Reindustrialize 2026: Leaving Detroit More Optimistic
I flew back to Bogotá from Reindustrialize 2026 more optimistic about the United States’ chances of successful reindustrialization than I was when I landed in Detroit. That was not the outcome I expected.
The headlines still read like a country running late: supply chain shocks, skilled-trades shortages, permitting delays, China holding export levers on metals the US barely produces. All of that is true. But half a mile from Henry Ford’s first workshop, in a room of 1,500 people across government, defense, manufacturing, academia, and capital markets, a different picture came into focus. There is a broad acceptance across all of these players that outsourcing a vast majority of manufacturing in pursuit of earnings growth was a grave strategic error for the US, and that reindustrialization is the solution.
A strategic error, finally named
I have to admit that reindustrialization felt a bit nostalgic to me. But Reindustrialize 2026 convinced me otherwise.
Across Ford executives, university commercialization offices, JPMorgan capital markets, defense primes, and administration speakers, the same premise surfaced without much argument: outsourcing the vast majority of manufacturing in pursuit of quarterly earnings growth was a grave strategic error for the United States. Reindustrialization is not sentimental. It is a strategic imperative, and this was the first time I had heard the business world, academia, investors, and government aligned on that fact.
That alignment matters. Industrial policy debates used to fracture along predictable lines: free traders versus protectionists, climate versus jobs, Silicon Valley versus Detroit. In 2026, the room felt more like a working session than a trade show. Defense procurement in one track, additive manufacturing demos upstairs, SBA matchmaking for drone reshoring in another. A Harbinger truck was parked near the hotel. The mood was less “bring back the past” and more “we broke something important, and we need to fix it before it’s too late.”
Smart people, deep pockets, and real progress
A lot of very smart people and deep pockets are pointed at the same problem.
JPMorgan talked about $1.5 trillion in security-resiliency capital. The DoW’s Office of Strategic Capital has $200 billion to deploy. SBA lending is at record highs, and is running vertical matchmaking events for defense, drones, and space. These numbers reflect a capital base that has decided manufacturing capacity is a national asset again.
The floor showed what that capital is starting to buy:
Divergent opened Factory 2 in Long Beach with 64 Monolith 1 printers, with a capacity framed around tens of thousands of munitions a year.
Formlabs kept circling back to Ukraine. Max Lobovsky’s Fuse X1 launch (~$100K machine, ~500K small parts per year) sat next to a case study nobody could ignore: Skyfall drones iterating in weeks while US programs measured progress in years. BMW running hundreds of thousands of parts on Formlabs printers because injection-mold lead time dominates unit cost at volumes that would have sounded small on a prime’s deck.
Machina Labs and ntop built a Group 3 drone from requirements to flying hardware in days on stage. Boom is building the Symphony engine in-house after supplier lead times broke them. Harbinger vertically integrated motors and batteries because waiting nine months for a part that prints in a day stops being acceptable once you’ve lived it.
Anduril was the surprise on software. Nine years building Arsenal X inside Lattice, sixteen facilities, product definition wired through to yield. They quoted Roadrunner at ~$150M napkin-to-test against a $250–500M PAC-3 equivalent. Take the ratio with salt. The integration ambition is real.
Ribbon-cuttings were boring this year. Iteration was not. These are not slide-deck companies. They are shipping, iterating, and rewriting cost structures in public. That is the progress that made me optimistic.
Four bottlenecks that still need to be solved
1. Supply chain independence, especially for critical minerals
A Ford supply chain executive on day one put the problem in plain terms: twenty thousand direct suppliers, EV platforms with a third fewer parts than ICE, and still no reliable view past tier one. The line that landed in the room was about data, not dashboards: speed wins if you actually know what your suppliers’ suppliers are doing.
Robert Friedland’s session made the stakes obvious on the materials side. Copper demand will roughly triple while China holds export levers on metals the US barely produces domestically. Scandium is showing up in every advanced manufacturing conversation, whether you planned for it or not. Project Vault, stockpiles, fifty-four countries waving raw-material deals. The policy response is mobilizing. The physical supply chains are not there yet.
Supply chain independence does not mean self-sufficiency in every widget. It means enough visibility and domestic capacity that a geopolitical shock or an export restriction does not halt production.
2. Labor force training
The skilled-trades gap may be the harder constraint day to day. Labor force training showed up in almost every manufacturing panel, usually without anyone calling it a workforce strategy. Companies are vertically integrating partly because they cannot hire fast enough to trust the supply chain they have. Permitting and litigation still slow everything down. But even when the permits clear, the people have to exist.
3. Robotics deployment
The warehouse automation talks were thin. Mining and quarries were not. That matters because robotics deployment will not start in polished final-assembly lines.
I met Maximilian Rolf from Sensmore on the floor. German team, vision-language-action models, full-site quarry automation. Dust, uneven ground, heavy equipment, safety cases written for extraction sites instead of clean rooms. That felt closer to the deployment problem than humanoid demos.
CreateMe automates bonded apparel in San Francisco, twenty times the sewing throughput of conventional lines.
The US will not automate its way out of a labor shortage with warehouse pick-and-place alone. The interesting robotics bets are in unstructured environments where the work is dangerous, repetitive, and already happening at scale.
4. Patience from investors
There is a tension in the room that nobody resolves cleanly. Industrial bets need ten or twenty years. Rubio, Waltz, Ford, First Solar, and the SBA speakers all flagged election-cycle whiplash. Builders respond by chasing commercial revenue first and designing exportable COTS hardware that survives a funding cliff.
Investors need the same patience as the physical build requires. Robotics hardware, mineral processing, factory tooling: these are not SaaS multiples and eighteen-month liquidity events. The capital is showing up. The discipline to stay through the boring middle of a reshoring cycle is still being tested.
China is way ahead and we need even more urgency
Every session that touched geopolitics carried the same subtext: China spent decades building factory clusters, supplier depth, and process knowledge that the US let atrophy. Ukraine and Iran showed up as live labs for iteration speed, but the baseline manufacturing ecosystem advantage still sits across the Pacific.
The Formlabs/Ukraine comparison landed because it was about cadence, not patriotism. Skyfall iterated faster. The US had more capital and still moved slower. That gap should bother anyone who thinks capital deployment alone closes the race.
What this looks like from Latin America
Detroit is not thinking much about LATAM, but it should be for two reasons.
First, LATAM is an important source of the critical minerals that the US needs to rebuild its manufacturing base.
Second, LATAM companies and consumers can be buyers of the products and solutions designed and manufactured in the US. LATAM companies also need to automate, making them potential buyers of the robotics solutions that will drive the US’ reindustrialization efforts. This is what we as güil are excited about; helping robotics companies deploy into the LATAM market.
Leaving Detroit
I came back with a better understanding of the importance of reindustrialization and the bottlenecks that need to be solved; e.g., where we as güil should invest and partner in the value chain.
The US is late, but there is finally a shared diagnosis and many smart people and deep pockets backing reindustrialization. Physical AI will be key to making it happen.
— Drew Hallman, Investment Director at güil