How is AI Transforming the Semiconductor Industry: Top Use Cases and Benefits

Oct 20, 2021
High-Tech | 5 min READ
This article was originally published in Electronics Maker magazine - Source link
Electronics Maker magazine Logo
The State of the Semiconductor Industry
As per the latest figures by Semiconductor Industry Association (SIA), worldwide sales of semiconductors were $41.8 billion in April 2021, an increase of 1.9% from the March 2021 total of $41.0 billion and 21.7% more than the April 2020 total of $34.4 billion. Additionally, as per the estimates, the annual global sales will increase 19.7% in 2021 and 8.8% in 2022.
Nitesh Mirchandani
Nitesh Mirchandani

VP and Global Head

Communications, Media & Technology (CMT) Vertical


As per the 16th annual KPMG global semiconductor industry outlook, which surveyed top global semiconductor executives, 50% of industry leaders stated that COVID-19 had accelerated digital transformation of the semiconductor industry. Despite this acceleration, the adoption of digital initiatives lags the tech sector overall (89%) and other industries (81%).
According to a McKinsey study, the application of AI/ML use cases delivers the most value—about 40% in optimizing semiconductor manufacturing efficiencies.
These use cases significantly improve the throughput of a Fabrication Plant (Fab). With consistent application, a Fab can expect cost-cutting to the range of 17%.
AI in Semiconductor Industry: Use Cases
The applications of AI span across the length and breadth of the semiconductor industry. Here's a look at the growing expanse of AI in the semiconductor industry:
Chip Development and Design
The procession time for chip production is one of the key challenges in the semiconductor supply chain. During this time, up to 30% of production costs are lost to testing and yield losses. Hence, embedding AI applications into the production cycle allows companies to systematically analyze losses at every production stage to assist manufacturers in optimizing operations.
Stay Ahead
Visit our High-Tech page
Working with next-generation semiconductor materials, this ability to assess losses becomes even more valuable as these materials tend to be more expensive (and volatile) than traditional silicon. If recent news is anything to go by, we have global companies joining the AI bandwagon in AI, including Google, Nvidia, and now Samsung. According to Wired, Samsung is using Synopsys AI software to design its Exynos chips, which are used in smartphones, including its own branded handsets, as well as other gadgets.
Visual Inspection of Wafers
Visual inspection of wafers helps ensure quality by detecting defects early in the front-end and back-end production process. As per McKinsey, with the help of advancements in deep learning technologies for computer vision, wafer-inspection systems are enhanced to automatically identify and classify defects at an accuracy rate that is on-par or better than human inspections.
This approach helps companies gain early insights on potential process or tool deviations, allowing them to detect problems earlier and improve yields, all while reducing costs.
Inventory Optimization
According to the 2020 McKinsey Global Survey on AI, 50% of businesses have adopted AI in at least one business function.
Findings also point out that the product or service-development and service-operations functions have the highest rate of AI adoption.
The largest shares of businesses have reported revenue increases for inventory and parts optimization, pricing and promotion, customer-service analytics, and sales and demand forecasting within these functions. AI enables physical tasks such as relocating and tracking items or more complicated processes requiring advanced insights for error-free planning or demand forecast.
 Top Use Cases and Benefits
How is AI Transforming the Semiconductor Industry: Top Use Cases and Benefits
Supplier Risk Management
COVID-19 has been a great teacher for semiconductor companies as it broke the shackles of primitive practices and opened the eyes of companies to diversify their supplier base. The perfect marriage of artificial intelligence will only strengthen their race to win the major market share.
Innovative digital and machine learning solutions (for example, multi-tier, multi-factor supplier sensing) that improve visibility into companies' supplier base, drive real-time decision-making for optimized capacity, and proactively manage supply chain risks will be instrumental in helping the industry address the business cycles, states Kearney report. Here artificial intelligence would play a critical role in the future.
 Top Use Cases and Benefits
Tail spend is an area that many sourcing managers find difficult to address regularly and with as much depth as they'd like. The 80% of suppliers that fall in this area are typically not diligently sourced. However, supplier selection and management, especially those positioned on the operational front line, is key in ensuring a high standard of service and customer perception. Using AI to help make the right choices helps the semiconductor manufacturers grow confidence, drive value, protect against risk and free up time at all levels to think more tactically.
Procurement and Material Planning
The procurement function has sadly often been a late adopter of modern technologies. Leveraging AI algorithms to track the supply and demand and analyze the production data provides actionable insights for the procurement team to meet market demands. The synergy of the procurement function with the other systems for determining the spend analytics improves the manufacturers' planning cycle, enhances the supply chain resilience, and leads the way to tap the hidden efficiency and savings potential.
Predictive Maintenance
As per McKinsey, the most critical use case for artificial intelligence/machine learning in the semiconductor industry is predictive maintenance.
Semiconductor manufacturers have hundreds of tools in each fab, with each tool generating terabytes of data of its own. It is impossible to analyze this data for any anomalies. Hence, AI turns out to be a boon in such situations for analyzing this data for identifying the slightest of functional anomalies.
Sales and Marketing Transformation
In the 'New Normal,' as digitalization picks up pace, semiconductor chips transcended all industries. The end-user penetration of smart devices is on the rise. Semiconductor manufacturers need to understand their customers' demand and buying patterns and be well prepared to cater to those demands. As per Mckinsey, AI/ML helps these manufacturers to analyze the data to study the usage patterns of their consumers to streamline the demand-supply gap.
It offers AI/ML models for effective sales and Marketing across lead management, budgeting, channel management, pricing and promotion, and providing buying recommendations to the customers. The servitization model offers huge opportunities for these manufacturers to explore end-consumer services through partnerships with product companies.
Fitch Ratings predicted that a global semiconductor shortage and increased demand for microchips are bound to drive the cash flows from operations for foundries and outsource assembly and testing (OSAT) companies in 2021. Foundries like TSMC and Samsung will be the biggest gainers in this regard.
To compete, semiconductor companies need to be more rapid and agile than ever. The advent of AI and big data will only strengthen their performance and profitability as AI can allow semiconductor companies to capture 40 to 50% of the total value from the technology stack.
Was this article helpful?