In the silicon industry, there are many contradictions at play as governments navigate the geopolitical landscape and companies compete to out-innovate each other. In this blog, I’ll explore two ways these tensions are playing out in chip design.
Semiconductor self-sufficiency vs. global collaboration
For the past few decades, the silicon industry has relied on a global supply chain to keep driving innovation forward while pushing costs down. However, the U.S. and Europe have become increasingly concerned with the lack of leading-edge chips manufactured domestically. Relying too much on any one country to manufacture chips makes the entire supply chain vulnerable. This was especially evident during the recent chip shortage, which started in 2020 due to the COVID-19 pandemic and delayed the production of cars, gaming consoles, laptops, and many other products. Many regions, including China, Europe, India, and the U.S., are now investing billions of dollars—through legislation such as the CHIPS and Science Act in the U.S. and the Chips Act in the EU—to boost their domestic semiconductor industries and become less reliant on foreign countries.
Even with this push for self-sufficiency, governments are finding ways to collaborate to further their shared interests. For example, the U.S.-EU Trade and Technology Council met in early April to discuss how to build more resilient semiconductor supply chains and identify potential disruptions. Open standards, like RISC-V, are also helping to foster collaboration, bringing together governments, the commercial sector, and academia to collaborate on the building blocks of silicon design. Semiconductor companies from the U.S. and Japan recently joined a roundtable with U.S. Secretary of Commerce Gina Raimondo and representatives from Japan’s Ministry of Economy, Trade and Industry (METI) to discuss opportunities for collaboration, including the role of RISC-V. SiFive, which participated in the roundtable, highlighted how RISC-V is pivotal for advancing innovation in AI, automotive, consumer devices, and beyond.
In the coming years, we’ll continue to see this push and pull between self-sufficiency and collaboration play out amid geopolitical tensions and the growing computing demands of the silicon industry.
Maximizing performance vs. efficiency
Silicon companies are in a constant battle between optimizing for performance vs. efficiency in chip design. While this has been an ongoing challenge for decades, the rise of AI applications is making chip design more complicated as performance demands skyrocket.
To meet today’s design challenges, companies are using a variety of new techniques to deliver better performance and efficiency. Foundries are leveraging backside power delivery techniques and gate-all-around transistors to maximize efficiency, along with High-NA EUV lithography. The silicon industry is rethinking the way chips are designed, turning to specialized solutions such as chiplets and AI accelerators, along with new approaches such as in-memory computing. Companies are using AI tools to design chips more quickly. Even as they introduce new techniques, chip designers still need to keep in mind the unique computing requirements of different applications to find the optimal balance of performance and efficiency.
Companies are exploring ways to harness external energy sources to power connected devices. For example, Atmosic is using controlled energy harvesting technology to capture energy from light, motion radio frequency, and thermal sources. Energy harvesting technology can be used to extend the battery life of IoT devices by years and to enable some devices to operate battery-free. Energy harvesting offers a sustainable, efficient solution to the battery waste that has become a growing issue in the IoT. The caveat is that energy harvesting isn’t yet feasible for high-performance, power-hungry devices like smartphones and laptops, but it is effective for low-power applications like sensors and beacons used in smart homes and enterprise and industrial environments.
While most consumers are not closely tracking these technical developments, not to mention the ins and outs of semiconductor geopolitics, it’s important to recognize how the decisions that are made will impact us all. I look forward to seeing how things continue to progress and what our everyday computing experiences will look like five years from now and further down the line.