The expanding $280 billion The CHIPS and Science Act is finally here, with applications for open funding February 28. US president Joe Biden has already claimed the Act as one of his signature accomplishments, giving it away main place in his recent State of the Union last month. And Washington isn’t the only one throwing money at semiconductors: The European Union, Japan and India have also launched their own semiconductor subsidy programs.
So now that governments are pledging hundreds of billions of dollars to boost their chip supply chains, could we see another chip shortage in the future?
The bad answer is: Yes, we do.
Chip shortages are the result of supply and demand mismatches that cannot be quickly addressed by chip manufacturers by increasing production or by markets by adjusting the chip production profile. The challenge of addressing both is not going away—and may even grow in size.
Semiconductor demand is unpredictable. The consensus is that future chip demand will be driven by AI, electric and autonomous vehicles, the Internet of Things, and 5G/6G. However the exact nature, speed and magnitude of the increase in demand is not yet known.
we cannot predict what kind of AI will dominate the next few years. The development of autonomous vehicles is slower than predicted. 5G deployment rates are very high affected by geopolitics—and the future of 6G is the same it’s darker. Entirely new product areas may emerge: Look at Bitcoin mining, which no one can predict will be a key driver in semiconductor demand.
This uncertainty makes demand planning difficult for chips of different types—logic vs. memory, digital vs. analog, high voltage vs. low power, etc.—and raises the specter of shortages. .
Supply is also unpredictable. Chip factories are typically run at near full utilization. That makes chip fabrication notoriously vulnerable to natural disasters (earthquakes, floods, etc.), accidents (fires, power outages, etc.), sabotage, and geopolitical and military crises (such as tensions between the US and China, or between Beijing and China.Taipei, or even in between Japan and Korea)
In theory, sufficient redundancy and excess capacity can overcome these problems. But the amount of investment required to build a large enough buffer can be very high. It will also take years for capacity building, in technology and human resources, to support this investment. And chip manufacturing is a low-margin business: It can’t sustain excess capacity or change for long.
Nor can we assume that the current investment in the chip will inevitably succeed. The economy of chip production is CRUEL, and most countries and companies cannot make the one-time and repeated investment required to support chip manufacturing, especially cutting edge. In addition, most of these investments are easily mitigated by cheap chips from regions of the world that benefit from large subsidies, cheap labor, and tighter regulations.
Because of this, only a small number of winners will emerge. Countries and companies with failed investments will resist future investments, again creating the possibility of a supply shortage. The political will for future investments may also disappear as officials realize that chip manufacturing uses fewer workers than they expected (especially with advances in automation).
Current spending is primarily focused on manufacturing, just one part of a very complex semiconductor supply chain. Some aspects of the supply chain can be a bottleneck.
Packaging, for example, has been restricted. After the chips are manufactured, they are placed in a protective casing (“package”) for reliability, which also provides an interface to communicate with other chips. Recent supply chain difficulties and increased chip demand have led to shortages of the necessary equipment and upstream materials needed for packaging, increasing lead times for most types of packaging. for many months.
It is also becoming more difficult to solve shortages when they occur, and it is more expensive to increase production with advanced technology every year. That means longer and deeper disruptions.
It is even difficult to increase the production of less advanced chips. Today’s cheap semiconductors are made using advanced equipment that has already been paid for, meaning pure profit for the manufacturers involved. But anyone thinking of expanding the production of less powerful semiconductors will be stuck trying to pay for more expensive tools and machinery with a cheaper product.
The increasing complexity and reach of the semiconductor industry means that supply and demand mismatches are becoming more common. And the high cost of entry and scaling of productions will make it difficult to solve these problems, even in an active way.
That means any company involved in semiconductors must be prepared for chip deficiencies to become the new norm. They can create a buffer through stockpiles or have multiple suppliers. They can monitor the supply chain, ideally in real time, to predict and plan for disruptions. And they can prepare management and engineering methods to quickly adapt to changing conditions.
That’s all easier said than done, and these decisions have trade-offs in terms of cost and efficiency.
But the cost of doing nothing is higher. Finding the right balance between competitiveness and stability should be the goal for any company that uses semiconductors.
Rakesh Kumar is a Professor in the Department of Electrical and Computer Engineering at the University of Illinois and author of Reluctant Technophiles: India’s Complex Relationship with Technology.
The opinions expressed in Fortune.com Commentary pieces are solely the views of their authors, and do not reflect the opinions and beliefs of Fortune.
Learn how to navigate and build trust in your business with The Trust Factor, a weekly newsletter that explores what leaders need to succeed. Sign up here.