Accenture Fab Future
Accenture Fab Future
Accenture Fab Future
the future:
Velocity with smart
manufacturing
1
Overview
With the rapid pace of technology innovation, it’s no surprise that today’s
fabs are continuously adapting, innovating, and transforming. They see huge
opportunities to leverage and advance the latest emerging high-growth areas
of Enhanced Reality, Internet of Things (IoT), Artificial Intelligence (AI),
Automotive, and 5G. At the same time, they still need to keep pace with the
conventional growth in demand as computing becomes pervasive in all aspects
of day-to-day life.
However, fabs also face a sobering reality: promise of sustainability, require fabs to not
Their complexity, scale, and variability make it simply adapt, but to do so quickly.
a constant challenge for fabs to balance critical
priorities: keeping manufacturing operating costs How can fabs achieve their priorities of
low, supply chains robust, processes integrated, providing cutting-edge innovations, optimizing
throughput high, quality uncompromised, operations, and creating sustainable values? The
and cycle time minimized. Further, challenges answer is to embrace digital technologies across
including increased regulatory challenges, the end-to-end supply chain to become a truly
pressure to stay ahead of the curve in a highly digitally connected and digital-led smart fab of
competitive environment, and delivering on the the future.1
Fabs and the ecosystem must A study by the Semiconductor Industry Association, With the development of advanced nodes, circuit
evolve from their conventional in partnership with Oxford Economics, found that designs are becoming more complicated, which in
structure and leverage modern building new fabs can create nearly 200,000 jobs turn means more resources to simulate, optimize,
and adds $25 billion annually to the US economy.2 and verify circuits are required. The industry
paradigms of technology and
Yet according to a recent survey from SEMI, 82% faces numerous challenges, including higher
manufacturing to realize their full of executives reported a shortage of qualified design costs, slower maturity of silicon processes
potential. Doing so means taking technical candidates,3 which adversely affects and design rules, and higher turnaround time
aim at several pivotal focus areas overall productivity. With growing demand for from the fab to achieve quality and yield targets.
digitally skilled resources, fabs need to find the Additionally, manual testing of physical designs
where the fab and its ecosystem
right people with right expertise, or equip existing leads to significantly greater cost and takes a lot
face major challenges today: talent with the skills to handle ever-changing digital of time. Fabs are also struggling to strike a balance
technologies and their use in manufacturing, between high R&D costs and ROI in designing
operations, and quality. A centrally managed and advanced nodes for manufacturing. Automation
technologically enabled collaborative learning of design processes leveraging generative design
and training digital center to upskill and reskill tools is desperately needed, along with exploring
workforces, at scale, has become a big priority. the potential of AI to boost design success rates,
superior performance and shorter payback period.
Fab wafer manufacturing is very capital intensive, Production of a new advanced node chip requires Covid-19 has led to a chip shortage, which has
and the development of advanced nodes more than 1,000 process steps. If something goes a cascading effect across industries, including
lower than 7nm will only increase the cost of wrong at any process step from oxidation to metal semiconductor manufacturers. In fact, the
manufacturing. However, developing new process deposition and etching, it can result into lower shortage has extended lead times for many
technologies and deploying them to production performance or even chip failure. At the same time, semiconductors out to more than a year. It’s difficult
phases takes considerable time. The disaggregated automotive, medtech, IoT, and other industrial to source multiple direct and indirect materials
design strategy is already mired in challenges applications require highly reliable chips over a long due to complex supplier networks spread across
involving combining components with different period of time. But traditional optical inspection geographies, and it takes at least five years to fund
yields, throughput times, and capacity constraints tools are ineffective in detecting defects. The quality and build a fab. As we have seen, these demand-
into a single package. Smaller and smaller chip sizes function is ripe for interconnected tools, equipment, and supply-side pressures, coupled with quality
will compound these challenges in the future due to and processes with advanced machine learning, issues and price fluctuations, can disrupt the supply
limitations in Moore’s Law. analytics, and AI capabilities that can identify chain—making a flexible and agile supply chain an
defects and process drift on a real-time basis.4 urgent priority for semiconductor manufacturing.5
PACKAGING SUSTAINABILIT Y
Packaging is becoming more complex as a result Information and computing technology is expected
of innovations in semiconductor fabrication. to account for as much as 20% of global energy
Multiple integration technology platforms have demand by 2030, and chip manufacturing
emerged that provide cost, size, performance, and accounts for most of the carbon output. In fact,
power optimizations to cater to different segments chip manufacturers have a larger carbon footprint
such as mobile computing, automotive, 5G, AI, than traditionally more polluting industries like
AR/VR, and IoT. However, these packages pose automotive.7 According to research at United
a challenge for traditional package design tools Nations University, 50% of the energy consumed
and methodologies struggle to keep up with new by a semiconductor chip over its useful life is
packaging requirements, and design teams must attributable to the original manufacturer. Moreover,
verify and optimize the entire system in such cases. the semiconductor manufacturing process is
The industry as a whole needs to make far greater responsible for significant water consumption as
progress in 3D modelling, prototyping, and digital well as chemical emissions. Smart and connected
twins for advanced packaging solutions.6 plants will help the semiconductor companies
to track and control the consumption of power,
emissions and water usage.
