As many of us in this industry know, our devices are powered by software and that software is powered by semiconductor chips—and it is predicted to be a trillion-dollar industry 2030, driven by demand for higher quality products and digital experiences in our lives.
This is a conversation I recently had with Tim Ryder, general manager, U.S. Manufacturing Sector, Microsoft, on The Peggy Smedley Show about the state of the semiconductor sector and how the industry is always evolving to power the continuous connectivity that we demand in our lives.
“They are always under continuous pressure to innovate around power, performance, and cost. That pace of innovation is increasing in an ever-more complex environment,” he says.
We took a few minutes to look back over the years and discuss how the industry evolved from the early days of computing until now and trends have emerged.
“The 1990s was really the era of the mainframe…. early 2000s, we really started to see the client-server architecture merge with the use of PCs,” he says. “By 2010, everybody has a smartphone, and we are starting to use more and more online services in our daily lives. We are used to seeing these 10-year increments of change and growth in the industry and for the last two-to-three-to-four-to-five years, we have really seen that advance in a big, big way and that has been fueled in part by demand for more cloud services.”
From that conversation, we unpacked some significant trends that are impacting the semiconductor sector. Perhaps it might help to review each of these points a little bit more now.
Edge computing: We see the trend toward big data centers run by large cloud providers like Microsoft and an ever-increasing demand for more chips to run those big data centers. As such, we are starting to see more proliferation at the edge.
Sustainability in the value chain: Companies are considering how to make materials safer; how to reuse materials; and how to effectively and efficiently use resources like water.
Lifecycle of a product: Ryder suggests there is a big focus on the lifecycle of a product, saying, “Traditionally, you would have a new generation of compute coupled with a new generation of memory. We might have situations where for longer lifecycles we would have a new generation of compute that is able to leverage an older generation of memory, allowing that product to stay out in the ecosystem a little bit longer, extend its lifecycle, and become a little bit more sustainable.”
Supply-chain shock: There is a finite number of OEMs (original-equipment manufacturers). There is a finite number of chip designers, chip manufacturers. He says, “This value chain, this ecosystem, is sensitive, as we saw in the pandemic. Most customers were wanting to run on short supply for efficiencies and cost and we had some forecast hiccups in, let’s say, the automotive industry that then led to the supply-chain shock that we all saw during the pandemic of chips.”
The need for a common data model: There is a greater need today for a common data model that allows for a secure sharing of data across the ecosystem. He explains, “To your point, the more you really understand the use of your product in your customer’s customers’ hands, the more you can work with your customer in terms of where you put your R&D dollars; how you source your raw materials; where you put buffer points in your supply chain; how do you react to changes and consumer taste and things like that.”
Shift in geographies: We are currently seeing a lot of production moving back closer to the customer. He explains, “A lot of domestic policies of various countries are wanting to pull production back closer to the customer for national security reasons.”
Workforce shortage: With the current talent crisis, companies can leverage technology to gain better insights in terms of how their employees learn and collaborate, ultimately improving the habits of high-performing employees.
All in all, these big trends provide an opportunity for manufacturers to think about how we produce products a little bit differently, according to Ryder. With the CHIPS and Science Act, we will see an infusion of funds being invested into this industry, much of which will incentivize the implementation of new fabs in the United States, where we are actually producing and manufacturing silicon. We will also see move investments with academia.
At the end of the day, we will need technology and AI (artificial intelligence) to make better decisions and be more efficient about how we are producing. At this point, the increased demand and the shift in where things will be produced are necessitating it.
Want to tweet about this article? Use hashtags #IoT #sustainability #AI #5G #cloud #edge #futureofwork #digitaltransformation #green #ecosystem #environmental #circularworld