The pandemic has highlighted the challenges of relying on goods manufactured thousands of miles away from the U.S. by a human workforce. It has accentuated the collision of existing cracks in the supply chain and the immediate delivery online shoppers have become used to.
Because we also have ongoing ambiguity around sanctions on commerce between the U.S. and China, we are starting to see a return of the manufacturing of certain goods back to U.S. soil.
The rise in U.S. manufacturing will be characterized by three elements:
Sustainability: Using less energy to create products and creating less waste in the process.
Customization/personalization: U.S. companies are unlikely to compete on a pure cost basis. But I believe users will pay more for goods that have their own personal touch.
Localization: This is similar to the farm-to-table movement that promotes serving local foods at restaurants. It is an interesting parallel, yet I don’t see it playing out quite to that degree.
Implemented correctly, this will involve establishing local (and far more sustainable) production facilities that are adept at adjusting what they make.
It will require a combination of human workers and a new type of robot. Instead of very large, expensive robots that can create specific goods in extremely high volume, the new wave of robots will be smaller, less costly and easy to reprogram.
The push to drive down cost, power and footprint is causing companies to implement what was previously multiple discrete subsystems down to a single board that incorporates multiprocessor technology.
One of the more interesting recent trends we have picked up on is for OEMs to download their “recipe” for creating a widget into the manufacturing system. The manufacturer provides a secured virtual machine for the OEM to run its code from. The manufacturer must be certain that this code cannot deliberately or accidentally cause the factory equipment to misbehave and put humans in danger.
The software architecture for the system must be able to isolate the core functions that run the machine, away from the applications, with the latter being unable to directly access those resources.
In this new world, though, it is not a case of robots versus humans. We will see humans working alongside robots, with the outcome being a significantly better result for the factory.
This is less about the value robot technology offers and more about getting away from the thought of replacing humans with robots. We need to look at the complete task holistically and partition it between a robot and a human.
Adjusting robots for new process flows is challenging, and downtime impacts factory output and effectiveness.
For all the claims of artificial intelligence (AI), only a human can predict problems before they occur and think on their feet to make decisions in real-time.
As opposed to a “Cambrian explosion,” which some journalists have used to describe the onslaught of AI solutions into the workplace, actual adoption will be slower.
Businesses are starting to recover from the pandemic—some more than others—and the need to deliver a seamless experience from online to the store (in order to compete with giants like Amazon) is requiring a rethink of entire supply chains.
This will be slow and steady. A piece of equipment that promises to accelerate functions in a factory does not aid the overall plant if it is unreliable, inflexible and/or costly.
But we are seeing well-thought-out proofs of concepts and trials of technology to validate the return-on-investment benefits before broader deployments are forthcoming.