Construction spending looks almost frozen from the top. Equipment demand does not.

The U.S. Census Bureau’s May construction spending report, released July 1, estimated total construction spending at a seasonally adjusted annual rate of $2.210 trillion. That was only 0.1 percent above April’s revised rate and 1.5 percent below May 2025. Through the first five months of 2026, spending was down 2.7 percent from the same stretch last year.

That sounds soft, and in some parts of the market it is. But the same report showed public construction up 0.5 percent from April, with highway construction at a $150.6 billion annual rate and educational construction at $113.4 billion. Private residential spending rose slightly. Private nonresidential spending slipped.

The headline is not “construction is dead.” It is stranger than that. The market has enough work to keep equipment busy, but the work is moving into narrower lanes. Public work, utilities, large industrial projects, healthcare, data centers, and certain infrastructure jobs are still pulling machines. Residential, warehouses, and some private building categories are more uneven.

That split matters more than the total spending number.

FieldFix Editor’s Note: A flat market is where weak fleet tracking gets expensive. FieldFix helps equipment owners track service history, downtime, repair spend, hours, fuel, and cost per hour by machine, so buying and replacement decisions are based on the units that actually earn money.

The spending number hides the equipment story

A $2.210 trillion annualized spending rate is huge. It also does not tell an owner whether to buy another excavator, rent a dozer, replace a compact track loader, or delay a telehandler purchase.

Construction spending measures dollars put in place. Equipment owners live in machine hours, mobilization windows, crew availability, job mix, fuel, parts, and downtime. A small shift inside the spending number can change equipment demand faster than the headline suggests.

If highway work is stronger, contractors need excavators, dozers, graders, pavers, compactors, rollers, haul trucks, water trucks, loaders, milling machines, trenchers, pumps, light towers, and support equipment. If education work is steady, demand may lean toward site prep, utilities, concrete, lifting, aerial access, material handling, and smaller support machines. If private nonresidential slips, local rental yards tied to warehouses, offices, retail, and light commercial work may feel pressure in different categories.

The spending total smooths all of that into one number. That is useful for economists. It is not enough for fleet planning.

The better question is simple: what kind of work is actually funded, permitted, awarded, and mobilizing in your market?

That is where the equipment signal lives.

Starts are stronger, but they are not broad

The starts data gives the market a different shape.

Dodge Construction Network reported that total construction starts improved 34.1 percent in May to a seasonally adjusted annual rate of $1.78 trillion. Nonbuilding starts jumped 91.9 percent for the month. Nonresidential building starts rose 17.8 percent. Residential starts fell 2.1 percent.

Year to date through May, Dodge said total starts were up 12.7 percent. Nonbuilding starts were up 32.9 percent. Nonresidential starts were up 12.3 percent. Residential starts were down 4.9 percent.

That is a good demand signal, but it is not broad enough to justify sloppy buying. Dodge tied much of the monthly strength to megaproject activity in healthcare, manufacturing, utilities, and data centers. The same report pointed to weakness in institutional construction pockets, warehouses, and residential work.

That is the equipment market in one paragraph.

Some contractors are looking at a strong pipeline. Others are staring at a slower phone. Some rental branches will be short on heavy earthmoving, power, compaction, pumps, and aerial equipment. Others may have too many compact units chasing small jobs that are not penciling out like they did two years ago.

This is why national optimism can feel useless on the ground. A contractor in the path of a utility buildout or bridge project may need iron immediately. A contractor tied to speculative warehouses or soft single-family work may need to protect cash.

Both can be true in the same national market.

Public work changes the fleet mix

Public construction does not behave like private work. The bid cycle is different. Payment timing is different. Specs are different. Compliance is different. The equipment mix is different too.

Highway, bridge, utility, water, sewer, school, and public building work can keep machines busy for long stretches, but the margins are not automatic. Public jobs can be schedule-heavy and paperwork-heavy. They often require tighter documentation, certified payroll, traffic control, safety compliance, inspection windows, bonding, and more coordination than smaller private jobs.

For equipment owners, the upside is utilization. A long public job can justify owned equipment if the machine is core to the work and the operator plan is solid. A contractor with steady road, utility, or municipal work may be able to carry excavators, loaders, compactors, rollers, trench boxes, pumps, and trucks with more confidence than a contractor chasing one-off private jobs.

The risk is overreading one project.

A large public job can make a machine look necessary for six months. The harder question is what happens after that project. If the machine can roll into similar work, ownership may make sense. If it only fits one contract, rental or subcontracting may be cleaner.

