Council Post: How Data Science Can Help Overcome The Global Chip Shortage

How Data Science Can Help Overcome The Global Chip Shortage

The modern world, which relies on silicon chips for almost everything, has been hit by a massive shortage of semiconductor chips. This is attributed to excessive demand in the market and supply chain challenges. The situation has birthed concerns over the supply chain and the need to diversify it, scale-up production, and maintain a good buffer. 

Even though the automotive industry makes up five to seven per cent of the total semiconductor industry, the shortage has caused automotive giants to slow down production and shut down their plants. Consumer electronics and power industries are some of the other major players hit by the shortage. The nature of these industries is interwoven with designers, chip makers, contract manufacturers and fab makers for semiconductor plants. 

The supply chain architecture integrates a lengthy value chain consisting of chip design, software, fabrication, equipment, chemicals, wafer and assembly. All the processes need to be efficiently working for the industry to work smoothly and meet the demands. 


Sign up for your weekly dose of what's up in emerging technology.

But, this chain is undergoing a recurring challenge given its diamond-shaped structure. As a result, the industry manufacturers are at the mercy of Tier 1 component integrators supplied at a limited rate by the semiconductor manufacturers. 

Leading to the Shortage

Download our Mobile App

The semiconductor chip industry is quite cyclical, and the increase in demand is not a new phenomenon. The chip industry has faced the long-term challenge of having a tightly balanced supply-demand ratio given the dynamic changes in demand. In comparison, building a semiconductor chip factory takes a couple of years, making it harder to meet the demands. But, this challenge just got exacerbated due to a combination of reasons. The waiting time for chip delivery gap has astronomically increased for Auto OEMs. This, to the extent that fresh capacity commissioning and allocation is still about a year away.

During the pandemic, there was a huge increase in demand for electronic devices that the industry couldn’t meet. Automotive manufacturers did not consider the few months it takes to reinstate orders that had been cancelled during the pandemic, leading to a larger delay in orders. 

Geopolitical factors have also played a huge role, specifically the trade war between the US and China. The US and China are the largest semiconductor markets, each accounting for 25% of global consumption. The Trump administration’s tight regulation of semiconductor sales to Chinese firms led to stockpiling chips essential to 5G smartphones and other products. In return, the American firms were being cut off from China’s Semiconductor Manufacturing International Corporation chips. While Biden views the semiconductor shortage as a “top and immediate priority” in the US, India doesn’t have any operational wafer fabrication plants and depends heavily on semiconductor imports.

This was only accompanied by local disruptions to the supply chain, such as the high rejection rates in Italy, winter storms in Texas, fire at the semiconductor plants in Japan, draught in Vietnam and the Covid Delta Variant in the Indonesia STM plant. 

Recovery Plan

This semiconductor shortage isn’t likely to resolve in the short term, given that chip capacity can’t catch up with demand immediately. But practices such as capacity allocation and management can help manufacturers meet the short product life cycles and customer demands in the shorter run. Automakers and tier-one suppliers have also begun to collect sophisticated intelligence on the chip value chain and manufacturing locations to make more informed decisions regarding factors such as contract terms and reassess the landscape. Additionally, manufacturers can leverage technology and analytics to match supply with demand, manage variants and prevent manual errors. 

In the longer term, manufacturers will need to rethink how they structure the industry’s plans. This includes capacity addition in the industry and creating a strategic road map that makes AI the top priority and emphasises the company’s total value. Additionally, mapping the semiconductor value chain that focuses on every part of the supply chain process helps ensure that all the steps are flexibly running to meet the demands. 

Leveraging AI

There is a need for collation at a global level to map the entire supply value chain efficiently. This entails collaboration between countries on new areas of electromobility, autonomous and connected vehicles. 

The supply-demand imbalance is going to continue in the foreseeable future. This means that the production of any new fabrication unit for the semiconductor plant will cost around fifteen to twenty billion dollars, not to mention the time to set up a plant, which can extend from three to five years. AI can step in here and help parts of the semiconductor value chain to quicken the pace. 

The application of AI/ML can dramatically accelerate the semiconductor industry over the next few years. AI can increase the speed involved in R&D by eliminating defects, incorrect processes, accelerating yield ramp up and suggesting ways to avoid time-consuming iterations. Additionally, they also automate physical layout designing and the verification process. This allows for quick stimulation of the chip technology optimisation for fully capable auto-grade chips with higher nanometer wafers. 

AI also plays a huge role in risk assessment by allowing manufacturers to spot upcoming trends in the market and make the most of the current opportunities. This reduces the risk posed by market shifts and logistics.

While it takes humans around six months or so to build one processor, AI can develop faster and more efficient processors through algorithms, with lesser time and a better combination of power & performance, as proved by Google. 

The utilisation of AI and predictive systems add the last layer to a perfect semiconductor supply chain process – continuously scanning for defects in real-time. This allows manufacturers to significantly improve the quality of their products, increase yield to meet the demands and prevent wastage of material and money by identifying the defective materials. 

An AI computer vision-enabled perfect manufacturing process flow leads to better production yields, i.e., 100 per cent accurate parts, given its process inspection and quality control. This acts as a game-changer for chip manufacturers to improve their production yield, reduce wastage and optimise the design. It does so by streamlining the supply chain disruption by mapping the value chain, predicting risks arising out of potential disruptions, and offering data analytics for quick decision-making. In the situation of the present shortage, this will also help conserve the limited resources and costs. 

The Three Dimensions of Policy Making

The semiconductor value chain is interdependent and interwoven with several industries, making it critical for governments to create policies that address the three important dimensions in the longer run. 

To ensure the global level of collaboration, government policies should focus on ensuring & securing access to foreign technology providers through the means of trade and foreign policy. Meanwhile, strategic industrial policies should be put in place to assist domestic companies in building leverage and strengthening themselves. Lastly, it is most important to have policies that foster and support a more resilient supply chain that can meet the demands!

This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill the form here.

More Great AIM Stories

Pradeep Mishra
Pradeep leads the Purchasing and Supply Chain function at VECV and is responsible for direct and indirect spending, supply chain operations and supplier quality reliability. He is currently developing a D2D (Demand to Delivery) Model using Data Science and Artificial Intelligence as bases for transformation. This model integrates demand data with the supply chain network, uses AI to predict risks and deploys analytics for risk mitigation decisions and easy governance.

AIM Upcoming Events

Conference, in-person (Bangalore)
Rising 2023 | Women in Tech Conference
16-17th Mar, 2023

Early Bird Passes expire on 10th Feb

Conference, in-person (Bangalore)
Data Engineering Summit (DES) 2023
27-28th Apr, 2023

3 Ways to Join our Community

Telegram group

Discover special offers, top stories, upcoming events, and more.

Discord Server

Stay Connected with a larger ecosystem of data science and ML Professionals

Subscribe to our Daily newsletter

Get our daily awesome stories & videos in your inbox