up 6.5% · Up 51.7% · 0.36 · 0.32

Farm products · Up 3.2% · Up 19.5% · Up 3.9% · Up 19.4% · 1.14 · 1.14

Chemicals · Up 2.5% · Up 17.1% · Up 0.4% · Up 5.8% · 1.01 · 1.12

Apparel · Unchanged · Up 2.8% · Up 0.3% · Down 0.8% · 1.98 · 2.05

Drugs · Down 0.1% · Up 2.3% · Up 1.6% · Up 2.5% · 0.94 · 0.94

Groceries · Down 1.5% · Up 2.3% · Down 0.2% · Up 3.6% · 0.73 · 0.72

Paper · Down 1.3% · Down 1.6% · Up 1.6% · Up 2.8% · 1.06 · 1.01

Alcohol · Down 2.0% · Down 4.2% · Up 0.6% · Down 4.5% · 1.66 · 1.67

Miscellaneous nondurable · Down 1.2% · Down 4.2% · Up 1.2% · Up 1.5% · 1.88 · 1.77

The real-terms test. Nominal sales growth against the 5.4% producer price rate and the 3.7% consumer rate:

Kind of business · Sales, 12-month · Against 5.4% producer prices · Against 3.7% consumer prices

Alcohol · Down 4.2% · A real decline of about 9 points · A real decline of about 8 points

Miscellaneous nondurable · Down 4.2% · A real decline of about 9 points · A real decline of about 8 points

Paper · Down 1.6% · A real decline of about 7 points · A real decline of about 5 points

Drugs · Up 2.3% · A real decline of about 3 points · A real decline of about 1 point

Groceries · Up 2.3% · A real decline of about 3 points · A real decline of about 1 point

Automotive · Up 2.5% · A real decline of about 3 points · A real decline of about 1 point

Apparel · Up 2.8% · A real decline of about 3 points · A real decline of about 1 point

Note what those seven lines have in common: they are the consumer-facing half of the wholesale economy, and every one of them is shipping less in volume terms than a year ago. The categories growing fast are fuel, electrical equipment, metals, farm products and chemicals — inputs and capital goods, not shopping.

Impact on USD

• Slight bullish — an inventory build is an addition to third-quarter GDP by arithmetic, and 13.0% annual sales growth in nominal terms is not the profile of an economy rolling over.
• The ratio turning up off a fourteen-year low — 1.15 in May to 1.20 now, two consecutive increases — is the first sign in this series that accumulation is outpacing selling.
• June revised up from +0.2% to +0.4% adds a little more to the quarter than was previously on the books.
• Against it: none of this is real. Census does not adjust for prices, energy is up 24.4% over the year, and half the largest single contributor is fuel costing more.
• No trade in it. The print matched consensus, the monthly change was unrevised from the advance two weeks ago, and it lands at 10:00am ET behind producer prices.

Impact on US Indices (ES / NQ / YM)

• Mixed, lean bullish for the technology and industrial complex, bearish for consumer staples distribution.
• The AI thread is now visible in a fourth dataset: electrical goods sales up 30.8% over twelve months, computer equipment up 22.1%, and together with professional equipment they made over half the monthly inventory build.
• Metals is the sleeper — sales up 28.2% on the year with inventories up only 2.0%, and a ratio collapsing from 2.16 to 1.72. Distributors cannot restock fast enough.
• Staples and beverages take the clearest hit: alcohol shipments down 4.2% on the year, miscellaneous nondurables down 4.2%, groceries and drugs up 2.3% against 3.7% consumer inflation. Volume is falling across the household basket.
• Housing-linked distribution has already adjusted rather than being caught out — lumber inventories down 1.8% and furniture down 3.1% over twelve months, the two deepest cuts in the durable table.

Impact on Gold

• Mixed, lean bullish — a 51.7% annual rise in the dollar value of petroleum stock against 24.4% energy producer price inflation is another reading of the same inflation this feed has tracked all fortnight.
• Real yields are the transmission and nothing here moves them, but the composition argues that nominal growth is being flattered by prices rather than driven by volume.
• The consumer table is the durable support: seven wholesale categories covering most household spending are shipping less than a year ago in real terms.
• Against it, a rising inventories-to-sales ratio and a Q3 GDP contribution both argue mildly for a firmer growth path and a Fed with no reason to move.

