Enhancing the resilience of the semiconductor value chain requires a better understanding of the value chain and supply and demand of different semiconductor types. This enhanced understanding is not possible without better data to allow policymakers to identify bottlenecks, monitor the balance between demand and supply of specific semiconductor types, and manage disruptions in value chains.
A key objective of the OECD’s Semiconductor Informal Exchange Network (hereafter the SIEN), convened in June 2023, is to enhance the understanding of semiconductors and help participants move towards more resilient semiconductor value chains (OECD, n.d.[3]).1 This paper provides an overview of semiconductor front-end production capacity data, building on the semiconductor taxonomy developed by SIEN in 2024 (OECD, 2024[1]) and available data sources.
By focusing on improving transparency and information sharing on front-end manufacturing, this paper provides preliminary information based on the ongoing development of a database underpinning analyses to address the following key questions:
Where are the production facilities for front-end manufacturing located?
Analysis under this workstream sheds light on where various types of semiconductor production capacity are located and provides insights on market structure and thus potential bottlenecks. Moreover, data on the location, capacity and capabilities of front-end manufacturing facilities currently being planned or underway provide insights on whether recent policy efforts (e.g. policy strategies and instruments) are helping diversify semiconductor production and reduce the risks of bottlenecks in the future.
How is the balance between semiconductor demand and supply expected to evolve?
The planned approach is to combine the existing OECD Semiconductor Production Database with forthcoming data on chip demand to help identify segments where installed capacity is growing faster than demand, and therefore creating risks of excess capacity and unnecessary redundancies. While chip demand analysis is not the focus of this paper, future work combining chip capacity with demand data would help policymakers understand which types of chips investments should be encouraged to build a more resilient overall value chain.
By building on front-end manufacturing capacity information and semiconductor demand developments and forecasts, the analysis could help monitor industry cycles. Furthermore, the taxonomy facilitates distinguishing between different semiconductor product types, with the level of product type granularity depending on the availability of both production capacity and demand data at the same level of granularity.
What is the potential for substitution?
It is important to understand if and where production is substitutable between fabrication plants (fabs) in order to analyse gaps and consider policy approached focused on building more resilient value chains.
It would also help better understand the degree of flexibility in front-end chip manufacturing to cope with supply-demand imbalances for certain chips, including in the event of localised disruptions and other shocks.
Comprehending substitutability is critical when planning for managing worst-case disruption scenarios. There are two important dimensions to substitutability:
Chip substitutability describes the ability to replace one specific chip in an end-product with another that performs the same functions with a minimal loss in performance.
Fab substitutability describes the ability to shift the production of one chip to another plant, and if so, the range of chips that can be produced with the same available fab equipment and facilities (or with minor tweaks).
Further work towards addressing these questions will be necessary, including access to more comprehensive and granular data on chip types and process technologies per fab, as part of the development of the OECD semiconductor production database.
The paper is organised as follows: section 2 describes the various data sources, their limitations and how best to address them, section 3 provides the main findings. Finally, section 4 outlines future work in this workstream towards meeting the objectives and answering the policy questions described in section 1.