US core capital goods orders rose 1.6% month-over-month in August 2026, a print that pushed the year-on-year rate to roughly 14.1% and marked a five-year high in YoY growth momentum [S2][S9]. The Census Bureau advance report places nondefense capital goods new orders at $100.5 billion, with shipments at $95.5 billion and unfilled orders climbing to $956.7 billion [S1].
That is a meaningful divergence from the July reading of 0.2%, when transportation-driven aircraft lumpy bookings masked a soft underlying core, and where the market had expected 0.9% [S3]. August reversed that underperformance, with the headline durable goods aggregate at $338.6 billion and the core measure contributing the bulk of upside.
How Core Capital Goods Orders Are Defined and Why the Exclusions Matter
Core capital goods is the Census Bureau's formal category "nondefense capital goods excluding aircraft," a definition engineered to remove two noise sources: defense procurement, which follows appropriations rather than commercial demand, and commercial aircraft orders, which arrive in lumpy batches that can swing a monthly headline on a single widebody deal [S3]. The series includes machinery, industrial equipment, electrical gear, computers, and fabricated metal products, and represents a large share of total durable goods while behaving far more steadily month to month [S3].
The two exclusions are deliberate. Defense orders follow appropriations and contract awards rather than commercial demand. Aircraft orders arrive in lumpy batches that can swing a monthly headline on a single widebody deal [S3]. Strip those out and what remains is equipment companies buy because they expect more capacity, which is why the Federal Reserve and most macro desks treat this series as the cleanest available read on US business investment intent [S3][S5].
August 2026 Print in Numbers: Orders, Shipments, Backlog
The August 2026 advance report shows nondefense capital goods new orders at $100.5 billion, up $1.2 billion or 1.2% MoM, while shipments fell 1.3% to $95.5 billion [S1]. At the aggregate durable goods level, new orders held effectively flat at $338.6 billion, but excluding transportation, new orders rose 0.3 percent, and excluding defense, new orders rose 0.1 percent [S1].
The backlog ratio is the figure engineers should watch: unfilled orders for total manufactured durable goods reached $1,609.4 billion in August 2026, up twenty-five of the last twenty-six months, with transportation equipment at $1,009.8 billion driving most of the build [S1]. This widening gap between new orders and shipments is a leading signal: orders typically lead shipments by 1 to 3 months, and sustained strength in core orders translates into stronger GDP investment spending in subsequent quarters [S8].
Why This Print Beats the July Headline and the Sentiment Surveys
July 2026 looked strong on the surface: headline new orders rose 1.1 percent to $339.3 billion, roughly double the 0.5 percent consensus, while core capital goods orders rose only 0.2 percent [S3]. August flipped that composition: core capital goods orders jumped 1.6% MoM and YoY growth accelerated by about 80 basis points to roughly 14.1%, a five-year high in the YoY rate [S2][S9].
The divergence with the Philadelphia Fed's Manufacturing Business Outlook Survey is worth flagging. The Philly future activity index hit 73.6 in August 2026, the highest reading since 1983, while current activity stood at 47.4 [S3]. A survey captures executive feelings about the next six months; an order book captures money already committed. The two can diverge because a sentiment turn happens in a week, while a capital purchase order requires a business case, a budget cycle, and usually board or division approval [S3]. The August print is the more credible read because it is the one with money behind it.
Reading the Series: Three-Month Trend Beats Single Month
Single-month prints are noisy by design, which is why core capital goods orders should be read as a three-month trend rather than a point estimate [S3]. A three-month average smooths aircraft lumpiness and defense timing, and the August 2026 result lifts that trend materially. For procurement teams, the practical question is whether the backlog-to-shipment ratio is widening: August 2026 shows exactly that, with unfilled orders at $1,609.4 billion and shipments at $333.8 billion, a ratio that has expanded for most of the past 26 months [S1].
Capital planners should also watch inventories, which climbed 0.5% in August 2026 to $608.1 billion, up eleven consecutive months, indicating that manufacturers are not just booking orders but also building stock against expected demand [S1]. That combination, rising orders, rising unfilled orders, and rising inventories, has historically preceded equipment-led GDP contributions by two to three quarters.
What This Means for Capital Equipment Suppliers and Buyers
For an industrial supplier, the August 2026 print matters more than the July headline because it confirms the underlying core, not the aircraft lumpy noise, is in gear. A 1.6% MoM rise in core orders against a 0.4% prior three-month average, with the YoY rate at roughly 14.1%, is the kind of signal that converts quoting activity into booked purchase orders within one to two quarters [S2][S9]. For a buyer, the same data cuts the other way: if you are sourcing capital equipment in Q4 2026, expect longer lead times on items already in a tightening backlog, particularly anything tied to industrial machinery, electrical gear, or fabricated metal products.
The composition question matters. The series covers machinery, industrial equipment, electrical gear, computers, and fabricated metal products, so demand intensity varies by subcategory [S3]. Aggregate strength does not mean every subsegment is red-hot, but the August 2026 print is broad enough that spot shortages in motors, reducers, and switchgear are plausible by early 2027 if the trend holds. For reliability engineers tracking conveyor and drive systems, the lead-time data points feed directly into spare-parts planning and condition-monitoring budgets.
Limits, Caveats, and What Could Break the Signal
Three caveats are non-negotiable when reading this series. First, the Census Bureau M3 survey is not based on a probability sample, so sampling error cannot be measured and confidence intervals cannot be computed, which is why the August print carries the "statistical significance not measurable" footnote [S1]. Second, the Philly Fed survey is Third District only, covering eastern Pennsylvania, southern New Jersey, and Delaware, and is therefore regional rather than national [S3]. Third, a single month of core orders is noisy, and the three-month trend is the unit of analysis for serious forecasting [S3].
The structural risks to the signal are a sudden defense pullback, a credit tightening that delays capital approvals, or a transport-driven headline that distracts from a softer core. Trackable signals over the next two prints are the September 2026 advance release on October 27, 2026 [S1], the next Philly Fed future activity index release, and any revision to the July 2026 core figure beyond the 0.2% initial print.
Detailed specification references: core machine, signal calibrator, and signal conditioner.
This topic is covered further in Cold Chain Equipment Lead Times in 2026: Reefer, ULT Freezer, and Walk-In Room Spec Guide.