These advanced technologies are reshaping semiconductor fabs

These advanced technologies are reshaping semiconductor fabs

By Sofiane Boukhalfa

The semiconductor industry and the billions of products they fabricate are helping to pave the way for advanced emerging technologies like advanced analytics, machine learning, and artificial intelligence that are radically transforming the market.

A tremendous continuous need for more powerful integrated circuits as reflected in the well-known Moore’s Law coupled with the barriers and complexity of small scale semiconductor physics have led to increased lead times for bringing these circuits to market with each node.

In order to reduce this lead time while maintain a cost advantage over their competitors, modern fabs are employing the very technologies they are helping develop. Here, we will explore how some fabs make use of emerging technologies, namely advanced analytics, the Internet of Things (IoT), machine learning and artificial intelligence to remain competitive and relevant in this rapidly evolving marketplace.

Advanced analytic capabilities and IoT:

Intelligent fabs today now offer significant advantages over more traditional manufacturing environments. Advanced analytic capabilities permit modern fabs to reduce manufacturing errors by identifying and testing possible points of failure throughout the manufacturing chain, to optimize various processes, and to improve yield and reliability by simulating runs. Moreover, the use of IoT sensors that are becoming omnipresent on tools and manufacturing lines allows manufacturers to identify sources of chip or equipment failure (in some cases before they happen).

These advanced analytic capabilities allow semiconductor companies to identify problems in many cases without the need for bringing on several teams of highly specialized and expensive process engineers to run trial runs to identify these problems – which can be extremely hard to spot in a process that involves the study of thousands of parameters that may be used in various combinations.

Advanced data analytics have also found much use in supply chain optimization in several industries, and the semiconductor industry is benefiting from rapid advances there as well. Coupling low cost sensors with big data analytics can help track inventory (including information such as when and where they were manufactured, the route they took to get there, and the various ambient conditions the product has been subject to, among others). These capabilities are also allowing semiconductor firms to identify sources of error quicker, and to better optimize their supply chain to reduce costs and improve yield times.

Machine learning and artificial intelligence:

Today, artificial intelligence can be installed on various tools to aid with chip inspection, increase yields, and improve the quality assurance process.

Traditionally, inspection tools were used for very specific applications, and to analyze the output of distinct steps in the manufacturing process. Installing so many tools in the manufacturing environment was costly, took up space, and increased the risk for damage during the inspection process.

In order to counter these problems, Nanotronics is building automated telescopes that make use of artificial intelligence and machine learning to aid in the semiconductor manufacturing process. These microscopes can analyze and find defects automatically, and can be used interchangeably at several steps in the manufacturing process. In an interview with McKinsey, Nanotronics claims:

“Our microscope might analyze 100,000 chips within minutes, while a manual inspector could require 30 minutes to look at 50. Fabs can also inspect more layers if they use our microscopes, rather than manual inspections. We worked with one company that inspected 25 layers manually but increased that to 300 with our microscopes. Then there’s the improvement in yield and throughput—fabs also see increases when they move from manual inspections to our microscopes.”

Such start-ups are no longer novel, with many players targeting the expensive semiconductor manufacturing industry with their advanced machine learning capabilities to reduce costs for their clients.

Future outlook:

With lead time for each new node generation increasing, and with the average cost of a fab increasing rapidly, the competitive and capital-heavy semiconductor industry is in a consolidation period. 83 fabs have been closed since the Great Recession of 2008. According to 24/7 Wall Street and Joe Dingalls, the following companies were consumed during mergers in the past 2 years:

  • Freescale Semiconductor
  • Integrated Silicon Solution
  • Altera
  • Atmel
  • Broadcom (name changed from Avago back to Broadcom post-merger)
  • EZchip Semiconductor
  • Hutchinson Technology
  • Mattson Technology
  • OmniVision Technologies
  • PMC-Sierra
  • KLA-Tencor
  • Fairchild Semiconductor
  • SanDisk

High manufacturing costs and increased complexity are driving both fab closures and M&A activity. Companies that wish to remain competitive in this sector will be forced to employ the technologies mentioned above.


Interested in exploring more technologies reshaping the semiconductor industry? Feel free to reach out to  Sofiane Boukhalfa (sboukhalfa@prescouter.com), PreScouter Project Architect and high-tech industry thought leader.

More of Our Insights & Work