Preparing the semiconductor industry for a circular economy

Much of the talk about the circular economy (CE) revolves around only a part of that circle, and for good reason. In the Semiconductor and so many other industries, the greatest and most obvious hurdles to CE involve developing the standards and the major physical and data infrastructure needed to enable the reuse, repair, and recycling part of the equation.

Preparing the semiconductor industry for a circular economy

(Source: Shutterstock)

The explosion of the internet of things has added urgency to these discussions: By the middle of this decade, an estimated 27.1 billion connected IoT devices, from smartwatches to refrigerators to light bulbs, will be in play. They will add to the mountains of e-waste that’s already the world’s fastest-growing waste stream.

That infrastructure will be indispensable, and it’s a safe bet that laws and regulations devolving from the Paris Agreement and future treaties will demand it. But there’s much we can do in the meantime, particularly in the realm of data infrastructure. Before touching upon two key aspects of that digital infrastructure, a few words ON two ways in which the IoT revolution may, fortunately, be at least partially self-braking.

First, on the consumer side, unlike the products most obviously associated with semiconductors (PCs, tablets, and smartphones, primarily), embedded devices are often power products with longer life expectancies — thermostats, fridges, and light bulbs (a smart LED light bulb burning three hours a night is rated to last 23 years). Like so many Teslas, virtual “product” upgrades via over-the-air software updates can substitute for product replacement. The longer the product lifetime, the fewer replacements needed and the less waste CE must deal with.

Second, on the B2B side, the growth of everything-as-a-service (XaaS), outcome-driven business models should also extend product life cycles. While this also applies in the consumer realm (users of cloud-based data backup don’t know or care how old the servers and drives storing their data are, for example), for businesses, the bottom line is whether you are satisfying the service-level agreement. Older, less costly hardware coupled with occasional software updates can slow the replacement cycle.

Selling fewer chips due to the longevity improvements of existing products may sound like a losing proposition for semiconductor makers. But CE can be expected to become interwoven with carbon-reduction efforts, and there should be incentives across the value chain to do well economically by doing good environmentally. This will open the door to new business models to harvest one’s proper share of the benefits accrued. Those business models will depend on precise, secure information derived from data available throughout the supply chain. And that brings us to two key steps semiconductor makers can start taking now to position themselves for the CE future.

These steps, both involving data infrastructure, not only position the industry to tackle today’s chip shortage but establish a basis for the sort of detailed carbon tracking that emissions regulations now taking shape around the world will require.

The first is about upgrading from the peer-to-peer systems the industry has relied on for a half-century. The second harnesses blockchain’s immutable ledgers and other technologies to enable the tracking and tracing needed to establish reuse potential, recyclability, and end-of-life disposition.

First, while RosettaNet has been a valuable enhancement of aging EDI, it is overmatched by today’s challenges, not to mention those that await. For one thing, EDI is too expensive for the small companies that play such vital roles in this industry. For another, it’s a one-to-one system in an n-to-n business.

Consider the chip shortage. There’s untapped production capability in the form of devices that, while too imperfect for their slated uses, would be perfectly adequate for lesser applications (say, a flawed batch of aerospace-grade chips that could instead populate a consumer drone). If a fab can communicate with players up and down and fan out across the value web, the chipmaker and the drone producer can connect quickly, and a wafer that produced only 800 of 1,000 hoped-for aerospace-grade chips could yield the full 1,000 in a hierarchy of quality tiers. That, in turn, takes pressure off the nominal producer of the drone chips — and the industry at large. In aggregate, such yield gains could delay or even obviate the need for new multi-billion–dollar fabs.

Conversely, if there’s a yield bust in which a wafer falls far short, the aerospace firm can recognize that instantaneously rather than waiting for a series of one-to-one EDI updates to percolate through the system — as can the drone maker, which can enjoy an even greater haul.

The European auto industry is developing a network worth emulating: Catena-X, which aims to develop secure, continuous data chains with multiple partners across a multi-tier supply chain for a common value-creation process.

EDI’s successor in the semiconductor industry must be about more than just messaging. It will involve analyzing the massive amounts of data churned out by the manufacturing and testing of semiconductors and sharing that data across the value chain so the semiconductor supplier, the OEM, and the parties between and beyond the two can identify opportunities to improve quality, yields, designs, reuse, and proper end-of-life disposition.

Second, CE will call for track-and-trace capabilities well beyond those of current approaches. A company aiming to repurpose an old semiconductor will need to understand its capabilities and configurability; a recycler needs to know its composition. Climate-related regulations will require a precise understanding of energy inputs and outputs for carbon-footprint calculations up and down the supply chain. Extended producer responsibility comes into play here, too, to ensure that a product and its semiconductors is properly disposed of at end of life.

The immutable ledger of blockchain appears to be well suited to this application. There’s also a clear near-term benefit: Digitally fingerprinting devices using blockchain could foil counterfeiters, solving both a business problem and, when you have counterfeit chips going into medical devices and airbag deployment systems, a public health issue.

Driven by a combination of environmental, regulatory, and competitive pressures, CE is coming to the semiconductor industry. While the chemical and other semiconductor recycling infrastructure we’ll ultimately see may be a few years away, work on the needed information infrastructure can start now. Such investments will ensure resilience and profitability while positioning semiconductor firms to tap into new revenue streams as the circular economy takes shape.