Quality Management
End-to End Quality Management Integrated Supply Chain
using sensors and advanced analytics Dynamic Production Planning/
for automated detection of defects Scheduling, Connected Suppliers/
and improving quality Distribution Centers, Remote Operations
approach to remotely monitor, assess,
guide and resolve issues from a central
Integrated Processes location driving optimization
Data from ERP, MES and key systems,
with seamless integration and Additional Sensors
& Visual data Sustainability
advanced analytics and sensors to
Use of technology as a catalyst to
derive actionable insights in real-time To provide additional data to
bring change in Environmental, Social
enable predictive models
and Governance ecosystem.
Centralized Energy & GHC Emission
Equipment PLCs Control Tower
Used for connected equipment for
smarter control and automation
Fab of the future 8
FAB OF THE FUTURE – MAKING INDUSTRY 4.0 LIVE
SMART BUILD
Leveraging digital in a fab’s “build” phase provides and coating to etching and metallization. This, in of manufacturing equipment. Production
the most convenient way of embedding digital turn, creates the foundational layer for developing a processes, down to the layer of real controllers,
DNA in the plant through connected systems and Digital Twin of the fab. can be modeled, simulated, and validated using
processes, smart layout, efficient planning, and Digital Twin to monitor, control, and optimize
integrated insights at every link of the fabrication Digital Twin can help virtually build and commission the relevant machines, production cells, and
value chain. The most important digital solution the fab processes and systems through digital production lines interaction.
from the perspective of fab in this phase is Digital modelling and simulation of the actual fab
Engineering – Digital Thread and Digital Twin. environment, with the learning from those exercises A Generative Design Studio, built on Digital Twin
being transferred to the actual implementation. and machine learning, can help create and optimize
Digital Engineering – Digital Twin and Digital Additionally, virtual plant simulation and production fab design plans. The optimized design generator
Thread and their role in building the fab of planning can help improve material flow planning helps populate initial fab designs by leveraging data
the future across the production facility, precisely detect and learnings from past projects, customization
manufacturing dynamic bottlenecks, increase needs, and cross-industry best practices, such
The right Digital Twin and Digital Thread strategy
throughput, and optimize energy consumption. as performance metrices, process requirements,
can open new and vast areas of opportunities for
Assembly operations simulations can help create locations, and budget.
the fab plant. Digital Thread connects processes
and optimize robot paths and efficient placement
across a fab operation’s value chain—from oxidation
Empowering operations and maintenance with However, with an enterprise-level, secure, By providing a consolidated platform for
digital can unlock enormous benefits and help get customized, and mature cloud solution now management, a cloud-enabled smart factory
the most out of a digital fab by unlocking a wide available for fabs, fabs need to embrace the can perform and accelerate decision making
range of new opportunities. Some of the key digital cloud for flexible, location-agnostic, and smarter on processing time, staging, raw material
solutions for smart operations and maintenance manufacturing. Cloud can create transparency management, and more—ultimately automating
include the following: in the value chain with increased visibility and manufacturing including setup changeovers, lay-
service consumption metrics. Real-time data outing and routings, and maximizing throughput.
Smart Factory enabled by Cloud and Digital accessibility and analysis of key business metrics Fabs can leverage the cloud to store and analyze
Twin for flexible production and operations on the cloud can drive better resource efficiency data and make the information readily available to
and cost stewardship. Cloud-accelerated innovation suppliers and vendors, and also standardize the
Cloud technology is nothing new to the
for semiconductor capabilities can also provide business processes across disparate sites for more
semiconductor industry;8 in fact, this game-
the highest levels of reliability and fastest pace insightful production and performance analysis.
changing industry has helped other sectors become
of innovation. Advanced analytics and AI allow interpreting and
the major disruptors they are today. However, when
connecting manufacturing (fab) on the cloud to
it comes to actual cloud adoption in a fab, progress
Fabs need to embrace the cloud help deliver relevant product and engineering
has been slow due to concerns over security or
insights for improving product development and
innovation capacity. for flexible, location-agnostic, and manufacturing quality.
smarter manufacturing.
