We’ve all heard this quote – or variations of it – “Those who cannot remember the past are condemned to repeat it,” from philosopher George Santayana‘s Life of Reason. But it’s still a lesson the semiconductor industry should take to heart.
We’re currently seeing the digital transformation of everything. And semiconductors – chips that process digital information — are now found everywhere: computers, cars, home appliances, toys, medical devices, industrial equipment, and more. In 2021, semiconductor companies were estimated to sell $600 billion worth of chips globally, according to the Semiconductor Industry Association. And the global semiconductor industry is projected to become a trillion-dollar industry by 2030, says McKinsey.
Chip manufacturers must plan how they will capture shares of the market. One view is to take history lessons from the industry to try and pave a path forward. In this article we will use the U.S. semiconductor industry as a case study and an attempt to illuminate timely learnings for the next business cycle.
Alex Williams and Hassan Khan wrote a splendid article (“A Brief History of Semiconductors: How The U.S. Cut Costs and Lost the Leading Edge”) detailing the semiconductor industry’s story, delving deep into many of the economic and technical issues that have arisen over the years. Here, with a bow of respect for their hard work is a quick summary of U.S. policy and economic conditions over the decades. The goal: to discern any patterns that might be useful for future strategizing.
The 1960s: Industrial policy ruled
Industrial policy refers to enacting government initiatives to support specific industries considered critical to a country’s economy and security. The United States pursued an ambitious industrial policy strategy in the 1960s for the semiconductor industry, pouring money into research and manufacturing and using its buying power – particularly that of the Department of Defense (DOD) – to steer the industry in a way that drove leadership research and innovation through mandated knowledge-sharing and “second source” contracts. Confident of the demand coming from the DOD, even smaller firms could make investments in capacity and research with low risk.
In effect, say Williams and Hassan, the U.S. government protected the semiconductor industry so it could concentrate on innovation and production, rather than maximizing shareholder value for any one firm.
The 1970s: A golden age for domestic semiconductor manufacturers
Williams and Hassan point out that although U.S. government industrial policy relaxed somewhat in the 1970s, it didn’t suppress overall demand as private sector firms began to ship high volumes of electronics through their supply chains. This “booming” economy in the non-defense sector meant small and large firms coexisted successfully without requiring much government support. As no international competitors had yet appeared, innovations in domestic semiconductor technology blossomed, and innovations in production processes also improved. It was indeed a golden age for U.S. chip makers, despite littler help from government investments.
The 1980s: International firms take over
However, trusting to private market dynamics proved disastrous to U.S. semiconductor firms in the 1980s, when Asian markets such as Japan emulated U.S. industrial policies of the 1960s to swiftly build capacity and fuel innovation, causing the United States to lose both market and technological dominance to firms abroad. This had a dramatic and negative effect on U.S. companies. Many smaller chip firms were unable to survive.
The 1990s and 2000s: “Science policy” becomes the paradigm
The U.S. government responded by introducing asset- and capital-light “science policy” to replace the former industrial policy that created a robust competitive ecosystem with strong supply chains. Instead, science policy meant nurturing public-private partnerships with specific firms, bringing industry research closer to academic research, and dividing research from production so that small firms could go “asset-light” through fabless operations.
The goal of this policy was to minimize research costs – making the United States competitive again globally – while getting the government off the hook for large investments.
Say Williams and Hasan, “In theory, this preserved flexibility for these ‘fabless’ firms to pursue innovative design strategies while minimizing overhead costs.”
A critical aspect of science policy was the drive to downsize employees, as labor accounted for a large proportion of chip unit cost. Thus, in the name of efficiency, there were massive layoffs throughout the industry, through which a lot of important domain engineering and production knowledge was lost.
This strategy proved successful in the short term, with the United States overtaking technology leadership by the late 1990s. The U.S. chip industry successfully innovating while staying competitive amidst overseas firms – without the heavy government investments required by industrial policy.
The 2010s: Cracks begin to appear in the supply chain
The new science policy favored the people doing research and design over production and deployment. The result was the multiplying of “fabless” companies that depended on foreign fabrication plants.
Over time, the U.S. semiconductor ecosystem gradually grew too fragile to sustain long-term competitiveness. Eventually, underinvesting opened up vulnerabilities in both innovation and cost-effectiveness.
For example, the policy of keeping labor costs to a minimum for decades dramatically shrunk the availability of trained technologists and engineers, while the corresponding lack of capital investment in capacity made it difficult for domestic firms to respond to market challenges.
There have been 10 economic downturns in the semiconductor industry since the early 1970s, and increasingly, the semiconductor industry was not resilient enough to cope with them.
Almost all these downswings were exclusively caused by demand shocks that sent prices downward. Each time it happens, the entire market feels it, but manufacturers are the ones left holding excess inventory or are underutilized
By prioritizing research over production, the U.S. government paradoxically stifled advancement, say Williams and Hasan, as “learning by doing” is an important step in pushing the technical innovation envelope. Offshoring production meant U.S. firms lost out on important learnings. To be competitive, semiconductor firms needed more capacity, talent – and money. Bringing us to the passing of the CHIPS & Science Act of 2022.
Read ‘what to do about it’ in Part Two of this article.
Learn how Model N can help your company regain your competitive edge when managing the balance between people, process and technology.
