Speed came back, price did not
ResearchReshoring activity rose in 2026 and satisfaction with it fell hard, from 96 percent to 65 percent in one year. The benefits manufacturers actually booked were time and proximity. The reason they still lose work is unit price. A survey of 249 manufacturers is quiet corroboration for the least comfortable part of this site's argument.
2026-09-04 · Field notes · 8 min read · By Sondre Hegerland KristiansenIn July this site argued that the reshoring boom was measurable and that on the physical numbers it was not happening: manufacturers were absorbing tariffs into prices rather than moving production. The second annual USA Reshoring Survey has arrived to complicate that.
It cuts both ways. Activity is genuinely up, and satisfaction with it has fallen off a cliff. The two stop being in tension once you look at which benefits manufacturers actually booked, and which ones they were expecting.
Two numbers moving in opposite directions
The Reshoring Initiative and Regions Recruiting surveyed 249 U.S. manufacturers: 118 original equipment manufacturers and 131 contract manufacturers. The report went up in late August 2026, and trade coverage followed on 2 September.
The activity numbers moved up consistently. 36 percent of OEMs had reshored or were actively engaged in further reshoring, against 29 percent in 2025. Among contract manufacturers, the share currently quoting reshoring projects doubled, from 16 percent to 32 percent. 79 percent of contract manufacturers said customers had raised reshoring in the previous year. 63 percent of OEMs planned U.S. capital expenditure in 2026 or 2027. This is not last year's flat picture.
Then the other number. The share of OEMs satisfied with their reshoring results fell to 65 percent, from 96 percent in 2025, as reported by trade coverage of the survey. Dissatisfaction rose from 4 percent to 25 percent. A near-unanimous verdict became a two-thirds majority in twelve months, while the underlying activity increased. The causes attributed in that coverage are labor costs, vendor gaps, inflation and implementation difficulty.
A collapse in satisfaction during an expansion is what a category looks like once the easy cases are done. The firms reshoring in 2025 were the ones for whom it was obvious. The 2026 cohort is the next tier down, where the real constraints live.
The gains landed on time, not on cost
Ask what the OEMs reporting impacts actually got, and the answer is unusually clean. 70 percent cited improved speed to market. 65 percent reported better fulfillment or on-time delivery. 60 percent noted logistics savings. Every one of those is a time or distance benefit. None of them is a unit-cost benefit.
The de-risking numbers point the same way. 53 percent of contract manufacturers said geopolitical risk was why their customers were reshoring, up from 24 percent a year earlier, which is the sharpest single move in the survey. 60 percent reported that customers importing from China or Taiwan were discussing de-risking, and 19 percent said transitions were actually underway.
Read that against the register and it is close to a restatement of P-07, which predicts that location comes to be decided by bill-of-materials structure, energy, logistics, tariffs and risk, rather than by wages. These manufacturers are describing decisions made on delivery time, distance to customer and political exposure, not on the hourly rate.
Ninety-four percent still lose on price
94 percent of contract manufacturers said price is the primary reason they lose orders to imports. That near-unanimity explains the satisfaction collapse. A firm that reshored expecting the tariff wall to make its quote competitive found that the wall changed the landed cost of a competitor's part without changing its own cost of production.
57 percent of OEMs named policy uncertainty as their primary challenge, last year's finding in a harsher form: a tariff schedule that can be renegotiated is not a capital investment case. It moves prices immediately and capacity slowly, and a manufacturer siting a line has to underwrite a decade against a policy with an expiry date.
The honest reading is that reshoring in 2026 buys responsiveness at a price premium. That is a real product with a real market, and roughly what P-09 predicts customers will pay for. It is not the product most of the 2026 cohort thought they were buying, which is why a quarter of them are unhappy.
Technicians, and the vendor who is not there
66 percent of respondents rated hiring technicians, meaning welders, machinists and electrical or chemical technicians, as very difficult or at crisis levels. 60 percent said the same of maintenance and repair technicians. This site reached that conclusion from a single company last week, when Isembard turned out to be constrained by operators rather than machines or capital. A survey of 249 firms is a different quality of evidence, and the same finding.