Public work also tends to expose weak maintenance habits. A machine that misses a residential site for a day is annoying. A machine that shuts down a public road crew, utility crew, or bridge phase can create bigger consequences. Preventive maintenance, parts availability, field service, and backup plans matter more when the schedule is visible and penalties are real.

Prices are still irritating

Equipment buying discipline would be easier if machine prices had cooled in a clean way. They have not.

The FRED series for construction machinery and equipment producer prices, sourced from the Bureau of Labor Statistics, put the May 2026 index at 324.499, up from 323.759 in April and 322.601 in January. That is not a dramatic monthly move, but it is another reminder that new equipment costs are not giving contractors much relief.

High equipment prices change the rent-versus-buy math. They also raise the bar for replacement. A contractor may know an older machine is becoming expensive, but replacing it at today’s price can still hurt.

That is where a lot of 2026 fleet decisions will get uncomfortable.

Keeping an older machine can look smart until repairs, downtime, weak resale value, and missed work erase the savings. Buying new can look smart until utilization slips or the financed payment follows the company into a slower season. Renting can look smart until the same machine is needed every month and the rental invoices become their own kind of payment plan.

There is no universal answer. The answer depends on the machine, the work, the operator, and the maintenance record.

That sounds obvious. It is also where many contractors get lazy. They look at monthly payment, rental rate, or purchase price in isolation. The real comparison is cost per productive hour, including downtime risk.

Labor tells the same uneven story

Equipment demand is also tied to labor. A machine without a qualified operator is not capacity. It is a parked cost.

The Associated General Contractors of America said construction employment increased in only 152 of 360 metro areas from May 2025 to May 2026, while 161 metro areas lost construction jobs and 47 were flat, according to its June 29 analysis. The strongest gains were concentrated in places like Houston, St. Louis, Baton Rouge, Minneapolis, and Charlotte. Large losses showed up in several California and Pacific Northwest markets.

That spread matters. Equipment dealers and rental companies can see strong national demand and still have weak local labor conditions. Contractors can have funded work and still struggle to add crews. A market can have big projects and limited people to run the iron.

This is where utilization plans can get too optimistic. A contractor buys the machine because the work exists, then discovers the operator plan is thin. The owner starts moving the best operator between too many units. Production drops. Damage goes up. Preventive maintenance gets skipped. One new machine creates a scheduling problem instead of solving one.

The labor question has to come before the purchase order.

Who will run it? How many hours per week? What happens when that operator is on another job? Who services it? Who hauls it? What job class keeps it busy after the first project?

If those answers are vague, the machine is not ready to be owned.

Dealers and rental yards need local reads

For dealers, the May data argues against a one-size-fits-all sales push. The buyer tied to public work has a different problem than the buyer tied to residential site prep. The contractor near a data center corridor is not the same as the contractor in a market where warehouse starts are slowing.

The best sales conversations will be specific. Not “the market is strong.” More like: your rental history says this size excavator has become core fleet, your service records show the old unit is eating shop time, and your next six months of work can support the payment.

Rental yards need the same discipline. The starts data makes it tempting to add fleet around megaprojects, utilities, and infrastructure. Some yards should. But fleet bought for one phase of one big project has to go somewhere when that phase ends.

General-purpose machines with repeat customers are easier to justify. Specialized units can be excellent money, but only when the utilization path is real and the exit plan is honest.

The danger is buying yesterday’s demand. A data center, bridge, highway, hospital, or manufacturing project can pull a lot of fleet at once. It can also release a lot of fleet at once.

The practical read for contractors

The market is not bad. It is selective.

May spending was basically flat. Starts were strong, but heavily shaped by large projects and public-facing work. Equipment prices remain stubborn. Labor gains are uneven by metro. That combination rewards contractors who know their numbers and punishes contractors who buy because the calendar feels busy.

The next equipment decision should start with the work, not the machine.

Break the next six months into job types. List the machine classes each job needs. Separate owned core equipment from occasional equipment. Pull the last year of rental invoices by category. Review repair history by unit. Compare cost per hour against replacement cost and rental rates. Be honest about operator availability.

Then make the call.

Buy when the machine fills repeatable, profitable work and has an operator. Rent when the need is tied to a project window or uncertain backlog. Sell when a machine is staying busy but losing money through repairs and downtime.

Flat spending does not mean flat opportunity. It means the easy read is gone.

In 2026, equipment demand is following specific work: public infrastructure, utilities, healthcare, manufacturing, data centers, and local markets with real crews to put in the seat. Contractors who chase the headline will get whiplash. Contractors who follow the work will make better fleet decisions.

Sources