LG's state-of-the-art OLED TV with ThinQ artificial intelligence, bringing the control of fridge, air purifier and robot vacuum cleaner altogether by voice. Photo: Asia Times
LG's state-of-the-art OLED TV with ThinQ artificial intelligence, bringing the control of fridge, air purifier and robot vacuum cleaner altogether by voice. Photo: Asia Times

Manufacturing has been out of fashion in the US as much of it migrated offshore in search for lower costs. But technology, particularly artificial intelligence, is changing that calculation for high-tech industries that can use extensive robotics. This change makes manufacturing closer to home practical for some industries and creates a competitive advantage by allowing a much more nimble business compared with relying on offshore production.

Forget the old image of factories where workers are standing on an assembly line bolting parts together. We are talking of plants run with sophisticated information technology (IT) by highly trained technicians. Robots do the work.

Built around modular units, such plants can be expanded as needed, or partially not used, without impacting the operating modules. The production flexibility is enabled by tooling and software changes with the objective to move products quickly from development to production, thus closely integrating manufacturing with product development, marketing and sales. 

Flexible high-tech manufacturing relies on the creative application of IT through the use of massive timely data and artificial intelligence, robotics, sensor deployment and ubiquitous communications to link the factors bearing on manufacturing. Such plants with suitable interlinked sensors are well suited for a high level of in-process quality control and  documentation. 

The success of such flexible plants depends on a highly trained workforce of technicians able to monitor and adjust such sophisticated systems. Such people need good training. An example of a school offering a two-year program is Indian Hill Community College in Ottumwa, Iowa, which offers training in applied science, optics, electronics and software.

No two plants will be the same because of product differences. However, the concepts are the same for all industries, although a pharmaceutical plant will be very different from a microelectronics plant.

The model that has pioneered this concept is the chip industry. We are benefiting from continuous and remarkable innovations in device manufacturing since the 1960s when a single silicon switching transistor (the core device in computing) cost US$5. Today, built with a totally different technology, a silicon chip the size of a fingernail containing 1 billion interconnected transistors sells for the same price.

The corresponding economics of computing are all around us – a smartphone has more computing power than mainframe computers of some decades ago.

This is the result of amazing innovations (many in the US) in optics, robotics, chemistry and software to integrate the various parts of the production process. Workers are highly trained technicians able to control remotely very complex and costly machines. In fact, robots do the process work because the cleanliness levels are so high that humans cannot come in contact with the production in process lest they contribute dust particles. 

I mentioned the values of close integration of production with product development. My experience with a chip-producing company in the Warburg Pincus portfolio is relevant. This company was a pioneer in producing low-cost microcontroller chips for the consumer-electronics industry such as TV remote controllers and consumer appliances.

The company operated its own production plant in the US and was able quickly to produce products needed by its global customers that required fast response to consumer-market needs. Because of the close integration of engineering design and production, it achieved good sales growth, profitability and advantage over fierce competition.

This came to an end when the company was eventually acquired by another investment group that decided to cut capital costs by outsourcing production to a Taiwanese company. The close functional links were lost and the company lost its nimble response to market needs. As a result, it lost its competitive advantage as customer response deteriorated. The company could not compete profitably and eventually disappeared.

Building flexible high-tech companies is neither easy nor cheap, and capital availability is key. I don’t want to minimize the technical challenges in designing such flexible factories. Designers and technologists have to integrate many different disciplines, design sophisticated software and ensure that the proper parameters are controlled.

They will also discover how difficult it is to scale products from the laboratory to cost-effective production, as I learned from experience. At RCA I was once responsible for developing a new silicon industrial-market transistor. I successfully completed a large number of samples in my laboratory and then proceeded to set up a production line in a new factory.

I set up automated equipment that would replicate my laboratory equipment but at a much larger scale. The plant had to produce millions of devices annually – a far step from the dozens made in my laboratory. I had noted with great care the process steps, and the equipment was set up to replicate these conditions.

Then reality set in – production began but the yields were very low and erratic. As the operation proceeded, the cost targets and volume targets could not be met. I had to determine where the mechanization failed to replicate my successful laboratory process, and this is where I learned the importance of smart production people.

With a selected team of 12 technicians, we systematically checked each process step with careful measurements and compared them to the laboratory data. It took weeks of careful detective work to discover why the laboratory process was not being replicated – the automated equipment was not sufficiently temperature-controlled.

The effort was not wasted. This transistor, the 2N2102 type, is still commercially available decades after its development.

Henry Kressel is a technologist, inventor and long-term Warburg Pincus private equity investor. Among his technological achievements is the pioneering of the modern semiconductor laser device that enables modern communications systems. He is the author, with N Winarsky, of If You Want to Change the World: A Guide to Creating, Building and Sustaining Breakthrough Ventures (Harvard Business Press, 2015).