The two words worth more than either percentage are "vendor gaps," among the reported causes of dissatisfaction. A vendor gap is what a reshoring manufacturer finds in the second week: the part can be made here, and the fixture, the heat treat, the plating line or the tool that makes it cannot be bought here. That is P-13, which predicts toolmaking and fixture capability is recognized as one of Norway's critical industrial gaps and is being rebuilt. This is U.S. data, so it speaks to the recognition, not to Norway. Recognized now has evidence. Being rebuilt does not.
Forty percent now count the whole cost
The quietest number here is the one this site should care about most. Use of total cost of ownership analysis to compare domestic against offshore sourcing rose from 30 percent of OEMs to 40 percent in a year. That is the accounting change P-07 requires. Wage rate comparison sends work to the cheapest hour. Total cost analysis is what makes a nearer, faster, more expensive hour defensible on a spreadsheet. Ten points in one year, off a base of thirty, is that mechanism turning over.
Set against it, the survey's most deflating figure. Only 33 percent of manufacturers believe advances in AI and automation will make U.S. manufacturing competitive enough to cut imports by half by 2040. Two-thirds of the people who would have to buy, install, staff and maintain the robots do not believe the robots close the gap on that horizon.
That deserves to sit here without being explained away. This site's thesis assumes a flexible robotic execution layer changes where production can profitably sit. The practitioners are more skeptical of that than the vendors are, and they hold the purchase order. Either they are wrong, which would be the ordinary history of automation forecasts, or the register's timeline is optimistic. This survey cannot distinguish those, and neither can we today.
What this bears on, and what would prove it wrong
P-07 is the claim this note moves most. Its prediction that location is decided by structure rather than wages now has evidence on both sides of the mechanism: the reported benefits are time, distance and risk, and the accounting that lets firms act on those rose ten points in a year. P-06, on the erosion of the low-wage advantage, gets the opposite signal, since 94 percent of lost orders still go on price. P-09 gets a partial and unflattering version of itself: the premium is being paid, and a quarter of the payers are dissatisfied. P-08, on high-wage countries winning back categories, is untouched by a U.S.-only survey and must not be read as supported by it. P-13 gets its clearest evidence yet in "vendor gaps." P-14 gets the two hiring numbers. Nothing here is scored, and the register is untouched.
The falsification conditions. If next year's survey shows satisfaction recovering while price remains the loss reason, the premium product reading is right and the 2026 dip was a cohort effect. If satisfaction keeps falling and activity falls with it, tariffs bought a pause rather than a shift, and the July note was closer to correct than this one. If total cost of ownership adoption stalls near 40 percent, the accounting change is a minority practice rather than a transition. And if that 33 percent on automation does not rise as humanoids actually deploy, the execution layer this site is built around is failing to persuade the people who would have to install it.
One limit, stated plainly. This is a self-selected survey of 249 firms who answered a questionnaire from an organization that advocates for reshoring. The direction of that bias is toward more reshoring enthusiasm, not less. That makes the satisfaction collapse harder to dismiss and the activity increase easier to.
What this is built on
Reshoring Initiative: 2026 USA Reshoring Survey, U.S. reshoring momentum builds despite policy and workforce challenges ·Floor Covering News: Reshoring survey finds U.S. manufacturing momentum building ·Quality Magazine: 2026 Reshoring Survey, momentum builds despite policy and workforce challenges ·Business Facilities: Survey, reshoring interest rises despite cost and labor challenges ·The Fabricator: U.S. reshoring momentum builds despite policy and workforce challenges ·Plumbing Perspective: New survey, U.S. reshoring momentum builds despite policy and workforce challenges ·Global Sources: U.S. reshoring gains face hurdles of policy uncertainty and labor shortages ·Modern Machine Shop: Reshoring Initiative seeking survey responses to shape U.S. manufacturing policy decisions
The survey report is the primary source. The satisfaction, dissatisfaction and automation figures are taken from trade coverage of the report rather than from the announcement page, which does not carry them, and each is reported by more than one outlet. Two further figures circulating in coverage, a 30 percent threshold on losing import bids and a claim about the share of offshore production that would return if skilled labor existed, are omitted here because their provenance between the 2025 and 2026 surveys could not be established.
Add to this
Corrections, evidence, and disagreement are welcome. This is knowledge in the open. Anyone can read; sign in with GitHub only to post or react, and it appears here instantly.