A virtual production model for fab Virtual testing and simulation can improve overall quality deviations and faults. Sensor-based
equipment effectiveness (OEE), as well as enhance tracking, automated process steps validation,
operations and maintenance helps
resource and maintenance management. Finally, data visualization and analytics are also critical.
optimize throughput, yield, quality, a virtual production model for fab operations and Built across the end-to-end lifecycle of design
and cost with faster turnaround time. maintenance helps optimize throughput, yield, service, mask making, wafer manufacturing,
quality, and cost with faster turnaround time. backend service and packaging, they enable
Forging Digital Twin in fabs’ operations and proactive monitoring, faster potential failure
maintenance is yet another vital enabler for the End-to-End Quality Management for automated mode identification and effect analysis, improved
smart fabs of the future. Digital Twin-led9 realistic detection of defects and improving quality statistical process control, accurate measurement
simulations and tests in new product development system analysis, and optimal continual
With processes becoming more complex, rapid
can deliver useful information much earlier in improvement.10 This, in turn, makes a semiconductor
innovations compelling frequent changes, and
the design phase, leading to optimization of company “quality resilient”—a must for the fab
supply chains becoming more geographically
semiconductor material behavior, integration, of future to deliver the best services to the next
dispersed, end-to-end Quality Control and Quality
and advanced packaging. With real-time virtual generation of high-potential customers in highly
Assurance has be an integral part of each process
modelling of products and the production regulated industries.
step in a fab. Real-time, AI-driven monitoring
process, Digital Twin also provides insights into with integration of processes and built-in alarms
the condition of highly expensive semiconductor helps provide automated early detection of
equipment, products, and production environment.
Intelligent Insights from Integrated Processes processes is required to increase efficiency and
using ERP, MES and other key systems throughput, reduce costs, make faster decisions,
boost operations reliability, reduce lead time,
To achieve automation’s full potential of in
minimize variability, and assure quality.
operations and planning, fabs need to redesign their
legacy process flows with seamless cross-system
Integrated Supply Chain for end-to-end supply
and -application integration. Available data from
chain orchestration
Enterprise Resource Planning (ERP), Manufacturing
Execution System (MES), and other key systems— Advanced analytics implemented across the entire
coupled with seamless automated integration value chain provides insights and visibility into areas
with advanced digital technologies such as smart such as sales and operations planning, raw material
sensors, AI, advanced analytics, and cloud—can procurement, inventory management, supplier
generate key actionable, real-time, and intuitive management, and logistics. Analytics also enables
insights. These insights help a fab define, monitor, fabs to optimize their suppliers’ network in logistics
control, execute, and document best-in-class and implement remote operations via a shop floor
manufacturing operations, planning, scheduling, control tower. Overall, applying analytics to the
packaging, and shipping from anywhere, anytime.12 supply chain helps decrease time to market and
Seamless integration to remove siloed systems and increase demand and supply planning resiliency.
Sustainable Manufacturing for Net Zero such as carbon-neutral computing, carbon footprint Cyber Resilience
outflows and Circular Economy reduction, energy savings, water and wastewater
The fab of the future must be laser-focused on
management, industrial waste management,
Manufacturing sustainability is of paramount building cyber-resilience across all phases of fab
health and safety, and other natural resources
importance across any phase of the fab operations, operations through proactive monitoring of cyber
conservation. Additionally, a centralized energy and
be it build, operate, or maintain. Sustainability goals incidents and robust cybersecurity compliance
GHG emission control tower can help fabs address
for fabs include using technology as a catalyst to management. With products, processes, and plants
their sustainability challenges—particularly those
spur change in the Environmental, Social, and all connected, cybersecurity must be infused
focused on energy, emissions, water and waste
Governance (ESG) ecosystem. Digital technology from the very start—in the build phase of the fab
management—by providing accurate reporting,
with connected sensors, built-in KPIs, and live itself—and should extend its comprehensive, 24x7
increased visibility, and actionable insights related to
intuitive dashboards can serve as the foundation for coverage across all the value chain and life cycle
their sustainability initiatives and results.
innovation and collaboration in planning, execution, stages of the fab plant. Smart access management,
monitoring, analysis, and optimizing sustainability robust firewalls, secure remote access, IT security
measures across the entire enterprise—from Digital is the obvious enabler when vulnerability management, and integrated threat
strategy to operations to the supply chain—to it comes to monitoring, controlling, monitoring—along with a holistic, proactive
achieve end-to-end sustainable systems and cyber resiliency continuity plan for recovery and
and automating processes that
processes that drive business values. Digital is remediation—is what makes the fab of the future
the obvious enabler when it comes to monitoring, can help address environmental secure and safe.