What To Watch

• Consumer prices — Friday, 11 September, 8:30am ET. The deflator problem in this release resolves itself there. Every figure above is dollars; tomorrow says how many of them are volume.
• FOMC decision and dot plot — Tuesday and Wednesday, 15-16 September. Three dissents for a hike in July, and an inventory series that is no longer tightening.
• July business inventories — Wednesday, 16 September, 10:00am ET, which lands on FOMC decision day. That release folds today's wholesale figure together with manufacturers and retailers and gives the whole-economy ratio; wholesale is only one of the three layers.
• The Census benchmark revision — tentatively 26 October, 10:00am ET. Historical corrections plus the 2023 and 2024 Annual Integrated Economic Survey. The second statistical benchmark to hit this desk in two months.
• Whether the ratio makes it three. Two consecutive increases off a fourteen-year low is a turn only if it continues; at 1.20 the level is still inside the bottom quarter of the 1992-2026 record.
• Next wholesale trade — Thursday, 8 October, 10:00am ET, covering August.

TLDR

Census Monthly Wholesale Trade, July 2026 (released 10 September, 10:00am ET, CB26-136):

• Inventories $958.9 billion, up 1.3% (±0.2) against a 1.3% consensus — in line, and unrevised from the 27 August advance
• Annual rate up 5.7% (±1.2); sales up 0.8% (±0.4) to $801.3 billion and up 13.0% (±0.5) on the year
• Inventories-to-sales 1.20 against 1.28 a year ago — wholesalers still carrying eight fewer days of stock per dollar of sales
• But the ratio bottomed at 1.15 in May, the leanest since April 2012, and has risen two months running to 1.19 then 1.20
• June's inventory build was revised up from +0.2% to +0.4%; June sales revised only from −3.0% to −2.9%
• Petroleum and electrical goods were 11.6% of the stock a year ago and made 59.3% of the entire annual build — petroleum inventories +51.7%, electrical +22.7%
• Strip both out and inventories rose 2.6% on the year against sales up 7.4%, with the ratio at 1.40 from 1.46
• Petroleum is largely price. Census does not adjust for prices, and energy producer prices are up 24.4% over the same twelve months
• Electrical is not. Computer equipment sales +22.1% annually and +5.0% monthly — the AI build-out in a fourth dataset
• Metals sales up 28.2% on the year with inventories up 2.0%, ratio collapsing from 2.16 to 1.72
• Seven consumer-facing categories are in real decline: alcohol −4.2%, miscellaneous nondurable −4.2%, paper −1.6%, drugs, groceries, automotive and apparel all between +2.3% and +2.8% against 3.7% inflation
• Lumber inventories −1.8% and furniture −3.1% over twelve months, the two deepest cuts in the durable table, alongside a housing market that just broke below four million sales
• Every headline change clears its 90% confidence interval — unlike July's new home sales, where the 10.5% decline carried a ±14.0 band

The number was known two weeks ago and the release still had something in it. Wholesalers are carrying eight fewer days of stock per dollar of sales than a year ago, which is the opposite of an overhang — but the ratio bottomed at 1.15 in May, the leanest reading since April 2012, and has now risen two months running. What is growing is not a broad restock. Petroleum and electrical goods were one-ninth of the shelf a year ago and produced three-fifths of the entire build, and those two are different stories: petroleum is roughly half fuel-price inflation in a release Census explicitly does not deflate, while electrical and computer equipment is the AI build-out arriving in the distribution layer for the fourth time in a fortnight. Underneath both, seven consumer-facing categories are shipping less than they did a year ago once prices come out. Consumer prices tomorrow at 8:30am ET decide how much of the 13.0% sales growth was real; business inventories land on FOMC decision day, 16 September.

_For informational purposes only. Not investment advice._