Robotics offers route for US manufacturing renaissance
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
Aindrea Campbell knows more than most about high-tech production. In her previous role, she was senior director of iPad operations at Apple, helping to run the sophisticated assembly lines in China that produce tens of millions of the tablet computers each year.
But, now, as chief operating officer of Agility Robotics, Campbell will oversee the production of pioneering products in the US. In September, the company announced that its 70,000 sq ft RoboFab — the “world’s first factory” for building humanlike robots — would be located in Salem, Oregon, less than an hour’s drive from the start-up’s headquarters in Corvallis.
“What’s most attractive is being close to the engineering team — having that connection with engineering, so we can address issues and gain learnings, and really cycle the product through our production development process much faster,” says Campbell, who took up her role in January.
Within the next 12 months, Agility plans to start building its Digit model bipedal, autonomous robots, which are designed to handle warehouse and logistics work alongside human employees. Starting in the hundreds, the factory has capacity to produce 10,000 Digits a year, to work on the assembly lines of Fortune 100 companies.
The decision to build Digit at scale in the US might not have been fathomable two decades ago, and not just because robotics was not so advanced. These days, however, the country is once again being seen as a viable option because, as Council on Foreign Relations scholar Shannon K O’Neil puts it in her book The Globalization Myth, “many of the economic, technological, and political factors that once gave edge to the far-flung factories have begun to fade”.
Manufacturing wages in China have jumped tenfold since 2001. Shipping disruptions amid the pandemic have also compelled western companies to shorten their supply chains and de-risk from China. And boardrooms are now prioritising actions to mitigate climate change.
Vietnam, India and Mexico are among those primed to capitalise, especially in labour-intensive areas. Their opportunity is vast: there are $4.6tn worth of shipped goods that could “shift across regions” by 2026, according to consultants McKinsey.
The US is unlikely to compete in sectors such as high-volume electronics, because of higher wages and a widespread skills shortage. But advances in 3D printing and artificial intelligence are helping to build the pathway for a US manufacturing renaissance based on robotics and automation.
“The US is not going to build the same type of manufacturing China has built for the last four decades,” says Lior Susan, partner at Eclipse Ventures, which backs physical industries at the forefront of what it describes as the next industrial revolution.
“You can’t put 80,000 people on a campus in the US. That’s just not going to work,” he explains. “But we’re going to use technology — a lot of automation, a lot of software — and do manufacturing in Charlotte, Cincinnati, Salt Lake City and outside of Boston.” He points to President Joe Biden’s $1tn infrastructure bill and the Department of Energy’s $400bn clean technology loan programme as signs of momentum and bipartisan support.
The trend is being reinforced by a spate of start-ups developing “enabling technologies” to help factories run more efficiently.
One is Palo Alto-based Instrumental, which uses AI-enabled cameras and sensors to help factories cut down on waste and increase their manufacturing yield — the percentage of goods that are perfect. Its chief executive, Anna-Katrina Shedletsky, says clients are moving roughly 10 per cent of their production from China each year and rethinking every facet of their operations.
“Everybody that is building physical products has realised in the last two to three years — if they didn’t realise before — that the way that they’ve been doing it for the last two decades needs to change,” she says. “They are essentially restarting from a blank sheet.”
Another is Fictiv, in San Francisco. It connects different factories’ idle machines — remotely — to work as one digital contract manufacturer.
“We basically take the factories that are at 20-30 per cent utilisation, and we bring them up to be more like 50-60 per cent,” says Dave Evans, co-founder. “Or we take the factory that’s overwhelmed, and we redistribute that work to help them survive.”
VulcanForms, a Massachusetts-based start-up, last year raised $355mn to pioneer 3D-printing of metal parts at industrial scale. John Hart, chief executive, says its machines can be quickly reformatted with just a few lines of code, to build for different industries.
“So, in one factory, you can be making components for jet engines and rocket engines, consumer devices or medical implants,” he says. “That’s never been possible before. Think of what that means for managing supply chains, for restructuring the relationships between organisations.”
Industry experts say Elon Musk’s SpaceX and Tesla have played catalysing roles over the past decade, demonstrating that advanced manufacturing can be executed in the US. Alumni of the two have founded 154 companies, raising $18.9bn, according to Rick Davis, a climate tech investor who produces a database tracking launches.
Venture capital, however, still flows overwhelmingly into software, IT and cloud infrastructure, and fintech, rather than heavy-duty machinery and automation. According to advisers EY, just 2 per cent of $44bn raised by VCs in the first quarter of 2023 went to industrials.
For US manufacturing to thrive, Eclipse’s Susan says VCs need to re-evaluate their investments. “Nothing against crypto, but most of the world’s GDP is in industry,” he says. “VC is not funding any of that — and it’s a big mistake.”