controlling, and automating processes that can and climatic concerns.
help address environmental and climatic concerns,
Even better, many have already made clear Manufacturing Sustainability to adhere to Environmental, Social and Governance ecosystem
commitments to build and enable the smart
fabs using the solutions available today, Digital Manufacturing Analytics for actionable insights on key semiconductor manufacturing
yet with a vision for the future. Along their value drivers
journey, these companies will be driving
toward eight key imperatives a ”fab of the Manufacturing Collaboration Tower for seamless visibility in end-to-end manufacturing and
future” transformation requires, as illustrated intelligent decision making
in Figure 2. Two of these imperatives are
especially critical and worth calling out here. The possibilities are endless…
The first is Digital Manufacturing Analytics, which Importantly, Digital Manufacturing Analytics also performance indicators, automatic collection of
is needed to get the maximum outcome from the serves as the pivotal enabler of the second vital production and quality data, KPI visualization,
other seven imperatives. Digital Manufacturing imperative, a Manufacturing Collaboration analytics for OEE improvement, predictive insights
Analytics needs to be built as an integral capability Tower. Fabs must be able to proactively manage and proactive risk management. This holistic view
into all digital solutions to help drive actionable and control end-to-end internal manufacturing can help the fab more effectively manage demand,
insights on key semiconductor manufacturing operations, as well as ensure reliable and efficient supply, production, and maintenance, more
value drivers: machine performance (Overall collaboration with the external interfaces of design, efficiently use materials and assets, reduce costs,
Equipment Effectiveness (OEE), fault detection back-end services, and packaging. An overarching and minimize shop floor disruption.
and classification data), manufacturing quality Manufacturing Collaboration Tower spanning the
(metrology and defectivity), and yield (multivariate end-to-end build, operate, and maintain phases—
analysis and test time reduction).13 Additional assets and powered by the cloud, advanced analytics,
such as Bayesian modeling, AI algorithms, and AI, and proven simulation models—can provide
semantic layers can help create digital solutions a consolidated view and actionable insights for
and infrastructure that drive lower wafer costs, optimization across manufacturing operations. A
reduce yield variation and ramp up time, optimize Manufacturing Collaboration Tower helps achieve
depreciation/capital investment, and increase new efficiencies through the connected fab of the
overall machine utilization. future and enables real-time visualization of machine
1
Accenture research 6
https://www.eetimes.eu/5-keys-to-next-generation-ic- 10
https://www.tsmc.com/english/aboutTSMC/quality_
packaging-design/ and_reliability
2
Chipping In: The Positive Impact of the Semiconductor
Industry on the American Workforce and How 7
https://www.bloomberg.com/news/articles/2021-04-08/ 11
Intelligent PAM whitepaper 0.2 (Accenture)
Federal Industry Incentives Will Increase Domestic the-chip-industry-has-a-problem-with-its-giant-
Jobs – Semiconductor Industry Association carbon-footprint
12
Digital Transformation in Semiconductor
(semiconductors.org) Manufacturing, by Sophia Keil, Rainer Lasch, Fabian
8
https://www.accenture.com/_acnmedia/Accenture/ Lindner, Jacob Lohmer (page 22)
3
Engineering Talent Shortage Now Top Risk Factor Conversion-Assets/Secure/pdf-no-index-4/Accenture-
(semiengineering.com) Cloud-Imperative-Semiconductor-Industry.pdf
13
High Tech - Semiconductor Digital Manufacturing
Analytics (accenture.com)
4
https://semiengineering.com/defect-challenges-grow- 9
https://assets.new.siemens.com/siemens/assets/api/
for-ic-packaging/ uuid:763b9c75-0151-49bc-9595-082289e975fe/7795-
13-des-oem-ipdf-en-181009-2-150dpi.pdf
5
https://hbr.org/2021/02/why-were-in-the-midst-of-a-
global-semiconductor-shortage
Prasad Satyavolu
Managing Director
Industry X, North America
prasad.satyavolu@accenture.com
This document is produced by consultants at Accenture as general guidance. It is not intended to provide specific advice on your circumstances. If you require
advice or further details on any matters referred to, please contact your Accenture representative.
Copyright © 2021 Accenture.
All rights reserved. This document makes descriptive reference to trademarks that may be owned by others. The use of such trademarks herein is not an assertion of ownership
Accenture and its logo are registered trademarks of Accenture. of such trademarks by Accenture and is not intended to represent or imply the existence of an association between Accenture and the lawful owners of such
trademarks. No sponsorship, endorsement, or approval of this content by the owners of such trademarks is intended, expressed, or implied. 20