No Match Found
If your company is like most, it has products and services that depend on semiconductors — and right now, you can’t get enough of them at a reasonable cost. This shortage is, unfortunately, different from ones in the past, as it’s been fueled by unprecedented disruptions in both demand and supply. Sure, we’ve experienced this sort of thing before, but not in such a severe way that’s acutely impacted businesses in nearly every sector.
Systemic solutions may take months or years, but it’s possible to mitigate these challenges in the meantime. Here are some steps that may help you acquire the semiconductors you need, while helping reduce the risks of overpaying or missing out on high-in-demand supplies.
In the not so distant past, semiconductors were a cyclical business, with booms and busts that usually followed a predictable pattern: Tech companies offered a new generation of products that needed new semiconductors. Demand would boom and supply capacity would first struggle to keep up, then overshoot the mark, leading to a glut and price declines — until the next cycle began.
Today’s troubles stem from the fact that semiconductors aren’t just in traditional technology products. They’re everywhere: cars and light bulbs, children’s toys and kitchen faucets.
With so many different products in so many different markets, semiconductor demand is now less cyclical. Even the recent recession led to less of a bust than many had expected. At the height of the pandemic, for example, people bought fewer cars. But work-from-home trends boosted demand for computers, webcams and other technology that rely on semiconductors. The faster-than-expected global economic recovery then further increased demand — and the current supply gap may prove unusually hard to close.
As a highly sophisticated technology, semiconductors have a complex, global supply chain of their own. Of the companies that design and sell semiconductors, some are integrated device manufacturers (IDMs) that also own the fabricating facilities (“fabs”) that manufacture semiconductors. Other, “fabless” companies design and sell semiconductors, but partner with “foundries” to do the manufacturing.
In times of shortage, IDMs may not have sufficient fab capacity to meet demand. Building a new fab — a highly specialized facility — can take as long as three years. A fab requires specialized equipment that’s also caught in its own supply chain woes. Some more advanced equipment, such as extreme ultraviolet lithography tools, for instance, currently take up to three years to acquire. Even after fabs are built with the necessary equipment and software up and running, separate facilities (or companies) are needed to assemble them into device packages, test and distribute them. These different stages are typically performed by different companies in different countries, and it gets even more complicated if devices are assembled from different chips using different technologies or process nodes. This complexity makes it hard to ramp up production capacity in a hurry. It also makes the supply chain vulnerable to disruptions — which current trends may make more common.
Not too long ago, most customers had access to the same semiconductors so long as they were willing to pay for them. Now, some companies are designing their own semiconductors and hiring foundries to manufacture these custom products — for their own use only. Other companies are signing long-term agreements with semiconductor suppliers, implying that even many off-the-shelf products also may never reach the market. Geopolitical concerns and the rise in “semiconductor nationalism,” with countries seeking to boost domestic production, may further complicate supply.
Only a few companies make the tools and software needed to fabricate the highest performance semiconductors, and some components (such as silicon wafers and substrates) are in short supply.
Further, lower-end semiconductors (“lagging nodes”) common in some products (for example, those with long development lead times like automobiles) can also face supply challenges. Semiconductor companies aren’t generally eager to invest huge sums to ramp up production for lagging process technologies. These component shortages have been further exacerbated by the conflict in Ukraine, given that a significant percentage of neon gas and palladium required for semiconductor device manufacturing is sourced or processed in Russia and Ukraine.
If your company buys semiconductors, it will likely continue to be a challenge to acquire the supply you need at reasonable prices. But it’s not insurmountable. We’ve drawn on both leading practices and new technologies to develop steps you can take to help reduce semiconductor supply chain challenges while mitigating risks.
It would be wonderful to know exactly which semiconductors and how many of them your operations will need in coming years, but there’s a problem with forecasts: They’re generally wrong. It’s possible, however, to make your forecasts more accurate. Well-designed artificial intelligence (AI) models, if they have access to the right internal and external data sources, can create detailed forecasts for supply and demand under a variety of different conditions. You can then assess probabilities and risks, considering likely demand and supply factors in multiple scenarios.
With better forecasts for a variety of scenarios, you can help enhance sales and operational planning, but you’ll need more than AI to execute on these plans. Consider clearly defined roles with lower management empowered to handle routine challenges and a protocol to engage senior executives when mismatches in supply and demand cannot be resolved. It’s essential to fully integrate sales, finance and operations teams for alignment and joint ownership of decisions. Consider, too, regular reexaminations of supply and demand for tactical (e.g., quarterly) and strategic (e.g., 12- to 18-month) time frames.
One of the latest emerging trends is larger companies bringing their semiconductor design in-house, then contracting out manufacturing. Electronic design automation (EDA) tools, along with the ability of many foundries to support their customers’ design efforts, could put your company in the semiconductor design business. You could then acquire chips that do exactly what your products need them to do, while cutting out costly and power-draining capabilities that your products won’t use.
In-house design has its own challenges, starting with cost. You may have to open up your checkbook to compete for scarce and expensive talent. If large quantities of differentiated semiconductors are critical to your core business, this investment could pay for itself. But you’ll also need new management and partnership skills. Even with your own design team in-house, you’ll still need to navigate and negotiate the semiconductor ecosystem, including not just foundries but also IP providers, EDA tool companies, and assembly and test houses.
Odds are, your company no longer has an option that it once had: of treating semiconductor vendors as commodity suppliers that could accommodate new or canceled orders at the last minute. Semiconductors have become too business critical for you to leave your supply to the mercy of the latest market trends. Instead, you may need to move from transactional relationships to partnerships with both semiconductors vendors and their distributors.
You could, for example, share your forecasts and demand scenarios with a few key vendors. You could then work out a plan to share the costs and risks involved in producing the supplies that you’ll most likely need. If your semiconductor needs are both business critical and foreseeable, you may wish to sign a long-term purchasing agreement. You could also consider collaborations that go beyond purchasing agreements. If you present a concept, for instance, many semiconductor companies can work with you to incorporate your needs into the design and mass production of their chips without requiring you to build up a costly in-house design team.
It’s true for any supply chain, but especially so for those with long lead times: If you and your suppliers can know in advance what you need, you’re more likely to get it. With semiconductor companies investing far more in leading-edge devices than lagging ones, you may also benefit from quickly rolling out new, higher-margin products that use these leading-edge devices.
To help enhance product development velocity and predictability, consider a product and cycle time excellence framework with key elements that include cross-functional executive oversight to allocate budget and people. You’ll then need cross-functional core teams to execute projects quickly, without different functions and lines of business crossing signals. You also want to consider a structured, consistent development process with formal reviews from your cross-functional teams at specific milestones as well as clearly defined goals and accountability.
Semiconductor companies are working hard to ramp up production and, in coming years, shortages in different parts of the semiconductor market may be gradually resolved. Government measures, such as the Creating Helpful Incentives to Produce Semiconductors (CHIPS) for America Act, could also help — and companies may wish to lobby policymakers to support such measures.
But there’s no need to wait for broader market trends to come to your rescue. With the right technology tools and leading industry practices, you can better forecast semiconductor demand and supply, make thoughtful decisions about how much to do yourself, build mutually beneficial partnerships, and increase the velocity and predictability with which you roll out new products that use semiconductors. Together, these steps can help you navigate today’s semiconductor shortage and prepare for tomorrow’s supply scenarios, too.