Three predictions for manufacturing in 2026.

We talk to roughly 40 mid-market manufacturers a month. The conversations got noticeably tighter in November and December. People are planning. The plans are not adding up.

Three things will define 2026 for shops in the 50 to 1,500 employee range. None of them are speculative. They're already happening. The question is how each shop responds.

1. Reshoring demand will keep arriving. Back offices will not keep up.

The reshoring conversation in 2024 was abstract. The conversation in 2025 turned operational. In 2026, it becomes a capacity question.

The numbers from Deloitte's 2026 outlook are concrete. 38% of manufacturers are actively moving production back to North America. Another 27% are in advanced evaluation. Defense, aerospace, industrial components, and specialty chemicals are all seeing sustained inbound volume from OEMs that spent 2022 to 2024 building contingency supply chains and are now committing to them.

The shop floor side of this is well understood. New equipment investment is up, capacity utilization is up, and most operators have a clear plan for absorbing more production.

The back office side isn't.

Jeff runs a 250-person contract manufacturer in South Carolina doing precision machined components. His RFQ volume is up 40% year over year. His three estimators are handling the work of five, and they're losing ground every week. He's been trying to hire two more for eight months. He cannot find them.

This pattern is going to break a lot of shops in 2026. Not the floor. The front office. The work coming in will outrun the capacity to quote it. The shops that can't scale response without scaling headcount will turn away business they're qualified to handle. The shops that can scale response will pick up that business.

The planning implication: every operator should look at their RFQ volume run rate and ask the honest question. If volume grew another 30% this year, would your back office handle it? If the answer is "we'd hire more people," start the second question. Can you actually hire them, at the rate you'd need, in your geography? If both answers are no, the constraint is back-office throughput, and headcount alone won't fix it.

2. Tariff volatility is now a quarterly event. Repricing speed is a competitive moat.

The effective U.S. tariff rate sits at 13.7%, the highest since 1946. That's the headline number. The operational reality is harder to summarize.

What manufacturers will deal with in 2026 isn't a single tariff shock. It's a continuous flow of changes. Rates get adjusted. Exemptions get granted, then revoked. Trading partners retaliate and renegotiate. Companies with global supply chains will be running repricing cycles four to six times a year, every year, for the foreseeable future.

The CFO who used to update master pricing every January now updates it every March, June, September, and December. Every quarter triggers a cascade. Identify affected SKUs. Calculate cost delta. Update price lists. Communicate to customers. Reissue quotes for open orders. Each cycle takes weeks of distributed work across sales, ops, and finance. Each cycle interrupts the routine work that should be happening in the same period.

The shops that can execute a repricing in days, not weeks, will hold customer relationships under pressure. The shops that can't will lose business at each cycle, because their slower response will read to customers as either disorganization or hidden margin grab.

The planning implication: build the repricing process as a permanent capability, not as a quarterly fire drill. Know which SKUs are exposed to which tariffs. Have customer-specific pricing in a system that can be updated programmatically, not in PDFs in email threads. Have communication templates ready for the most common scenarios. Velocity is the strategy.

3. Mid-market manufacturers without AI will lose ground to bigger competitors with AI.

This is the prediction most operators want to argue with. Everyone is tired of hearing about AI. The hype has been overcooked for two years. The vendor pitches are exhausting.

The reality is that something changed in 2025. AI in manufacturing back offices, the boring kind, the kind that reads RFQs and pulls pricing and drafts quotes, started actually working. Not in demos. In production, at real shops, on real volume. Lawton Standard, Excal Foundry, Steel-Con, Premix, and dozens of other mid-market operators went live with AI Inside Sales Reps in 2025 and are running them in production today.

The gap between the early adopters and the rest of the market is now visible. The shops with AI in the back office are quoting in 4 hours and closing 35% of inbound RFQs. The shops without are quoting in 4 days and closing 20%. Over a year, that gap compounds into 30 to 50% revenue differences on identical underlying capability.

In 2026, that gap will widen. The 2025 early adopters are already on year two. Their pricing logic is captured. Their customer history is in the system. Their quote velocity is structural, not project-based. They're going to take share from the shops still operating the old way.

The planning implication: this isn't a 2027 problem. The mid-market manufacturers who don't have a path to AI in the back office by mid-2026 will be playing defense against competitors who do. The question is no longer whether AI is real for this segment. The question is whether your shop is going to be on the front foot or the back foot when the customers start asking about quote velocity.

What 2026 won't be

This isn't a year where the labor market loosens, where reshoring slows, where tariffs settle into a stable regime, or where AI fades from the conversation. The trajectories are clear, and they all push in the same direction.

The shops that came into 2026 with a plan to scale their back office without scaling headcount will be fine. The shops that planned to hire their way through this year will spend most of the year discovering they can't.

Start there.