Authored by Ee Huei Sin, SVP and President, Electronic Industrial Solutions Group, Keysight Technologies
Competition and innovation continually force businesses to consume more and more data, but the radio frequency spectrum remains finite. The Internet of Things, autonomous cars, and the rollout of 5G and coming 6G networks are already upon us with enormous high-speed data requirements. In addition, future new applications which have not yet been invented will create a need for greater bandwidth and the implied reliance on advanced semiconductors to support these new technologies. How can technology companies keep up with the demand to innovate amid the instability we’ve recently seen in supply chains?
Today’s technologies are more complex than ever. They must be developed in less time and with fewer people – all while trying to be first to market. Components once considered simple, such as a cellular power amplifier (PA), are now complex integrated modules with 1500 different nodes to simulate and test. What is driving this growing product complexity across industries?
Everything is going or has already gone digital. Twenty years ago, the estimated global data storage was just over 50 exabytes; today, it has increased 140 times to 7000 exabytes. The world is generating more than 1000 petabytes — that’s 1 million terabytes — per day. New devices and business models have emerged, providing more business and personal value as big data and new insights improve outcomes across a wide range of use cases. Five years ago, McKinsey estimated that only 40 percent of businesses were digitized, despite deep penetration into tech, media, and retail. Today, business digitization has increased to 65 percent. This massive growth in just a few years means we are moving a lot of data around on the network, creating the need for higher throughput with less latency, even with the gains of the cloud
And the changes are far from over. The use of distributed processing, with apps leaning on massive data centers for heavy processing loads, is transforming how applications work. Technology now supports our just-in-time needs, from instant voice recognition to real-time translation or turn-by-turn directions. As customized chips and compute engines process more data faster, the use cases and value of artificial intelligence (AI) and machine learning (ML) to control complex systems also increase.
Addressing semiconductor demand
Demands arising from these new technologies like silicon photonics, millimeter-wave (mmWave), and high-power devices have led to new semiconductor manufacturing facilities, or fabs, being planned around the world. This is to build enough capacity for developing these emerging technologies and address the chip shortage that occurred during the pandemic. As chip supply chain shortages are expected to linger into 2024, manufacturers will continue to struggle with supply chain risks.
According to IHS Markit, roughly four in five global semiconductor fabs are located in Taiwan, Korea, Japan, and China. In the U.S., lawmakers and manufacturers are starting to focus on reshoring semiconductor technology manufacturing to the Americas, as confirmed by the recent passage of the bi-partisan, congressional CHIPS and Science Act. For example, Intel is opening a new fab in Ohio, and Qualcomm and Micron are also investing in additional domestic manufacturing capacity.
Most fabs have historically operated at about 80 percent utilization, but supply chain disruptions over the past few years have made supplier pipelines and deliveries less predictable. So, to ensure greater flexibility for manufacturing processes and supply chains, organizations should begin prioritizing partnerships with semiconductor manufacturers to gain access to technologies they lack in-house or invest in developing their own custom design and fab capabilities.
Solving the challenges for future technology innovations not only requires manufacturing capacity, but also engineering breakthroughs in test and measurement processes. Custom-designed semiconductors can enable these breakthroughs, creating new measurement systems to test use cases that require greater spectrum and bandwidth.
Additionally, building custom chipsets can help relieve supply chain problems. Early in the Covid pandemic, some supplier lead times grew to 50 weeks or more. Companies that used their own semiconductor technology were better able to manage product flows and protect shipments from external gridlock.
Staying ahead of today's technology and supply chain challenges requires strong, collaborative partnerships within the right ecosystem, enabling the advancement of custom semiconductors that will help us all create a more innovative and sustainable future.
Ee Huei Sin is President of the Electronic Industrial Solutions Group at Keysight Technologies. Based in Penang, Malaysia, she manages a broad portfolio of businesses that addresses the automotive, energy, general electronics, manufacturing, education, and semiconductor industries around the world. Huei Sin has been with Keysight and its predecessor companies, Agilent, and Hewlett-Packard, for more than 30 years and has held multiple leadership positions spanning engineering, marketing, manufacturing, order fulfillment, and business management. She holds a bachelor’s degree in Microelectronics & Physics from Campbell University, North Carolina, and an Executive Education certification from Columbia Business School.