Why America Must Automate or Stagnate
I spent part of my childhood in China, and later, part of my career working to build factories there. Coming out of this Labor Day weekend, I’ve been reflecting on two things.
First: China’s industrialization and market reforms from the mid-70s through the 1990s raised the standard of living for hundreds of millions of people. While entire careers can be spent on the hows and the whys, it strikes me as a meaningful touchstone as our countries continue moving towards an era of strategic competition.
The second thing is that our own growth, here in the U.S., will depend on doing a few things very well in a short amount of time. And one of those is how quickly and pervasively we can build automation and robotics infrastructure for our own manufacturing base.
It’s no secret that American manufacturing needs a boost. There are major headwinds:
The local networks that once helped small and mid-sized factories modernize have largely disappeared. Financing to expand production is harder to secure. Our pipeline of skilled workers is thin and unprepared; and demographic trends and immigration policy show no signs of a magical “labor boost” coming.
These gaps add up to a simple but serious problem. Smaller manufacturers, which account for nearly half of U.S. manufacturing output, are struggling to adopt the technologies they need to keep pace with oscillating demand. Those are the firms making the parts that get assembled by companies announcing massive manufacturing investments in the U.S. and underpinning our entire industrial base.
In recent months, this tension has played out in political debates and headlines: even if tariffs pushed demand towards American-made goods, without a change in how we do things, American firms may not be able to rise to the challenge. Because at the very moment when reshoring is accelerating and labor shortages are biting deeper, too many of these firms are left without the tools, the capital, or the support to compete.
These gaps are measurable. The rate of U.S. manufacturing productivity growth has been in decline for more than a decade. In fact, since 2009, annual U.S. manufacturing productivity growth has averaged less than a quarter of a percent, far behind Germany, Taiwan, the United Kingdom, South Korea, France, and Italy. And China has out-deployed the U.S. in robots at roughly 4:1, and last year, Chinese firms deployed seven robots for each one deployed in the U.S. Walk into a typical American factory today, and chances are you won’t find a single robot. Roughly 90% of U.S. factories have never automated at all.
In Shenzhen or Suzhou, by contrast, you can feel the urgency of adopting new techniques to build. Automation is messy, sometimes clunky, but the commitment is clear: if a robot could reliably do the work, it was deployed. Walking those shop floors left an impression on me. Robots weren’t futuristic novelties. They were everywhere: welding, palletizing, tending machines. Not because they were glamorous, but because they are necessary.
Why? China is running into the same demographic wall that many advanced economies, including the U.S., have already hit. The country’s working-age population has fallen to about 858 million people in 2024, down from more than 900 million at its peak in 2011, and it is projected to shrink to roughly 700 million by mid-century. That is nearly a quarter of the labor force gone in just a few decades.
In the U.S., filling manufacturing roles has become increasingly difficult for a variety of complex social and economic reasons, including workforce demographics, COVID-19, and evolving expectations around work. The National Association of Manufacturers estimates 2.1 million unfilled jobs by 2030. Our own research, summarized in Automation NOW, shows line workers in “dull, dirty, and dangerous” roles lose an average of 67 days a year to injury. In one Midwestern food plant I visited, turnover in palletizing jobs topped 200%. It’s not that people don’t want to work, they just don’t want to wreck their backs stacking 40-pound boxes night after night. These are not just survey statistics. In research for the book we just published, we heard this over and over again:
One of Formic’s California customers told us, “The workers are just not out there.”
One of our Illinois customers told us that while they could find workers, “very low retention” was their biggest challenge.
One of our North Carolina customers emphasized that it was stuck in a never-ending recruiting/hiring/training/churning loop in an increasingly challenging labor market.
Which brings me back to Labor Day, first celebrated in New York in 1882. The Worker’s holiday was meant to honor the contributions of workers and demand better working conditions, which were particularly harsh in factories. The hours were long, typically ten to twelve hours a day, seven days a week. Working conditions were frequently unsafe and led to deadly accidents. Tasks tended to be divided for efficiency’s sake, which led to repetitive and monotonous work for employees. The very definition of dull, dangerous, and dirty work.
That same year, Thomas Edison’s Pearl Street Station started supplying electrical power and the “Current Wars” began in earnest. Yet, when electricity first reached American factories, it did not spark an instant revolution. By 1900, fewer than 5 percent of U.S. factories had been electrified. Many owners simply swapped out steam engines for motors and were disappointed when little changed in terms of production. That’s much of what we have seen with early attempts to automate factories in the U.S. Change was costly, risky, and complex, especially for smaller firms.
The real transformation only came decades later, when firms reorganized their buildings, their workflows, and their labor around what electricity uniquely allowed. That redesign unlocked the surge of productivity in the 1920s and set the foundation for modern industry.
We are at the same inflection point with robotics. Dropping a robot into a factory does not change much on its own, just as replacing a steam engine did not. The gains come when companies can rebuild processes around robotic ubiquity: flexible deployment, real-time sensing, data-driven decision-making, and people working alongside machines. That requires more than hardware. It calls for the infrastructure layer, software, integration advice, skills development, and financing models that make robotics practical and universal, in the same way power grids and demonstration projects once made electricity universal.
If America gets this right, we have the chance to leapfrog China’s path to industrialization. China’s rise was built on abundant labor and state-driven investment. But our future can be built on markets that deliver robotic capacity at scale. Just as electricity became invisible but indispensable, robotics can become the infrastructure that powers an economy capable of producing more, faster, and better than any competitor still bound to the old model.
Which brings me back to Labor Day. We celebrate the dignity of work, and that legacy and history is bigger than this newsletter. But I believe the best way to honor it is to give workers tools that make their jobs safer, more creative, and more sustainable. And to give American communities the means to make persistent poverty a thing of the past. Robots aren’t about replacing people; they’re about replacing churn, injury, and wasted potential.
That’s why I’m relaunching this newsletter alongside our new eBook from Formic. In this newsletter, I’ll share what we’re learning from factory floors across America — what works, what doesn’t, and why robotic ubiquity is the only path back to competitiveness. You’ll also get some behind-the-scenes content on how we’re building Formic, how our customers are building the products that you know, love, and use, and perhaps some robotic geekery along the way.
Because the choice is becoming simple: we either automate, or we stagnate.