The winners will be companies that design for flexibility, treat AI as core infrastructure, and build globally intelligent supply networks that let them keep making things no matter what the world throws at them.
For most of the last five years, reshoring has been one of the most energizing words in American manufacturing—and the momentum behind it is real. After the pandemic and two years of tariff whiplash, companies are moving with genuine urgency to build more of what they need closer to home, and that resurgence is opening opportunities we haven’t seen in a generation. What excites me most is where it goes next. The leading manufacturers are rethinking how products get designed, sourced, and built from the ground up—and engineers are right at the center of that shift.
Success in 2026 means moving past ‘reshore everything’ toward something more durable: supply chains built for resilience through flexibility and diversification, with intelligence layered on top. That pattern is exactly what we found in our 11th Annual State of Manufacturing & Supply Chain Report, a survey of more than 300 senior supply chain and manufacturing leaders across sectors including MedTech, EV, robotics, and climate tech. The goal is to keep building resiliency into sourcing decisions.
From reshoring to diversification
The macro backdrop is genuinely encouraging. U.S. policy has tilted decisively toward domestic investment: the recent tax package restored full expensing for new equipment and made R&D incentives permanent. Semiconductor commitments crossed half a trillion dollars as of last year, per Deloitte’s 2026 Manufacturing Industry Outlook. Apple is already assembling AI servers in Houston—hardware it once built overseas—running ahead of its own schedule, with mass production slated for 2026. The momentum is real, and our own data reflects it—93% of the leaders we surveyed named moving production back to the U.S. a top priority.
There are good reasons to believe reshoring to the US is a smart move. U.S. manufacturing reached a record $2.91 trillion in nominal value in 2024. At the same time, a projected 3.8 million skilled-worker gap threatens the next decade of growth.
So, while the US is a good bet, shortages in workers underscore the importance of diversification. In our State of Manufacturing Report findings, 81% of leaders want to increase U.S. manufacturing and 59% want to grow North American production—yet 49% still plan to diversify their global operations, and those goals are not in conflict. The United States ranks as the top preferred sourcing region at 89%, but Canada (47%), Mexico (39%), and the European Union (36%) all feature prominently in the same strategies. Friendshoring, nearshoring, selective domestic production for mission-critical parts, and continued global sourcing where it still makes sense are coexisting inside one plan.
The share of leaders who see geopolitical instability is also a significant factor in long-term supply chain strategy. It jumped from 51% in 2025 to 71% in 2026, and a striking 99% now consider a supplier’s tariff and trade expertise essential when choosing a partner. That is why sophisticated buyers have largely abandoned unit-price comparisons in favor of total cost of ownership—factoring in tariff exposure, lead-time variance, quality yield, and the option value of being able to switch. On a spreadsheet, the cheapest part often stops looking cheap once you price in the risk of not getting it. Yet 81% told us that supplier sourcing and manufacturing remain too time-consuming and costly, and 77% said trade compliance has grown too complex to manage without outside expertise. The demand for resilience is outrunning the tools most teams have to deliver it.
AI is becoming the operating layer
That gap is precisely where artificial intelligence is moving from pilot to implementation. A diversified, multi-region network generates far more decisions—more suppliers to weigh, more routes to model, more scenarios to run—than any procurement team can process manually. That load is exactly what machine intelligence is good at.
In fact, 95% of leaders now say implementing AI is vital to their company’s future success, and 97% say it is already embedded across core manufacturing and supply chain workflows. A year ago, AI was framed as transformational; now it is treated as foundational, and the differentiator is no longer whether you adopt it but how fast you deploy it. Ninety-five percent also said AI and automation are helping them address workforce shortages—while being clear-eyed that these tools augment specialized expertise rather than replace it.
The returns are no longer speculative. McKinsey’s analysis puts AI-driven logistics cost reductions in the 15–20% range and inventory reductions anywhere from 10% to 35%, largely through better demand forecasting and real-time management. For engineers, the more interesting shift is where the intelligence is being applied. We are past dashboards that merely describe what happened. The frontier now is agentic systems that act—reordering inventory against a forecast, rerouting freight around a disrupted lane, flagging a supplier whose financial signals suggest trouble before a line goes down—and digital twins that let a team pressure-test a plant or a network against disruption before committing capital. Nobody who is serious about their business is handing the whole supply chain to an autonomous agent yet, and they shouldn’t. But the direction is clear.
Design for scale before you design for the prototype
Here is where I want to speak directly to the people who actually make things, because this is the part the conversation usually skips. Resilience and AI-enabled sourcing only pay off if the product was designed to take advantage of them—and most scalability problems are baked in at the CAD stage, long before a single supplier is contacted.
I have watched too many promising products stall in the gap between a working prototype and a manufacturable one. A part gets designed around a tolerance a single shop can hit, or a custom component only one vendor stocks, or a geometry that is trivial to 3D print in a lab and a nightmare to injection-mold at volume. Each is a decision that quietly forecloses your options later. When demand arrives—or your primary supplier lands on the wrong side of a tariff line—you discover your design has no second source because it was never engineered to have one. Our data shows how expensive that drag has become: 83% of engineers now spend four or more hours a week on procurement-related workflows, and 93% of leaders say productivity climbs when that administrative load is offloaded. That is engineering talent spent chasing parts instead of designing them.
Getting supply chain into the conversation early—before architecture, tolerances, and materials are locked—is how the best programs now run. It means asking hard questions during design review, not after: Can this be made by more than one supplier on more than one continent? Are we specifying standard, off-the-shelf components where function allows, and reserving custom work for where it truly differentiates? It also means leaning on the digital tools that make flexibility practical; 97% of leaders in our survey called digital manufacturing platforms essential to production, because they collapse the handoffs between design intent and a manufacturable part. Diversification is far cheaper to build into a drawing than to retrofit into a mature product.
The infrastructure builders are the winners
There is a pattern worth naming in all of this. When manufacturing surges, the most reliable beneficiaries are not always the brands whose logos end up on the finished goods. They are the companies supplying the picks, shovels, and increasingly the intelligence: the automation vendors, the component and tooling suppliers, the platforms that connect design to production.
The numbers bear it out. North American robot orders reached roughly $2.25 billion in 2025 by the Association for Advancing Automation’s count, and the U.S. factory automation market is tracking well past $170 billion this year. Yet by recent estimates something like 80% of U.S. factories still operate with little or no automation. That is not a saturated market—it is an enormous runway, especially as reshored and newly built plants come online, because the fastest way to make a domestic factory competitive against lower-wage economies is to automate it from day one. The demand is concentrated exactly where you would expect: among the leaders we surveyed, 90% of EV companies, 87% in climate tech, 82% in MedTech, and 69% in robotics are moving to expand U.S. production. Reshoring and robotics have become two sides of the same coin.
That intersection—physical automation hardware meeting AI-driven decision-making, precision components meeting on-demand manufacturing capacity—is where I spend my days, and it is why I am optimistic. The next phase of American manufacturing will not be won by whoever repatriates the most production. It will be won by the companies that design for flexibility, treat AI as core infrastructure, and build the resilient, globally intelligent supply networks that let them keep making things no matter what the world throws at them.
